System Administration Guide: Devices and File Systems

Sun Microsystems, Inc. 4150 Network Circle Santa Clara, CA 95054 U.S.A.
Part No: 817509311 June 2005

Copyright 2005 Sun Microsystems, Inc.

4150 Network Circle, Santa Clara, CA 95054 U.S.A.

All rights reserved.

This product or document is protected by copyright and distributed under licenses restricting its use, copying, distribution, and decompilation. No part of this product or document may be reproduced in any form by any means without prior written authorization of Sun and its licensors, if any. Third-party software, including font technology, is copyrighted and licensed from Sun suppliers. Parts of the product may be derived from Berkeley BSD systems, licensed from the University of California. UNIX is a registered trademark in the U.S. and other countries, exclusively licensed through X/Open Company, Ltd. Sun, Sun Microsystems, the Sun logo, docs.sun.com, AnswerBook, AnswerBook2, JumpStart, Sun Ray, Sun Blade, PatchPro, SunOS, Solstice, Solstice AdminSuite, Solstice DiskSuite, Solaris Solve, Java, JavaStation, OpenWindows, NFS, iPlanet, Netra and Solaris are trademarks or registered trademarks of Sun Microsystems, Inc. in the U.S. and other countries. All SPARC trademarks are used under license and are trademarks or registered trademarks of SPARC International, Inc. in the U.S. and other countries. Products bearing SPARC trademarks are based upon an architecture developed by Sun Microsystems, Inc. FireWire and the FireWire logo are trademarks of Apple Computer, Inc, used under license. X/Open is a registered trademark of X/Open Company, Ltd. DLT is claimed as a trademark of Quantum Corporation in the United States and other countries. The OPEN LOOK and Sun Graphical User Interface was developed by Sun Microsystems, Inc. for its users and licensees. Sun acknowledges the pioneering efforts of Xerox in researching and developing the concept of visual or graphical user interfaces for the computer industry. Sun holds a non-exclusive license from Xerox to the Xerox Graphical User Interface, which license also covers Suns licensees who implement OPEN LOOK GUIs and otherwise comply with Suns written license agreements. U.S. Government Rights Commercial software. Government users are subject to the Sun Microsystems, Inc. standard license agreement and applicable provisions of the FAR and its supplements. DOCUMENTATION IS PROVIDED AS IS AND ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT, ARE DISCLAIMED, EXCEPT TO THE EXTENT THAT SUCH DISCLAIMERS ARE HELD TO BE LEGALLY INVALID. Copyright 2005 Sun Microsystems, Inc. 4150 Network Circle, Santa Clara, CA 95054 U.S.A. Tous droits rservs.

Ce produit ou document est protg par un copyright et distribu avec des licences qui en restreignent lutilisation, la copie, la distribution, et la dcompilation. Aucune partie de ce produit ou document ne peut tre reproduite sous aucune forme, par quelque moyen que ce soit, sans lautorisation pralable et crite de Sun et de ses bailleurs de licence, sil y en a. Le logiciel dtenu par des tiers, et qui comprend la technologie relative aux polices de caractres, est protg par un copyright et licenci par des fournisseurs de Sun. Des parties de ce produit pourront tre drives du systme Berkeley BSD licencis par lUniversit de Californie. UNIX est une marque dpose aux Etats-Unis et dans dautres pays et licencie exclusivement par X/Open Company, Ltd. Sun, Sun Microsystems, le logo Sun, docs.sun.com, AnswerBook, AnswerBook2, JumpStart, Sun Ray, Sun Blade, PatchPro, SunOS, Solstice, Solstice AdminSuite, Solstice DiskSuite, Solaris Solve, Java, JavaStation, DeskSet, OpenWindows, NFS et Solaris sont des marques de fabrique ou des marques dposes, de Sun Microsystems, Inc. aux Etats-Unis et dans dautres pays. Toutes les marques SPARC sont utilises sous licence et sont des marques de fabrique ou des marques dposes de SPARC International, Inc. aux Etats-Unis et dans dautres pays. Les produits portant les marques SPARC sont bass sur une architecture dveloppe par Sun Microsystems, Inc. FireWire et le logo de FireWire sont des marques d?pos?es de Applex Computer, Inc., utilis? sous le permis. X/Open est une marque de fabrique ou une marque d?pos?e de X/Open Company, Ltd.Quantum Corporation riclame DLT comme sa marque de fabrique aux Etats-Unis et dans dautres pays. Linterface dutilisation graphique OPEN LOOK et Sun a t dveloppe par Sun Microsystems, Inc. pour ses utilisateurs et licencis. Sun reconnat les efforts de pionniers de Xerox pour la recherche et le dveloppement du concept des interfaces dutilisation visuelle ou graphique pour lindustrie de linformatique. Sun dtient une licence non exclusive de Xerox sur linterface dutilisation graphique Xerox, cette licence couvrant galement les licencis de Sun qui mettent en place linterface dutilisation graphique OPEN LOOK et qui en outre se conforment aux licences crites de Sun. CETTE PUBLICATION EST FOURNIE EN LETAT ET AUCUNE GARANTIE, EXPRESSE OU IMPLICITE, NEST ACCORDEE, Y COMPRIS DES GARANTIES CONCERNANT LA VALEUR MARCHANDE, LAPTITUDE DE LA PUBLICATION A REPONDRE A UNE UTILISATION PARTICULIERE, OU LE FAIT QUELLE NE SOIT PAS CONTREFAISANTE DE PRODUIT DE TIERS. CE DENI DE GARANTIE NE SAPPLIQUERAIT PAS, DANS LA MESURE OU IL SERAIT TENU JURIDIQUEMENT NUL ET NON AVENU.

050418@11223

Contents
Preface 19

Managing Removable Media (Overview) 25 Whats New in Removable Media? 25 DVD+RW and DVD-RW Support 25 New cdrw Options 26 Listing Removable Media Devices 26 Where to Find Managing Removable Media Tasks 27 Removable Media Features and Benets 27 Comparison of Manual and Automatic Mounting 28 What You Can Do With Volume Management 29

Accessing Removable Media (Tasks) 31 Accessing Removable Media (Task Map) 31 Accessing Removable Media 32 Using Removable Media Names 32 Guidelines for Accessing Removable Media Data 34 Playing a Musical CD 34 M How to Add a New Removable Media Drive 35 M How to Stop and Start Volume Management (vold) 36 M How to Access Information on Removable Media 36 M How to Copy Information From Removable Media 37 M How to Play a Musical CD 38 M How to Determine If Removable Media Is Still in Use 38 M How to Eject Removable Media 39 Accessing Removable Media on a Remote System (Task Map) 40
3

M How to Make Local Media Available to Other Systems M How to Access Removable Media on Remote Systems

40 43

Formatting Removable Media (Tasks) 47 Formatting Removable Media (Task Map) 47 Formatting Removable Media 48 Formatting Removable Media Guidelines 48 Removable Media Hardware Considerations 48 M How to Load Removable Media 49 M How to Format Removable Media (rmformat) 51 M How to Format Removable Media for a File System 52 M How to Check a File System on Removable Media 54 M How to Repair Bad Blocks on Removable Media 54 Applying Read or Write Protection and Password Protection to Removable Media 55 M How to Enable or Disable Write Protection on Removable Media 55 M How to Enable or Disable Read or Write Protection and Set a Password on Iomega Media 55

Writing CDs and DVDs (Tasks) 59 Working With Audio CDs and Data CDs and DVDs 59 CD/DVD Media Commonly Used Terms 60 Writing CD and DVD Data and Audio CDs 61 Restricting User Access to Removable Media With RBAC 62 M How to Restrict User Access to Removable Media With RBAC M How to Identify a CD or DVD Writer 63 M How to Check the CD or DVD Media 63 Creating a Data CD or DVD 64 M How to Create an ISO 9660 File System for a Data CD or DVD M How to Create a Multi-Session Data CD 65 Creating an Audio CD 67 M How to Create an Audio CD 67 M How to Extract an Audio Track on a CD 68 M How to Copy a CD 69 M How to Erase CD-RW Media 70

62

64

Managing Devices (Tasks) 71 Whats New in Device Management?

71

System Administration Guide: Devices and File Systems June 2005

USB Device Enhancements 72 1394 (FireWire) and Mass Storage Support on x86 Systems 72 Device File System (devfs) 72 Power Management of Fibre Channel Devices 73 Where to Find Device Management Tasks 74 About Device Drivers 74 Automatic Conguration of Devices 75 Features and Benets of Autoconguration 76 What You Need for Unsupported Devices 76 Displaying Device Conguration Information 77 driver not attached Message 77 M How to Display System Conguration Information 77 Adding a Peripheral Device to a System 81 M How to Add a Peripheral Device 81 M How to Add a Device Driver 82

Dynamically Conguring Devices (Tasks) 85 Dynamic Reconguration and Hot-Plugging 85 Attachment Points 87 x86: Detaching PCI Adapter Cards 88 SCSI Hot-Plugging With the cfgadm Command (Task Map) 89 SCSI Hot-Plugging With the cfgadm Command 90 M How to Display Information About SCSI Devices 90 M How to Uncongure a SCSI Controller 91 M How to Congure a SCSI Controller 91 M How to Congure a SCSI Device 92 M How to Disconnect a SCSI Controller 93 M SPARC: How to Connect a SCSI Controller 94 M SPARC: How to Add a SCSI Device to a SCSI Bus 94 M SPARC: How to Replace an Identical Device on a SCSI Controller M SPARC: How to Remove a SCSI Device 96 Troubleshooting SCSI Conguration Problems 97 M How to Resolve a Failed SCSI Uncongure Operation 99 PCI Hot-Plugging With the cfgadm Command (Task Map) 99 PCI Hot-Plugging With the cfgadm Command 100 M How to Display PCI Slot Conguration Information 100 M How to Remove a PCI Adapter Card 101 M How to Add a PCI Adapter Card 102

95

Troubleshooting PCI Conguration Problems What Is an RCM Script? 104 104 104

103 103

Reconguration Coordination Manager (RCM) Script Overview What Can an RCM Script Do? RCM Script Tasks 105 106 106

How Does the RCM Script Process Work?

Application Developer RCM Script (Task Map) System Administrator RCM Script (Task Map) Naming an RCM Script 107 107 108 108 109 108 Installing or Removing an RCM Script M How to Install an RCM Script M How to Remove an RCM Script M How to Test an RCM Script Tape Backup RCM Script Example

Using USB Devices (Overview) Whats New in USB Devices? Overview of USB Devices USB Bus Description USB 2.0 Features 119 114

113 113 114 114

Solaris Support for USB Devices Commonly Used USB Acronyms 115 119 About USB in the Solaris OS

USB Keyboards and Mouse Devices USB Host Controller and Hubs SPARC: USB Power Management Guidelines for USB Cables 124 123

121 123

Using USB Devices (Tasks)

125 125 126

Managing USB Devices in the Solaris OS (Roadmap) Using USB Mass Storage Devices (Task Map) Using USB Mass Storage Devices Using USB Diskette Devices 127 129

Using Non-Compliant USB Mass Storage Devices Hot-Plugging USB Mass Storage Devices 130

129 130 131

M How to Add a USB Mass Storage Device With vold Running M How to Add a USB Mass Storage Device Without vold Running
6 System Administration Guide: Devices and File Systems June 2005

M How to Add a USB Camera 131 M How to Remove a USB Mass Storage Device With vold Running 132 M How to Remove a USB Mass Storage Device Without vold Running 133 Preparing to Use a USB Mass Storage Device With vold Running 133 M How to Prepare to Use USB Mass Storage Devices Without vold Running 134 M How to Display USB Device Information (prtconf) 135 M How to Format a USB Mass Storage Device Without vold Running 135 M How to Mount or Unmount a USB Mass Storage Device With vold Running 137 M How to Mount or Unmount a USB Mass Storage Device Without vold Running 138 Troubleshooting Tips for USB Mass Storage Devices 139 Disabling Specic USB Drivers 139 M How to Disable Specic USB Drivers 140 M How to Remove Unused USB Device Links 140 Using USB Audio Devices (Task Map) 141 Using USB Audio Devices 141 Hot-Plugging Multiple USB Audio Devices 142 M How to Add USB Audio Devices 143 M How to Identify Your Systems Primary Audio Device 143 M How to Change the Primary USB Audio Device 144 Troubleshooting USB Audio Device Problems 144 Hot-Plugging USB Devices With the cfgadm Command (Task Map) 145 Hot-Plugging USB Devices With the cfgadm Command 146 M How to Display USB Bus Information (cfgadm) 147 M How to Uncongure a USB Device 148 M How to Congure a USB Device 148 M How to Logically Disconnect a USB Device 149 M How to Logically Connect a USB Device 149 M How to Logically Disconnect a USB Device Subtree 150 M How to Reset a USB Device 150 M How to Change the Default Conguration of a Multi-Conguration USB Device 150

Using InniBand Devices (Overview/Tasks) 153 Overview of InniBand Devices 153 Dynamically Reconguring IB Devices (Task Map) 155 Dynamically Reconguring IB Devices (cfgadm) 156
7

M How to Display IB Device Information 157 M How to Uncongure an IOC Device 159 M How to Congure an IOC Device 159 M How to Uncongure an IB Port, HCA_SVC, or a VPPA Device 159 M How to Congure a IB Port, HCA_SVC, or a VPPA Device 160 M How to Uncongure an IB Pseudo Device 161 M How to Congure an IB Pseudo Device 161 M How to Display Kernel IB Clients of an HCA 161 M How to Uncongure IB Devices Connected to an HCA 162 Conguring an IB HCA 163 M How to Update the IB p_key Tables 163 M How to Display IB Communication Services 163 M How to Add a VPPA Communication Service 164 M How to Remove an Existing IB Port, HCA_SVC, or a VPPA Communication Service 164 M How to Update an IOC Conguration 165 Using the uDAPL Application Interface With InniBand Devices 165 M How to Enable uDAPL 166 Updating the DAT Static Registry 167 M How to Update the DAT Static Registry 167 M How to Register a Service Provider in the DAT Static Registry 167 M How to Unregister a Service Provider from the DAT Static Registry 168

10

Accessing Devices (Overview) 169 Accessing Devices 169 How Device Information Is Created 169 How Devices Are Managed 170 Device Naming Conventions 170 Logical Disk Device Names 171 Specifying the Disk Subdirectory 171 Direct and Bus-Oriented Controllers 172 x86: Disks With Direct Controllers 172 Disks With Bus-Oriented Controllers 173 Logical Tape Device Names 174 Logical Removable Media Device Names 174

11

Managing Disks (Overview) 175 Whats New in Disk Management? 175

System Administration Guide: Devices and File Systems June 2005

Multiterabyte Disk Support With EFI Disk Label 175 Common SCSI Drivers for SPARC and x86 Systems 180 New fdisk Partition Identier 180 Where to Find Disk Management Tasks 180 Overview of Disk Management 181 Disk Terminology 181 About Disk Slices 182 format Utility 185 About Disk Labels 189 Partition Table Terminology 189 Displaying Partition Table Information 190 Partitioning a Disk 191 Using the Free Hog Slice 192

12

Administering Disks (Tasks) 193 Administering Disks (Task Map) 193 Identifying Disks on a System 194 M How to Identify the Disks on a System 194 Formatting a Disk 196 M How to Determine if a Disk Is Formatted 196 M How to Format a Disk 197 Displaying Disk Slices 199 M How to Display Disk Slice Information 199 Creating and Examining a Disk Label 201 M How to Label a Disk 201 M How to Examine a Disk Label 203 Recovering a Corrupted Disk Label 204 M How to Recover a Corrupted Disk Label 205 Adding a Third-Party Disk 207 Creating a format.dat Entry 208 M How to Create a format.dat Entry 208 Automatically Conguring SCSI Disk Drives 209 M How to Automatically Congure a SCSI Drive 209 Repairing a Defective Sector 211 M How to Identify a Defective Sector by Using Surface Analysis M How to Repair a Defective Sector 213 Tips and Tricks for Managing Disks 213 Debugging format Sessions 213

211

Labeling Multiple Disks by Using the prtvtoc and fmthard Commands

214

13

SPARC: Adding a Disk (Tasks)

217 217 218 218 219 220 226

SPARC: Adding a System Disk or a Secondary Disk (Task Map) SPARC: Adding a System Disk or a Secondary Disk M SPARC: How to Connect a System Disk and Boot M SPARC: How to Connect a Secondary Disk and Boot M SPARC: How to Create Disk Slices and Label a Disk M SPARC: How to Create a UFS File System 225 M SPARC: How to Install a Boot Block on a System Disk

14

x86: Adding a Disk (Tasks)

227 227 228 229 230 231 232 233 238 240

x86: Adding a System Disk or a Secondary Disk (Task Map) x86: Adding a System Disk or a Secondary Disk M How to Change the Solaris fdisk Identier x86: Guidelines for Creating an fdisk Partition M x86: How to Create a Solaris fdisk Partition M x86: How to Create Disk Slices and Label a Disk M x86: How to Create File Systems 239 M x86: How to Install a Boot Block on a System Disk M x86: How to Connect a System Disk and Boot M x86: How to Connect a Secondary Disk and Boot

15

Conguring Solaris iSCSI Initiators (Tasks) The iSCSI Technology (Overview) 241

241 242 243 245 246

iSCSI Software and Hardware Requirements Setting Up Solaris iSCSI Initiators (Task Map) Conguring Solaris iSCSI Initiators 243

M How to Prepare for a Solaris iSCSI Conguration

Conguring Authentication in Your iSCSI-Based Storage Network M How to Congure CHAP Authentication for Your iSCSI Conguration 246

Using a Third-Party Radius Server to Simplify CHAP Management in Your iSCSI Conguration 247 M How to Congure RADIUS for Your iSCSI Conguration M How to Congure iSCSI Target Discovery M How to Remove Discovered iSCSI Targets
10 System Administration Guide: Devices and File Systems June 2005

247

248 249

Accessing iSCSI Disks

250 250 251 252 255 255 256 256

M Monitoring Your iSCSI Conguration

Modifying iSCSI Initiator and Target Parameters Troubleshooting iSCSI Conguration Problems 254

M How to Modify iSCSI Initiator and Target Parameters No Connections to the iSCSI Target From the Local System M How to Troubleshoot iSCSI Connection Problems iSCSI Device or Disk Is Not Available on the Local System General iSCSI Error Messages 256

M How to Troubleshoot iSCSI Device or Disk Unavailability

16

The format Utility (Reference)

263 263 264

Recommendations and Requirements for Using the format Utility format Menu and Command Descriptions partition Menu x86: fdisk Menu analyze Menu defect Menu format.dat File 266 267 268 269 270 271 271 271 275 274 275 275 276 277

Contents of the format.dat File Syntax of the format.dat File Partition Tables (format.dat) Rules for Input to format Commands Keywords in the format.dat File

Specifying an Alternate Data File for the format Utility Specifying Numbers to format Commands Specifying format Command Names Getting Help on the format Utility 277 276

Specifying Block Numbers to format Commands Specifying Disk Names to format Commands

17

Managing File Systems (Overview) Whats New in File Systems? NFS Version 4 libc_hwcap 280 279 UFS Logging Is Enabled by Default

279 279 281

64-bit: Support of Multiterabyte UFS File Systems 283

11

Where to Find File System Management Tasks Overview of File Systems Types of File Systems 284 284

283

Commands for File System Administration

290 291 291

How File System Commands Determine the File System Type Manual Pages for Generic and Specic File System Commands Default Solaris File Systems UFS File System UFS Logging UFS Snapshots 293 293 294 295 295 296 297 298 Planning UFS File Systems 291

UFS Direct Input/Output (I/O) The Mounted File System Table The Virtual File System Table The NFS Environment Automounting or AutoFS 299 300

Mounting and Unmounting File Systems

Determining a File Systems Type

300 300

How to Determine a File Systems Type

18

Creating UFS, TMPFS, and LOFS File Systems (Tasks) Creating a UFS File System 303 304 M How to Create a UFS File System

303

M How to Create a Multiterabyte UFS File System M How to Expand a Multiterabyte UFS File System

306 307 308 309

M How to Expand a UFS File System to a Multiterabyte UFS File System Troubleshooting Multiterabyte UFS File System Problems Creating a Temporary File System (TMPFS) M How to Create a TMPFS File System Creating a Loopback File System (LOFS) M How to Create an LOFS File System 310 310 311 312

19

Mounting and Unmounting File Systems (Tasks) Overview of Mounting File Systems Commonly Used Mount Options 315

315 316

Commands for Mounting and Unmounting File Systems 317 319 Field Descriptions for the /etc/vfstab File
12 System Administration Guide: Devices and File Systems June 2005

Mounting File Systems

320 320 321 322 323

How to Determine Which File Systems Are Mounted M How to Add an Entry to the /etc/vfstab File M How to Mount a File System (/etc/vfstab File)

M How to Mount a UFS File System (mount Command)

M How to Mount a UFS File System Without Large Files (mount Command) 324 M How to Mount an NFS File System (mount Command) 325 M x86: How to Mount a PCFS (DOS) File System From a Hard Disk (mount Command) 326 Unmounting File Systems 327 327 328 328 Prerequisites for Unmounting File Systems How to Verify a File System is Unmounted M How to Unmount a File System 329

M How to Stop All Processes Accessing a File System

20

Using The CacheFS File System (Tasks) Overview of the CacheFS File System How a CacheFS File System Works

331 331 332 332 333 334

High-Level View of Using the CacheFS File System (Task Map)

CacheFS File System Structure and Behavior M How to Create the Cache 335 335

Creating and Mounting a CacheFS File System (Task Map) Mounting a File System in the Cache

M How to Mount a CacheFS File System (mount) M How to Mount a CacheFS File System (AutoFS) Maintaining a CacheFS File System (Task Map) Maintaining a CacheFS File System Modifying a CacheFS File System 340 340 339

336 338 339

M How to Mount a CacheFS File System (/etc/vfstab)

M How to Display Information About a CacheFS File System Consistency Checking of a CacheFS File System M How to Delete a CacheFS File System Packing a Cached File System (Task Map) Packing a CacheFS File System 345 346 M How to Pack Files in the Cache 342 344 345 342 M How to Specify Cache Consistency Checking on Demand M How to Check the Integrity of a CacheFS File System

341 342

13

M How to Display Packed Files Information Using Packing Lists 348 348 M How to Create a Packing List

347

M How to Pack Files in the Cache With a Packing List Unpacking Files or Packing Lists From the Cache Troubleshooting cachefspack Errors Collecting CacheFS Statistics (Task Map) Collecting CacheFS Statistics 354 355 356 357 M How to Set Up CacheFS Logging M How to Locate the CacheFS Log File How to Stop CacheFS Logging Viewing CacheFS Statistics 358 358 357 350 354 349 M How to Unpack Files or Packing Lists From the Cache

349 349

M How to View the Working Set (Cache) Size M How to View CacheFS Statistics

21

Conguring Additional Swap Space (Tasks) About Swap Space 361 362 Swap Space and Virtual Memory Swap Space as a Dump Device

361

Swap Space and the TMPFS File System 363 Swap Space and Dynamic Reconguration How Do I Know If I Need More Swap Space? Swap-Related Error Messages TMPFS-Related Error Messages How Swap Space Is Allocated Planning for Swap Space Monitoring Swap Resources Adding More Swap Space Creating a Swap File 368 365 366 367 364 364 364

362 363 363

Swap Areas and the /etc/vfstab File

365

M How to Create a Swap File and Make It Available Removing a Swap File From Use 369 369 M How to Remove Unneeded Swap Space

368

22

Checking UFS File System Consistency (Tasks) File System Consistency 372

371

14

System Administration Guide: Devices and File Systems June 2005

How the File System State Is Recorded 372 What the fsck Command Checks and Tries to Repair 373 Why UFS File System Inconsistencies Might Occur 373 UFS Components That Are Checked for Consistency 374 fsck Summary Message 379 Interactively Checking and Repairing a UFS File System 380 M How to Check the root (/) or /usr File Systems From an Alternate Boot Device 381 M How to Check Non-root (/) or Non-/usr File Systems 383 Preening UFS File Systems 384 M How to Preen a UFS File System 385 Fixing a UFS File System That the fsck Command Cannot Repair 385 Restoring a Bad Superblock 386 M How to Restore a Bad Superblock 386 Syntax and Options for the fsck Command 388

23

UFS File System (Reference) 389 Structure of Cylinder Groups for UFS File Systems 389 Boot Block 390 Superblock 390 Inodes 390 Data Blocks 391 Free Blocks 392 Customizing UFS File System Parameters 392 Logical Block Size 393 Fragment Size 393 Minimum Free Space 394 Rotational Delay 394 Optimization Type 395 Number of Inodes (Files) 395 Maximum UFS File and File System Size 396 Maximum Number of UFS Subdirectories 396

24

Backing Up and Restoring File Systems (Overview) 397 Where to Find Backup and Restore Tasks 397 Introduction to Backing Up and Restoring File Systems 398 Why You Should Back Up File Systems 399 Planning Which File Systems to Back Up 399
15

Choosing the Type of Backup 401 Choosing a Tape Device 401 High-Level View of Backing Up and Restoring File Systems (Task Map) Considerations for Scheduling Backups 403 Guidelines for Scheduling Backups 404 Using Dump Levels to Create Incremental Backups 406 Sample Backup Schedules 407

402

25

Backing Up Files and File Systems (Tasks) 415 Backing Up Files and File System (Task Map) 415 Preparing for File System Backups 416 M How to Find File System Names 416 M How to Determine the Number of Tapes Needed for a Full Backup Backing Up a File System 417 M How to Back Up a File System to Tape 418

417

26

Using UFS Snapshots (Tasks) 425 Using UFS Snapshots (Task Map) 425 UFS Snapshots Overview 426 Why Use UFS Snapshots? 427 UFS Snapshots Performance Issues 427 Creating and Deleting UFS Snapshots 428 Creating a Multiterabyte UFS Snapshot M How to Create a UFS Snapshot Deleting a UFS Snapshot Backing Up a UFS Snapshot 431 432 433 433 432 429 430 428

M How to Display UFS Snapshot Information M How to Delete a UFS Snapshot

M How to Create a Full Backup of a UFS Snapshot (ufsdump) M How to Back Up a UFS Snapshot (tar) 434 434

M How to Create an Incremental Backup of a UFS Snapshot (ufsdump) Restoring Data From a UFS Snapshot Backup

27

Restoring Files and File Systems (Tasks) Preparing to Restore Files and File Systems Determining the File System Name

435 435 436

Restoring Files and File System Backups (Task Map) 437

16

System Administration Guide: Devices and File Systems June 2005

Determining the Type of Tape Device You Need Determining the Tape Device Name Restoring Files and File Systems 437 438 439 437

437

M How to Determine Which Tapes to Use M How to Restore Files Interactively

M How to Restore Specic Files Noninteractively M How to Restore a Complete File System 443

441 446

M How to Restore the root (/) and /usr File Systems

28

UFS Backup and Restore Commands (Reference) How the ufsdump Command Works Determining Device Characteristics Detecting the End of Media 452 451 451

451

Copying Data With the ufsdump Command Purpose of the /etc/dumpdates File Backup Device (dump-le) Argument Specifying Files to Back Up 455 455 455 Specifying Tape Characteristics 452 453

452

Limitations of the ufsdump Command Default ufsdump Options 456

Specifying ufsdump Command Options and Arguments The ufsdump Command and Security Issues 456 457

456

Specifying ufsrestore Options and Arguments

29

Copying UFS Files and File Systems (Tasks) Commands for Copying File Systems Copying File Systems Between Disks Making a Literal File System Copy M How to Copy a Disk (dd) 463 459 462 462

459

Copying Directories Between File Systems (cpio Command) M How to Copy Directories Between File Systems (cpio) Copying Files and File Systems to Tape Copying Files to Tape (tar Command) M How to Copy Files to a Tape (tar) M How to List the Files on a Tape (tar) 467 468 468 469 470 471

466 466

M How to Retrieve Files From a Tape (tar) Copying Files to a Tape With the pax Command

17

M How to Copy Files to a Tape (pax)

471 472 472 473 474 475 476

Copying Files to Tape With the cpio Command M How to List the Files on a Tape (cpio)

M How to Copy All Files in a Directory to a Tape (cpio) M How to Retrieve All Files From a Tape (cpio) Copying Files to a Remote Tape Device 476 477

M How to Retrieve Specic Files From a Tape (cpio)

M How to Copy Files to a Remote Tape Device (tar and dd) M How to Extract Files From a Remote Tape Device Copying Files and File Systems to Diskette 478

What You Should Know When Copying Files to Diskettes M How to Copy Files to a Single Formatted Diskette (tar) M How to List the Files on a Diskette (tar) Archiving Files to Multiple Diskettes 481 480 480 M How to Retrieve Files From a Diskette (tar)

478 479

30

Managing Tape Drives (Tasks) Choosing Which Media to Use Backup Device Names 484

483 483 485 486

Specifying the Rewind Option for a Tape Drive Specifying Different Densities for a Tape Drive Displaying Tape Drive Status 486 486 487 488 487 M How to Display Tape Drive Status Handling Magnetic Tape Cartridges

Retensioning a Magnetic Tape Cartridge Rewinding a Magnetic Tape Cartridge

Guidelines for Drive Maintenance and Media Handling

488

Index

489

18

System Administration Guide: Devices and File Systems June 2005

Preface
System Administration Guide: Devices and File Systems is part of a set that includes a signicant part of the Solaris system administration information. This guide contains information for both SPARC based and x86 based systems. This book assumes you have completed the following tasks:
I I

Installed the SunOS 5.10 Operating System Set up all the networking software that you plan to use

The SunOS 5.10 release is part of the Solaris product family, which also includes many features, including the Solaris Common Desktop Environment (CDE). The SunOS 5.10 operating system is compliant with AT&Ts System V, Release 4 operating system. For the Solaris 10 release, new features of interest to system administrators are covered in sections called Whats New in ... ? in the appropriate chapters.
Note This Solaris release supports systems that use the SPARC and x86 families of processor architectures: UltraSPARC, SPARC64, AMD64, Pentium, and Xeon EM64T. The supported systems appear in the Solaris 10 Hardware Compatibility List at http://www.sun.com/bigadmin/hcl. This document cites any implementation differences between the platform types.

In this document these x86 terms mean the following:

I I I

x86 refers to the larger family of 64-bit and 32-bit x86 compatible products. x64 points out specic 64-bit information about AMD64 or EM64T systems. 32-bit x86 points out specic 32-bit information about x86 based systems.

For supported systems, see Solaris 10 Hardware Compatibility List at http://www.sun.com/bigadmin/hcl.

19

Note Sun is not responsible for the availability of third-party web sites mentioned in this document. Sun does not endorse and is not responsible or liable for any content, advertising, products, or other materials that are available on or through such sites or resources. Sun will not be responsible or liable for any actual or alleged damage or loss caused by or in connection with the use of or reliance on any such content, goods, or services that are available on or through such sites or resources.

Who Should Use This Book

This book is intended for anyone responsible for administering one or more systems running the Solaris 10 release. To use this book, you should have 12 years of UNIX system administration experience. Attending UNIX system administration training courses might be helpful.

How the System Administration Volumes Are Organized

Here is a list of the topics that are covered by the volumes of the System Administration Guides.

Book Title

Topics

System Administration Guide: Basic Administration

User accounts and groups, server and client support, shutting down and booting a system, managing services, and managing software (packages and patches) Printing services, terminals and modems, system resources (disk quotas, accounting, and crontabs), system processes, and troubleshooting Solaris software problems Removable media, disks and devices, le systems, and backing up and restoring data

System Administration Guide: Advanced Administration

System Administration Guide: Devices and File Systems

20

System Administration Guide: Devices and File Systems June 2005

Book Title

Topics

System Administration Guide: IP Services

TCP/IP network administration, IPv4 and IPv6 address administration, DHCP, IPsec, IKE, IP lter, Mobile IP, IP network multipathing (IPMP), and IPQoS DNS, NIS, and LDAP naming and directory services, including transitioning from NIS to LDAP and transitioning from NIS+ to LDAP NIS+ naming and directory services Web cache servers, time-related services, network le systems (NFS and Autofs), mail, SLP, and PPP Auditing, device management, le security, BART, Kerberos services, PAM, Solaris cryptographic framework, privileges, RBAC, SASL, and Solaris Secure Shell Resource management topics projects and tasks, extended accounting, resource controls, fair share scheduler (FSS), physical memory control using the resource capping daemon (rcapd), and dynamic resource pools; virtualization using Solaris Zones software partitioning technology

System Administration Guide: Naming and Directory Services (DNS, NIS, and LDAP) System Administration Guide: Naming and Directory Services (NIS+) System Administration Guide: Network Services System Administration Guide: Security Services

System Administration Guide: Solaris Containers-Resource Management and Solaris Zones

Documentation, Support, and Training

Sun Function URL Description

Documentation http://www.sun.com/documentation/

Download PDF and HTML documents, and order printed documents Obtain technical support, download patches, and learn about Sun courses

Support and Training

http://www.sun.com/supportraining/

21

What Typographic Conventions Mean

The following table describes the typographic conventions used in this book.
TABLE P1 Typographic Conventions Typeface or Symbol Meaning Example

AaBbCc123

The names of commands, les, and directories; on screen computer output

Edit your .login le. Use ls -a to list all les. machine_name% you have mail.

AaBbCc123 AaBbCc123 AaBbCc123

What you type, contrasted with on screen machine_name% su computer output Password: Command-line placeholder: replace with a real name or value Book titles, new words or terms, or words to be emphasized To delete a le, type rm lename. Read Chapter 6 in Users Guide. These are called class options. Do not save changes yet.

Shell Prompts in Command Examples

The following table shows the default system prompt and superuser prompt for the C shell, Bourne shell, and Korn shell.
TABLE P2 Shell Prompts Shell Prompt

C shell prompt C shell superuser prompt Bourne shell and Korn shell prompt

machine_name% machine_name# $

Bourne shell and Korn shell superuser prompt #

22

System Administration Guide: Devices and File Systems June 2005

General Conventions
Be aware of the following conventions used in this book:
I

When following steps or using examples, be sure to type double-quotes ("), left single-quotes (), and right single-quotes () exactly as shown. The key referred to as Return is labeled Enter on some keyboards. The root path usually includes the /sbin, /usr/sbin, /usr/bin, and /etc directories. So, the steps in this book show the commands in these directories without absolute path names. Steps that use commands in other, less common, directories show the absolute paths in the examples. The examples in this book are for a basic SunOS software installation without the Binary Compatibility Package installed and without /usr/ucb in the path.
Caution If /usr/ucb is included in a search path, it should always be at the end of the search path. Commands such as ps or df are duplicated in /usr/ucb with different formats and options from the SunOS commands.

I I

23

24

System Administration Guide: Devices and File Systems June 2005

CHAPTER

Managing Removable Media (Overview)

This chapter provides general guidelines for managing removable media in the Solaris OS. This is a list of the overview information in this chapter.
I I I I I

Whats New in Removable Media? on page 25 Where to Find Managing Removable Media Tasks on page 27 Removable Media Features and Benets on page 27 Comparison of Manual and Automatic Mounting on page 28 What You Can Do With Volume Management on page 29

Whats New in Removable Media?

The following sections describe new removable media features in the Solaris 10 release.

DVD+RW and DVD-RW Support

In this Solaris release, you can use the cdrw command to create DVDs on DVD+RW or DVD-RW drives by rst creating the data with the mkisofs command. However, you cannot make multi-session data DVDs. These disks are then mounted and accessed as HSFS le systems. DVD+RW and DVD-RW devices are dened as follows:
I

DVD+RW Digital video disk (recordable/rewritable) drives can write both DVD-R discs, which can play back on most DVD players, and computer drives and DVD-RW rewritable disks
25

DVD-RW Digital video disk (rewritable) drives can be read only by DVD-RW drives

Both DVD+RW and DVD-RW devices are described generally as DVD devices in this guide unless specic information is required for DVD+RW devices or DVD-RW devices. The cdrw command uses Disk-At-Once (DAO) mode when writing DVDs, which does the following:
I I

Closes the media when writing is completed Prevents any further sessions from being added

The cdrw -d option must be used when writing the image to the DVD media since DAO mode requires that the size of the image needs to be known in advance. Keep the following key points in mind when working with DVD+RW devices:
I I

You cannot blank (or erase) DVD+RW media. You can re-use a DVD+RW media by writing a new image onto the media. The cdrw command automatically formats and overwrites the existing media.

For instructions on adding a USB mass storage class-compliant CD or DVD-RW device to your system, see scsa2usb(7D).

New cdrw Options

The following cdrw options have been added in this Solaris release to improve the management of media:
I

The -b fast option does a quick erase of the media. Instead of taking 10-15 minutes to erase the media, this option erases the media in about 30 seconds. Using this option erases the TOC of the media. If the media is damaged, you need to use -b all to clear the whole media. Use the -L option to unclose a previously closed CD-RW media. This option erases the last leadout and enables you to add more sessions to a multi-session CD-RW.

Listing Removable Media Devices

You can use the new rmformat -l option to list the removable media devices on the system. Using this option provides detailed information about the device such as the name used by vold and both the logical and physical device names. For example:
# rmformat -l Looking for devices... 1. Volmgt Node: /vol/dev/aliases/rmdisk1 26 System Administration Guide: Devices and File Systems June 2005

Logical Node: /dev/rdsk/c5t0d0s2 Physical Node: /pci@1e,600000/usb@b/hub@2/storage@4/disk@0,0 Connected Device: TEAC FD-05PUB 1026 Device Type: Floppy drive

For more information, see rmformat(1).

Where to Find Managing Removable Media Tasks

Use these references to nd step-by-step instructions for managing removable media.
Removable Media Management Task For More Information

Access removable media Format removable media Write data CDs and DVDs and music CDs

Chapter 2 Chapter 3 Chapter 4

For information on using removable media with File Manager in the Common Desktop Environment, see Solaris Common Desktop Environment: Users Guide.

Removable Media Features and Benets

The Solaris release gives users and software developers a standard interface for dealing with removable media. Referred to as volume management, this interface provides three major benets:
I

Automatically mounts removable media. For a comparison of manual and automatic mounting, see the following section. Enables you to access removable media without having to become superuser. Allows you to give other systems on the network automatic access to any removable media on your local system. For more information, see Chapter 2.

I I

Chapter 1 Managing Removable Media (Overview)

27

Comparison of Manual and Automatic Mounting

The following table compares the steps involved in manual mounting (without volume management) and automatic mounting (with volume management) of removable media.
TABLE 11 Steps

Comparison of Manual and Automatic Mounting of Removable Media

Manual Mounting Automatic Mounting

1 2 3

Insert media. Become superuser. Determine the location of the media device.

Insert media. For diskettes, use the volcheck command. Volume management (vold) automatically performs many of the tasks previously required to manually mount and work with removable media.

4 5 6 7 8 9 10 11 12

Create a mount point. Make sure you are not in the mount point directory. Mount the device and use the proper mount options. Exit the superuser account. Work with les on media. Become superuser. Unmount the media device. Eject media. Exit the superuser account. Eject media. Work with les on media.

28

System Administration Guide: Devices and File Systems June 2005

What You Can Do With Volume Management

Essentially, volume management enables you to access removable media just as manual mounting does, but more easily and without the need for superuser access. To make removable media easier to work with, you can mount removable media in easy-to-remember locations.
TABLE 12

(vold)
Access

How to Access Data on Removable Media Managed by Volume Management

Insert Find the Files Here

Files on the rst diskette Files on the rst removable hard disk Files on the rst CD Files on the rst DVD Files on the rst PCMCIA

The diskette and type /floppy volcheck on the command line The removable hard disk and type volcheck on the command line The CD and wait for a few seconds The DVD and wait for a few seconds /rmdisk/jaz0 or /rmdisk/zip0

/cdrom/volume-name

/dvd/volume-name

The PCMCIA and wait for a few /pcmem/pcmem0 seconds

If your system has more than one type of removable device, see the following table for their access points.
TABLE 13

Where to Access Removable Media

Access File Systems With This Path Access Raw Data With This Path

Media Device

First diskette drive Second diskette drive First CD-ROM drive

/floppy/floppy0 /floppy/floppy1 /cdrom/cdrom0

/vol/dev/aliases/floppy0 /vol/dev/aliases/floppy1 /vol/dev/aliases/cdrom0 /vol/dev/aliases/cdrom1 /vol/dev/aliases/jaz0 or /vol/dev/aliases/zip0 /vol/dev/aliases/pcmem0

Second CD-ROM drive /cdrom/cdrom1 First removable hard disk First PCMCIA drive /rmdisk/jaz0 or /rmdisk/zip0 /pcmem/pcmem0

Chapter 1 Managing Removable Media (Overview)

29

30

System Administration Guide: Devices and File Systems June 2005

CHAPTER

Accessing Removable Media (Tasks)

This chapter describes how to access removable media from the command line in the Solaris OS. For information on the procedures associated with accessing removable media, see the following:
I I

Accessing Removable Media (Task Map) on page 31 Accessing Removable Media on a Remote System (Task Map) on page 40

For background information on removable media, see Chapter 1.

Accessing Removable Media (Task Map)

The following task map describes the tasks for accessing removable media.
Task Description For Instructions

1. (Optional) Add the removable media drive. 2. (Optional) Decide whether you want to use removable media with or without volume management (vold).

Add the removable media drive to your system, if necessary.

How to Add a New Removable Media Drive on page 35

How to Stop and Start Volume management (vold) runs by default. Decide Volume Management (vold) whether you want to use on page 36 removable media with or without volume management.

31

Task

Description

For Instructions

3. Access removable media.

Access different kinds of removable media with or without volume management running. Copy les or directories from the media as you would from any other location in the le system. You can congure a system to play musical CDs. However, you will need third-party software to play the media.

How to Access Information on Removable Media on page 36 How to Copy Information From Removable Media on page 37 How to Play a Musical CD on page 38

4. (Optional) Copy les or directories.

5. (Optional) Congure a system to play musical CDs.

6. Find out if the media is still in use. 7. Eject the media.

Before ejecting the media, nd How to Determine If out if it is still in use. Removable Media Is Still in Use on page 38 When you nish, eject the media from the drive. How to Eject Removable Media on page 39

Accessing Removable Media

You can access information on removable media with or without using volume management. For information on accessing information on removable media with CDEs File Manager, see Using Removable Media with File Manager in Solaris Common Desktop Environment: Users Guide. Starting in the Solaris 8 6/00 release, volume management (vold) actively manages all removable media devices. So, any attempt to access removable media with device names such as /dev/rdsk/cntndnsn or /dev/dsk/cntndnsn will be unsuccessful.

Using Removable Media Names

You can access all removable media with different names. The following table describes the different media names that can be accessed with or without volume management.

32

System Administration Guide: Devices and File Systems June 2005

TABLE 21

Removable Media Names

Volume Management Device Name Volume Management Device Alias Name Device Name

Media

First diskette drive

/floppy

/vol/dev/aliases/floppy0

/dev/rdiskette /vol/dev/rdiskette0/ volume-name

First, second, /cdrom0 third CD-ROM /cdrom1 or DVD-ROM drives /cdrom2 First, second, third Jaz drive /rmdisk/jaz0 /rmdisk/jaz1 /rmdisk/jaz2 First, second, third Zip drive /rmdisk/zip0 /rmdisk/zip1 /rmdisk/zip2 First, second, third PCMCIA drive /pcmem/pcmem0 /pcmem/pcmem1 /pcmem/pcmem2

/vol/dev/aliases/cdrom0 /vol/dev/aliases/cdrom1 /vol/dev/aliases/cdrom2 /vol/dev/aliases/jaz0 /vol/dev/aliases/jaz1 /vol/dev/aliases/jaz2 /vol/dev/aliases/zip0 /vol/dev/aliases/zip1 /vol/dev/aliases/zip2 /vol/dev/aliases/pcmem0 /vol/dev/aliases/pcmem1 /vol/dev/aliases/pcmem2

/vol/dev/rdsk/cntn[dn]/ volume-name

/vol/dev/rdsk/cntndn/ volume-name

/vol/dev/rdsk/cntndn/ volume-name

/vol/dev/rdsk/cntndn/ volume-name

Use this table to identify which removable media name to use with specic Solaris commands.

Solaris Command

Device Name

Usage Examples

ls, more, vi

/floppy /cdrom /rmdisk/zip0 /rmdisk/jaz0 /pcmem/pcmem0

ls /floppy/myfiles/ more /cdrom/myfiles/filea

fsck, newfs, mkfs

/vol/dev/aliases/floppy0 newfs /vol/dev/aliases/floppy0 /vol/dev/rdsk/cntndnsn mkfs -F udfs /vol/dev/rdsk/cntndnsn

Chapter 2 Accessing Removable Media (Tasks)

33

Guidelines for Accessing Removable Media Data

Most CDs and DVDs are formatted to the ISO 9660 standard, which is portable. So, most CDs and DVDs can be mounted by volume management. However, CDs or DVDs with UFS le systems are not portable between architectures. So, they must be used on the architecture for which they were designed. For example, a CD or DVD with a UFS le system for a SPARC platform cannot be recognized by an x86 platform. Likewise, an x86 UFS CD cannot be mounted by volume management on a SPARC platform. The same limitation generally applies to diskettes. However, some architectures share the same bit structure, so occasionally a UFS format specic to one architecture will be recognized by another architecture. Still, the UFS le system structure was not designed to guarantee this compatibility. To accommodate the different formats, the CD or DVD is split into slices. Slices are similar in effect to partitions on hard disks. The 9660 portion is portable, but the UFS portion is architecture-specic. If you are having trouble mounting a CD or DVD, particularly if it is an installation CD or DVD, make sure that its UFS le system is appropriate for your systems architecture. For example, you can check the label on the CD or DVD.

Accessing Jaz Drives or Zip Drives

You can determine whether accessing your Jaz drives or Zip drives changes from previous Solaris releases, depending on the following:
I

If you are upgrading from the Solaris 8 6/00 release to the Solaris 10 release, you can continue to access your Jaz drives and Zip drives in the same way as in previous releases. If you are freshly installing the Solaris 10 release, you cannot access your Jaz drives and Zip drives in the same way as in previous Solaris releases. Follow these steps if you want to access your Jaz drives and Zip drives in the same way as in previous Solaris releases: 1. Comment the following line in the /etc/vold.conf le by inserting a pound (#) sign at the beginning of the text. For example:
# use rmdisk drive /dev/rdsk/c*s2 dev_rmdisk.so rmdisk%d

2. Reboot the system.

Playing a Musical CD
To play musical media from a media drive attached to a system running the Solaris release, you need to access public domain software, such as xmcd, that is available from the following locations: http://www.ibiblio.org/tkan/xmcd
34 System Administration Guide: Devices and File Systems June 2005

This site includes frequent updates to the xmcd software. This software includes the version of xmcd that plays on newer Sun hardware, such as the Sun Blade systems. http://www.sun.com/software/solaris/freeware/pkgs_download.html Keep the following in mind when using the xmcd software with CDDA (CD Digital Audio format) support to play musical media:
I

Use xmcd, version 3.1 (or later) on Sun Blade systems. This version has CDDA support, which must be enabled in order to listen to CDs on these systems. Enable CDDA by launching xmcd, clicking on the options button, and then by clicking on CDDA playback. Note that the options button has a hammer and screwdriver on the button. When CDDA is enabled, audio is directed to the audio device. So, headphones and external speakers should be connected to the audio device and not to the media drive itself. CDDA can be enabled on other machines, too. Enabling CDDA is required for playing media on Sun Blade systems.

Consider the following issues as well:

I

If you are using xmcd with standard playback on a system that does not have an internal connection from the CD-ROM to the audio device, you must insert headphones into the CD-ROM drives headphone port. If you are using xmcd with standard playback on a system that does have an internal connection from the CD-ROM to the audio device, you can do either of the following:
I I

Insert headphones into the headphone port of the CD-ROM drive. Insert headphones into the headphone port on the audio device.

If you choose the second option, you must do the following from sdtaudiocontrols record panel:
I I

Select the internal CD as the input device. Make sure that Monitor Volume is non-zero.

How to Add a New Removable Media Drive

Generally, most modern bus types support hot-plugging. If your systems bus type supports hot-plugging, you might only need to do step 5 below. If your systems bus type does not support hot-plugging, you might have to do the following tasks, which are described in steps 1-6 below.
I I

Create the /reconfigure le. Reboot the system so that volume management recognizes the new media drive.

For more information about hot-plugging devices, see Chapter 6.

Chapter 2 Accessing Removable Media (Tasks) 35

Steps

1. Become superuser. 2. Create the /reconfigure le.

# touch /reconfigure

3. Bring the system to run level 0.

# init 0

4. Turn off power to the system. 5. Connect the new media drive. See your hardware handbook for specic instructions. 6. Turn on power to the system. The system automatically comes up to multiuser mode.

How to Stop and Start Volume Management (vold)

Occasionally, you might want to manage media without using volume management. This section describes how to stop and restart volume management.

Steps

1. Ensure that the media is not being used. If you are not sure whether you have found all users of the media, use the fuser command, seeHow to Determine If Removable Media Is Still in Use on page 38. 2. Become superuser. 3. Select one of the following:
I

Stop volume management.

# /etc/init.d/volmgt stop #

Start volume management.

# /etc/init.d/volmgt start volume management starting.

M
Steps

How to Access Information on Removable Media

1. Insert the media. The media is mounted after a few seconds.

36

System Administration Guide: Devices and File Systems June 2005

2. Check for media in the drive.

% volcheck

3. List the contents of the media.

% ls /media

Use the appropriate device name to access information by using the command-line interface. See Table 21 for an explanation of device names.
Example 21

Accessing Information on Removable Media

This example shows how to access information on a diskette.
$ volcheck $ ls /floppy myfile

This example shows how to access information on a Jaz drive.

$ volcheck $ ls /rmdisk jaz0/ jaz1/

This example shows how to access information on a CD-ROM.

$ volcheck $ ls /cdrom cdrom0@

sol_10_sparc/

This example shows how to view the symbolic links on a CD-ROM.

$ ls -lL /cdrom/cdrom0 total 24 dr-xr-xr-x 2 root drwxr-xr-x 18 root drwxr-xr-x 2 root drwxr-xr-x 2 root drwxr-xr-x 2 root drwxr-xr-x 2 root

sys root root root root root

2048 512 512 512 512 512

Dec Dec Dec Dec Dec Dec

3 3 3 3 3 3

11:54 13:09 13:10 13:10 13:10 13:10

s0/ s1/ s2/ s3/ s4/ s5/

This example shows how to access information on a PCMCIA memory card as follows
$ ls /pcmem/pcmem0 pcmem0 myfiles

How to Copy Information From Removable Media

You can access les and directories on removable media as with any other le system. The only signicant restrictions are related to ownership and permissions. For instance, if you copy a le from a CD into your le system, you are the owner. However, you wont have write permissions because the le on the CD never had them. You must change the permissions yourself.
Chapter 2 Accessing Removable Media (Tasks) 37

Steps

1. Ensure that the media is mounted.

$ ls /media

The ls command displays the contents of a mounted media. If no contents are displayed, see How to Access Information on Removable Media on page 36. 2. (Optional) Copy the les or directories. For example, for a CD, you would do the following:
$ cp /cdrom/sol_9_1202_sparc/s0/Solaris_9/Tools/add_install_client . $ ls -l -rwxr-xr-x 1 pmorph gelfs 59586 Jan 16 2004 add_install_client*

For example, for a PCMCIA memory card, you would do the following:
$ cp /pcmem/pcmem0/readme2.doc . $ cp -r /pcmem/pcmem0/morefiles .

How to Play a Musical CD

You will need to install public domain software to play musical CDs. For more information, see Playing a Musical CD on page 34. After you install the xmcd software, you can play a musical CD simply by inserting it into the CD-ROM drive and starting the xmcd control panel.

Steps

1. Install the xmcd software. 2. Insert the media into the media drive. 3. Start the media player.
% ./xmcd &

How to Determine If Removable Media Is Still in Use

1. Become superuser. 2. Identify the processes that are accessing the media.
# fuser -u /media

Steps

The -u displays the user of the media. For more information, see fuser(1M). 3. (Optional) Kill the process accessing the media.
# fuser -u -k /media 38 System Administration Guide: Devices and File Systems June 2005

The -k kills the processes accessing the media.

Caution Killing the processes that are accessing the media should only be used in emergency situations.

4. Verify that the process is gone.

# pgrep process-ID

Example 22

Determining If the Media Is Still in Use

The following example shows that the process 7258c, owner pmorph, is accessing the /cdrom/sol_9_1202_sparc/s0/Solaris_9/Tools/ directory.
# fuser -u /cdrom/sol_9_1202_sparc/s0/Solaris_9/Tools/ /cdrom/sol_9_1202_sparc/s0/Solaris_9/Tools/: 7258c(pmorph)

M
Steps

How to Eject Removable Media

1. Ensure that the media is not being used. Remember, media is being used if a shell or an application is accessing any of its les or directories. If you are not sure whether you have found all users of a CD (for example, a shell hidden behind a desktop tool might be accessing it), use the fuser command. See How to Determine If Removable Media Is Still in Use on page 38. 2. Eject the media.
# eject media

For example, for a CD, you would do the following:

# eject cdrom

For example, for a PCMCIA memory card, you would do the following:
# eject pcmem0

Chapter 2 Accessing Removable Media (Tasks)

39

Accessing Removable Media on a Remote System (Task Map)

The following task map describes the tasks need to access removable media on a remote system.

Task

Description

For Instructions

1. Make local media available to remote systems.

congure your system to How to Make Local Media share its media drives to make Available to Other Systems any media in those drives on page 40 available to other systems. How to Access Information on Removable Media on page 36

2. Access removable media on Access the remote media on remote systems. the local system.

How to Make Local Media Available to Other Systems

You can congure your system to share its media drives to make any media in those drives available to other systems. One exception is musical CDs. Once your media drives are shared, other systems can access the media they contain simply by mounting them. For instructions, see How to Access Removable Media on Remote Systems on page 43.

Steps

1. Become superuser. 2. Create a dummy directory to share.

# mkdir /dummy

The dummy mount point can be any directory name, for example, /dummy. This directory will not contain any les. Its only purpose is to wake up the NFS daemon so that it notices your shared media drive. 3. Add the following entry to the /etc/dfs/dfstab le.
share -F nfs -o ro /dummy

When you start the NFS server service, it will encounter this entry, wake up, and notice the shared media drive. 4. Determine whether the NFS server service is running.
# svcs *nfs* 40 System Administration Guide: Devices and File Systems June 2005

The following output is returned from the svcs command if NFS server service is running:
online 14:28:43 svc:/network/nfs/server:default

5. Identify the NFS server status, and select one of the following:
I I

If the NFS server service is running, go to Step 7. If the NFS server service is not running, go to the next step.

6. Start the NFS server service.

# svcadm enable -t network/nfs/server

Verify that the NFS daemons are running. For example:

# svcs -p svc:/network/nfs/server:default STATE STIME FMRI online Aug_30 svc:/network/nfs/server:default Aug_30 319 mountd Aug_30 323 nfsd

7. Eject any media currently in the drive.

# eject media

8. Assign root write permissions to the /etc/rmmount.conf le.

# chmod 644 /etc/rmmount.conf

9. Add the following lines to the /etc/rmmount.conf le:

share media*

These lines share any media loaded into your systems CD-ROM drive. You can, however, limit sharing to a particular CD or series of CDs, as described in share(1M). 10. Remove write permissions from the /etc/rmmount.conf le.
# chmod 444 /etc/rmmount.conf

This step returns the le to its default permissions. 11. Load the media. The media you now load, and all subsequent media, is available to other systems. Remember to wait until the light on the drive stops blinking before you verify this task. To access the media, the remote user must mount it by name, according to the instructions in How to Access Removable Media on Remote Systems on page 43. 12. Verify that the media is indeed available to other systems.

Chapter 2 Accessing Removable Media (Tasks)

41

If the media is available, its share conguration is displayed. The shared dummy directory is also displayed.
# share /dummy ro /cdrom/sol_9_1202_sparc/s5 /cdrom/sol_9_1202_sparc/s4 /cdrom/sol_9_1202_sparc/s3 /cdrom/sol_9_1202_sparc/s2 /cdrom/sol_9_1202_sparc/s1 /cdrom/sol_9_1202_sparc/s0

ro ro ro ro ro ro

"" "" "" "" "" ""

Example 23

Making Local CDs Available to Other Systems

The following example shows how to make any local CD available to other systems on the network.
# mkdir /dummy vi /etc/dfs/dfstab (Add the following line:) # share -F nfs -o ro /dummy # svcs *nfs* # svcadm enable -t network/nfs/server # svcs -p svc:/network/nfs/server:default # eject cdrom0 # chmod 644 /etc/rmmount.conf # vi /etc/rmmount.conf (Add the following line:) share cdrom* # chmod 444 /etc/rmmount.conf (Load a CD.) # share /dummy ro "" /cdrom/sol_9_1202_sparc/s5 /cdrom/sol_9_1202_sparc/s4 /cdrom/sol_9_1202_sparc/s3 /cdrom/sol_9_1202_sparc/s2 /cdrom/sol_9_1202_sparc/s1 /cdrom/sol_9_1202_sparc/s0

ro ro ro ro ro ro

"" "" "" "" "" ""

Example 24

Making Local Diskettes Available to Other Systems

The following example shows how to make any local diskette available to other systems on the network.
# mkdir /dummy # vi /etc/dfs/dfstab (Add the following line:) share -F nfs -o ro /dummy # svcs *nfs* # svcadm enable -t network/nfs/server # svcs -p svc:/network/nfs/server:default # eject floppy0

42

System Administration Guide: Devices and File Systems June 2005

# chmod 644 /etc/rmmount.conf # vi /etc/rmmount.conf (Add the following line:) share floppy* # chmod 444 /etc/rmmount.conf (Load a diskette.) # volcheck -v media was found # share /dummy ro "" /floppy/myfiles

rw

""

Example 25

Making Local PCMCIA Memory Cards Available to Other Systems

The following example shows how to make any local PCMCIA memory card available to other systems on the network.
# mkdir /dummy # vi /etc/dfs/dfstab (Add the following line:) # svcs *nfs* # share -F nfs -o ro /dummy # svcadm enable -t network/nfs/server # svcs -p svc:/network/nfs/server:default # eject pcmem0 # chmod 644 /etc/rmmount.conf # vi /etc/rmmount.conf (Add the following line:) share floppy* svc:/network/nfs/server:default# chmod 444 /etc/rmmount.conf (Load a PCMCIA memory card.) # volcheck -v media was found # share /dummy ro "" /pcmem/myfiles rw ""

How to Access Removable Media on Remote Systems

You can access media on a remote system by manually mounting the media into your le system. Also, the remote system must have shared its media according to the instructions in How to Make Local Media Available to Other Systems on page 40.

Steps

1. Select an existing directory to serve as the mount point. Or create a mount point.
$ mkdir /directory

where /directory is the name of the directory that you create to serve as a mount point for the remote systems CD.
Chapter 2 Accessing Removable Media (Tasks) 43

2. Find the name of the media you want to mount.

$ showmount -e system-name

3. As superuser, mount the media.

# mount -F nfs -o ro system-name:/media/media-name local-mount-point

system-name: media-name local-mount-point

Is the name of the system whose media you will mount. Is the name of the media you want to mount. Is the local directory onto which you will mount the remote media.

4. Log out as superuser. 5. Verify that the media has been mounted.
$ ls /media

Example 26

Accessing CDs on Remote Systems

The following example shows how to automatically access the remote CD named sol_9_1202_sparc from the remote system starbug using AutoFS.
$ showmount -e starbug export list for starbug: /dummy /cdrom/sol_9_1202_sparc/s5 /cdrom/sol_9_1202_sparc/s4 /cdrom/sol_9_1202_sparc/s3 /cdrom/sol_9_1202_sparc/s2 /cdrom/sol_9_1202_sparc/s1 /cdrom/sol_9_1202_sparc/s0 $ ls /net/starbug/cdrom/ Copyright Solaris_9

(everyone) (everyone) (everyone) (everyone) (everyone) (everyone) (everyone)

Example 27

Accessing Diskettes on Other Systems

The following example shows how to automatically access myfiles from the remote system mars using AutoFS.
$ showmount -e mars $ cd /net/mars $ ls /floppy floppy0 myfiles

Example 28

Accessing PCMCIA Memory Cards on Remote Systems

The following example shows how to automatically access the PCMCIA memory card named myfiles from the remote system mars using AutoFS.

44

System Administration Guide: Devices and File Systems June 2005

$ showmount -e mars $ cd /net/mars $ ls /pcmem pcmem0 myfiles

Chapter 2 Accessing Removable Media (Tasks)

45

46

System Administration Guide: Devices and File Systems June 2005

CHAPTER

Formatting Removable Media (Tasks)

This chapter describes how to format removable media from the command line in the Solaris OS. For information on the procedures associated with formatting removable media, see Formatting Removable Media (Task Map) on page 47. For background information on removable media, see Chapter 1.

Formatting Removable Media (Task Map)

The following task map describes the tasks for formatting removable media.
Task Description For Instructions

1. Load unformatted media. 2. Format the media.

Insert the media into the drive and type the volcheck command. Format removable media.

How to Load Removable Media on page 49 How to Format Removable Media (rmformat) on page 51 How to Format Removable Media for a File System on page 52 How to Check a File System on Removable Media on page 54

3. (Optional) Add a UFS le system. 4. (Optional) Check the media.

Add a UFS le system to use the media for transferring les. Verify the integrity of the le system on the media.

47

Task

Description

For Instructions

5. (Optional) Repair bad blocks on the media.

Repair any bad blocks on the media, if How to Repair Bad Blocks necessary. on Removable Media on page 54 How to Enable or Disable Write Protection on Removable Media on page 55

6. (Optional) Apply read Apply read or write protection or or write and password password protection on the media, if protection. necessary.

Formatting Removable Media

The rmformat command is a utility that you can use to format and protect rewritable removable media. This utility does not require superuser privilege. The rmformat command has three formatting options:
I

quick This option formats removable media without certication or with limited certication of certain tracks on the media. long This option completely formats removable media. For some devices, the use of this option might include the certication of the whole media by the drive. force This option formats completely without user conrmation. For media with a password-protection mechanism, this option clears the password before formatting. This feature is useful when a password is forgotten. On media without password protection, this option forces a long format.

Formatting Removable Media Guidelines

Keep the following in mind when formatting removable media:
I

Close and quit the File Manager window. File Manager automatically displays a formatting window when you insert an unformatted media. To avoid the window, quit from File Manager. If you prefer to keep File Manager open, quit the formatting window when it appears.

Volume management (vold) mounts le systems automatically. So, you might have to unmount media before you can format it, if the media contains an existing le system.

Removable Media Hardware Considerations

Keep the following restrictions in mind when working with diskettes and PCMCIA memory cards:
48 System Administration Guide: Devices and File Systems June 2005

SPARC and x86 UFS formats are different. SPARC uses little-endian bit coding, x86 uses big-endian. Media formatted for UFS is restricted to the hardware platform on which they were formatted. So, a diskette formatted for UFS on a SPARC based platform cannot be used for UFS on an x86 platform. Likewise, a diskette formatted for UFS on an x86 platform cannot be used on a SPARC platform. The same restriction is applies to PCMCIA memory cards. A complete format for SunOS le systems consists of the basic bit formatting in addition the structure to support a SunOS le system. A complete format for a DOS le system consists of the basic bit formatting in addition the structure to support either an MS-DOS or an NEC-DOS le system. The procedures required to prepare a media for each type of le system are different. Therefore, before you format a diskette or PCMCIA memory card, consider which procedure to follow. For more information, see Formatting Removable Media (Task Map) on page 47.

Diskette Hardware Considerations

Keep the following in mind when formatting diskettes:
I I

For information on diskette names, see Table 21. Diskettes that are not named (that is, they have no label) are assigned the default name of noname.

A Solaris system can format the following diskette types:

I I I

UFS MS-DOS or NEC-DOS (PCFS) UDFS

On a Solaris system (either SPARC or x86), you can format diskettes with the following densities.

Diskette Size

Diskette Density

Capacity

3.5 3.5

High density (HD) Double density (DD)

1.44 Mbytes 720 Kbytes

By default, the diskette drive formats a diskette to a like density. This default means that a 1.44 Mbyte drive attempts to format a diskette for 1.44 Mbytes, regardless of whether the diskette is, in fact, a 1.44 Mbyte diskette, unless you instruct it otherwise. In other words, a diskette can be formatted to its capacity or lower, and a drive can format to its capacity or lower.

How to Load Removable Media

For information about removable media hardware considerations, see Removable Media Hardware Considerations on page 48.
Chapter 3 Formatting Removable Media (Tasks) 49

Steps

1. Insert the media. 2. Ensure that the media is formatted. If you arent sure, insert the media and check the status messages in the system console window, as described in Step 3. If you need to format the media, go to How to Format Removable Media (rmformat) on page 51. 3. Notify volume management.
$ volcheck -v media was found

Two status messages are possible: media was found Volume management detected the media and will attempt to mount it in the directory described in Table 21. If the media is formatted properly, no error messages appear in the console. If the media is not formatted, the media was found message is still displayed. However, error messages similar to the following appear in the system console window: fd0: unformatted diskette or no diskette in the drive fd0: read failed (40 1 0) fd0: bad format You must format the media before volume management can mount it. For more information, see Chapter 3. no media was found Volume management did not detect the media. Ensure that the media is inserted properly, and run volcheck again. If unsuccessful, check the media, which could be damaged. You can also try to mount the media manually.

4. Verify that the media was mounted by listing its contents. For example, do the following for a diskette:
$ ls /floppy floppy0 myfiles

50

System Administration Guide: Devices and File Systems June 2005

Tip floppy0 is a symbolic link to the actual name of the diskette, In this case, myfiles. If the diskette has no name but is formatted correctly, the system refers to it as unnamed_floppy.

If nothing appears under the /floppy directory, the diskette was either not mounted or is not formatted properly. To nd out, run the mount command and look for the line that begins with /floppy (usually at the end of the listing): /floppy/name on /vol/dev/diskette0/name If this line does not appear, the diskette was not mounted. Check the system console window for error messages.

How to Format Removable Media (rmformat)

You can use the rmformat command to format the media. By default, this command creates two partitions on the media: partition 0 and partition 2 (the whole media).

Steps

1. Verify that volume management is running. If so, you can use the shorter nickname for the device name.
$ ps -ef | grep vold root 212 1 0 Nov 03 ? 0:01 /usr/sbin/vold

For information on starting vold, see How to Stop and Start Volume Management (vold) on page 36. For information on identifying media device names, see Using Removable Media Names on page 32. 2. Format the removable media.
$ rmformat -F [ quick | long | force ] device-name

See Formatting Removable Media on page 48 for more information on rmformat formatting options. If the rmformat output indicates bad blocks, see How to Repair Bad Blocks on Removable Media on page 54. 3. (Optional) Label the removable media with an 8-character label.
$ rmformat -b label device-name

For information on creating a DOS label, see mkfs_pcfs(1M).

Example 31

Formatting Removable Media

This example shows how to format a diskette.

Chapter 3 Formatting Removable Media (Tasks)

51

$ rmformat -F quick /dev/rdiskette Formatting will erase all the data on disk. Do you want to continue? (y/n) y .........................................................................

This example shows how to format a Zip drive.

$ rmformat -F quick /vol/dev/aliases/zip0 Formatting will erase all the data on disk. Do you want to continue? (y/n) y .........................................................................

M
Steps

How to Format Removable Media for a File System

1. Format the media.
$ rmformat -F quick device-name

2. (Optional) Create an alternate Solaris partition table.

$ rmformat -s slice-le device-name

A sample slice le appears similar to the following:

slices: 0 = 0, 30MB, "wm", "home" : 1 = 30MB, 51MB : 2 = 0, 94MB, "wm", "backup" : 6 = 81MB, 13MB

3. Become superuser. 4. Determine the appropriate le system type and select one of the following:
I

Create a UFS le system.

# newfs device-name

Create a UDFS le system.

# mkfs -F udfs device-name

Example 32

Formatting a Diskette for a UFS File System

The following example shows how to format a diskette and create a UFS le system on the diskette.
$ rmformat -F quick /vol/dev/aliases/floppy0 Formatting will erase all the data on disk. Do you want to continue? (y/n) y $ su # /usr/sbin/newfs /vol/dev/aliases/floppy0 newfs: construct a new file system /dev/rdiskette: (y/n)? y

52

System Administration Guide: Devices and File Systems June 2005

/dev/rdiskette: 2880 sectors in 80 cylinders of 2 tracks, 18 sectors 1.4MB in 5 cyl groups (16 c/g, 0.28MB/g, 128 i/g) super-block backups (for fsck -F ufs -o b=#) at: 32, 640, 1184, 1792, 2336, #

Example 33

Formatting a PCMCIA Memory Card for a UFS File System

The following example shows how to format a PCMCIA memory card and create a UFS le system on the card.
$ rmformat -F quick /vol/dev/aliases/pcmem0 $ su # /usr/sbin/newfs -v /vol/dev/aliases/pcmem0 newfs: construct a new file system /vol/dev/aliases/pcmem0:(y/n)? y . . . #

Example 34

Formatting Removable Media for a PCFS File System

This example shows how to create a PCFS le system with an alternate fdisk partition.
$ rmformat -F quick /dev/rdsk/c0t4d0s2:c Formatting will erase all the data on disk. Do you want to continue? (y/n) y $ su # fdisk /dev/rdsk/c0t4d0s2:c # mkfs -F pcfs /dev/rdsk/c0t4d0s2:c Construct a new FAT file system on /dev/rdsk/c0t4d0s2:c: (y/n)? y #

This example shows how to create a PCFS le system without an fdisk partition.
$ rmformat -F quick /dev/rdiskette Formatting will erase all the data on disk. Do you want to continue? (y/n) y $ su # mkfs -F pcfs -o nofdisk,size=2 /dev/rdiskette Construct a new FAT file system on /dev/rdiskette: (y/n)? y #

Chapter 3 Formatting Removable Media (Tasks)

53

M
Steps

How to Check a File System on Removable Media

1. Become superuser. 2. Identify the le system type and select one of the following:
I

Check a UFS le system.

# fsck -F ufs device-name

Check a UDFS le system.

# fsck -F udfs device-name

Check a PCFS le system.

# fsck -F pcfs device-name

Example 35

Checking a PCFS File System on Removable Media

The following example shows how check the consistency of a PCFS le system on media.
# fsck -F pcfs /dev/rdsk/c0t4d0s2 ** /dev/rdsk/c0t4d0s2 ** Scanning file system meta-data ** Correcting any meta-data discrepancies 1457664 bytes. 0 bytes in bad sectors. 0 bytes in 0 directories. 0 bytes in 0 files. 1457664 bytes free. 512 bytes per allocation unit. 2847 total allocation units. 2847 available allocation units. #

How to Repair Bad Blocks on Removable Media

You can only use the rmformat command to verify, analyze, and repair bad sectors that are found during verication if the drive supports bad block management. Most diskettes and PCMCIA memory cards do not support bad block management. If the drive supports bad block management, a best effort is made to rectify the bad block. If the bad block cannot be rectied despite the best effort mechanism, a message indicates the failure to repair the media.

54

System Administration Guide: Devices and File Systems June 2005

Steps

1. Repair bad blocks on removable media.

$ rmformat -c block-numbers device-name

Supply the block number in decimal, octal, or hexadecimal format from a previous rmformat session. 2. Verify the media.
$ rmformat -V read device-name

Applying Read or Write Protection and Password Protection to Removable Media

You can apply read protection or write protection, and set a password, on Iomega media such as Zip drives and Jaz drives.

How to Enable or Disable Write Protection on Removable Media

1. Determine whether you want to enable or disable write protection and select one of the following:
I

Steps

Enable write protection.

$ rmformat -w enable device-name

Disable write protection.

$ rmformat -w disable device-name

2. Verify whether the medias write protection is enabled or disabled.

$ rmformat -p device-name

How to Enable or Disable Read or Write Protection and Set a Password on Iomega Media
You can apply a password with a maximum of 32 characters for Iomega media that support this feature. You cannot set read protection or write protection without a password on Iomega media. In this situation, you are prompted to provide a password.

Chapter 3 Formatting Removable Media (Tasks)

55

You receive a warning message if you attempt to apply a password on media that does not support this feature.
Steps

1. Determine whether you want to enable or disable read protection or write protection and set a password. Select one of the following:
I

Enable read protection or write protection.

$ rmformat -W enable device-name Please enter password (32 chars maximum): xxx Please reenter password: $ rmformat -R enable device-name Please enter password (32 chars maximum): xxx Please reenter password:

Disable read protection or write protection and remove the password.

$ rmformat -W disable device-name Please enter password (32 chars maximum): xxx $ rmformat -R disable device-name Please enter password (32 chars maximum): xxx

2. Verify whether the medias read protection or write protection is enabled or disabled.
$ rmformat -p device-name

Example 36

Enabling or Disabling Read or Write Protection and Password Protection

This example shows how to enable write protection and set a password on a Zip drive.
$ rmformat -W enable /vol/dev/aliases/zip0 Please enter password (32 chars maximum): xxx Please reenter password: xxx

This example shows how to disable write protection and remove the password on a Zip drive.
$ rmformat -W disable /vol/dev/aliases/zip0 Please enter password (32 chars maximum): xxx

This example shows how to enable read protection and set a password on a Zip drive.
rmformat -R enable /vol/dev/aliases/zip0 Please enter password (32 chars maximum): xxx Please reenter password: xxx

This example shows to disable read protection and remove the password on a Zip drive.
56 System Administration Guide: Devices and File Systems June 2005

$ rmformat -R disable /vol/dev/aliases/zip0 Please enter password (32 chars maximum): xxx

Chapter 3 Formatting Removable Media (Tasks)

57

58

System Administration Guide: Devices and File Systems June 2005

CHAPTER

Writing CDs and DVDs (Tasks)

This chapter provides step-by-step instructions for writing and copying data CDs and DVDs and audio CDs with the cdrw command.
I I I I I I I I I

How to Restrict User Access to Removable Media With RBAC on page 62 How to Identify a CD or DVD Writer on page 63 How to Check the CD or DVD Media on page 63 How to Create an ISO 9660 File System for a Data CD or DVD on page 64 How to Create a Multi-Session Data CD on page 65 How to Create an Audio CD on page 67 How to Extract an Audio Track on a CD on page 68 How to Copy a CD on page 69 How to Erase CD-RW Media on page 70

Working With Audio CDs and Data CDs and DVDs

For new information about DVD support, please see Whats New in Removable Media? on page 25. You can use the cdrw command to write le systems for CDs and DVDs in ISO 9660 format with Rock Ridge or Joliet extensions on CD-R,CD-RW, DVD-RW, or DVD+RW media devices. You can use the cdrw command to perform the following tasks:
I I I I

Create data CDs and DVDs. Create audio CDs. Extract audio data from an audio CD. Copy CDs and DVDs.
59

Erase CD-RW media.

The cdrw command is available starting in the following releases:

I I

Software Supplement for the Solaris 8 Operating Environment 1/01 CD Part of the Solaris release starting in the Solaris 9 release

For information on recommended CD-R or CD-RW devices, go to: http://www.sun.com/io_technologies/ihvindex.html

CD/DVD Media Commonly Used Terms

This section denes commonly used terms related to CD/DVD media.

Term

Description

CD-R CD-RW DVD-RW DVD+RW

CD read media that can be written once and after that, can only be read from. CD rewritable media that can be written to and erased. CD-RW media can only be read by CD-RW devices. Digital video disk (rewritable) drives can be read only by DVD-RW drives. Digital video disk (recordable/rewritable) drives can write both DVD-R discs, which can play back on most DVD players, and computer drives and DVD-RW rewritable disks. ISO, an acronym for Industry Standards Organization, is an organization that sets standards for computer storage formats. An ISO 9660 le system is a standard CD or DVD le system that enables you to read the same CD or DVD on any major computer platform. The standard, issued in 1988, was written by an industry group named High Sierra, named after the High Sierra Hotel in Nevada. Almost all computers with CD or DVD drives can read les from an ISO 9660 le system.

ISO 9660

Joliet extensions Rock Ridge extensions

Adds Windows le system information. Adds UNIX le system information. (Rock Ridge is named after the town in the movie Blazing Saddles.)
Note These extensions are not exclusive. You can specify

both mkisofs -R and -j options for compatibility with both systems. (See mkisofs(1M) for details.)

60

System Administration Guide: Devices and File Systems June 2005

Term

Description

MMC-compliant recorder

Acronym for Multi Media Command, which means these recorders comply with a common command set. Programs that can write to one MMC-compliant recorder should be able to write to all other recorders. Acronym for Compact Disc Digital Audio, which is an industry standard method for storing digital audio on compact discs. Also known by the term Red Book format. The official industry specication calls for one or more audio les sampled in 16-bit stereo sound at a sampling rate of 44.1 kilohertz (kHz).

Red Book CDDA

Commonly used terms when writing to CD media are:

Term

Description

blanking mkisofs session track

The process of erasing data from the CD-RW media. The command to create ISO le system on a CD. A complete track with lead-in and lead-out information. A complete data or audio unit.

Writing CD and DVD Data and Audio CDs

The process of writing to a CD or DVD cannot be interrupted and needs a constant stream of data. Consider using the cdrw -S option to simulate writing to the media to verify that the system can provide data at a sufficient rate for writing to the CD or DVD. Write errors can be caused by one of the following problems:
I

The media cannot handle the drive speed. For example, some media are only certied for 2x or 4x speeds. The system is running too many heavy processes that are starving the writing process. Network congestion is causing delays in reading the image, and the image is on a remote system. The source drive is slower than the destination drive.
Chapter 4 Writing CDs and DVDs (Tasks) 61

If any of these problems occur, you can lower the writing speed of the device by using the cdrw -p option. For example, the following command shows how to simulate writing at 4x speed:
$ cdrw -iS -p 4 image.iso

You can also use the cdrw -C option to use the stated media capacity for copying an 80-minute CD. Otherwise, the cdrw command uses a default value of 74 minutes for copying an audio CD. For more information, see cdrw(1).

Restricting User Access to Removable Media With RBAC

By default, all users can access removable media starting in the Solaris 9 release. However, you can restrict user access to removable media by setting up a role through role-based access control (RBAC). Access to removable media is restricted by assigning the role to a limited set of users. For a discussion of using roles, see Role-Based Access Control (Overview) in System Administration Guide: Security Services.

How to Restrict User Access to Removable Media With RBAC

1. Become superuser or assume an equivalent role. 2. Start the Solaris Management Console.
$ /usr/sadm/bin/smc &

Steps

For more information on starting the console, see Starting the Solaris Management Console in System Administration Guide: Basic Administration. 3. Set up a role that includes the Device Management rights. For more information, see Chapter 9, Using Role-Based Access Control (Tasks), in System Administration Guide: Security Services. 4. Add users who need to use the cdrw command to the newly created role. 5. Comment the following line in the /etc/security/policy.conf le:
AUTHS_GRANTED=solaris.device.cdrw

If you do not do this step, all users still have access to the cdrw command, not just the members of the device management role.
62 System Administration Guide: Devices and File Systems June 2005

After this le is modied, the device management role members are the only users who can use the cdrw command. Everyone else is denied access with the following message:
Authorization failed, Cannot access disks.

M
Steps

How to Identify a CD or DVD Writer

1. Identify the CD or DVD writers on the system. For example:
$ cdrw -l Looking for CD devices... Node | Connected Device | Device type ----------------------+--------------------------------+----------------cdrom0 | YAMAHA CRW8824S 1.0d | CD Reader/Writer

2. Identify a specic CD or DVD writer. For example:

$ cdrw -a lename.wav -d cdrom2

3. Identify whether the media is blank or whether a table of contents exists on the media. For example:
$ cdrw -M Device : YAMAHA CRW8824S Firmware : Rev. 1.00 (26/04/00) Media is blank %

How to Check the CD or DVD Media

The cdrw command works with or without vold running. However, you must have superuser or role access to stop and start the vold daemon.

Steps

1. Insert a CD or DVD into the drive. The CD or DVD can be any CD or DVD that the drive can read. 2. Check that the drive is connected properly by listing the drive.
$ cdrw -l Looking for CD devices... Node Connected Device

Device type 63

Chapter 4 Writing CDs and DVDs (Tasks)

----------------------+--------------------------------+----------------cdrom1 | YAMAHA CRW8824S 1.0d | CD Reader/Writer

3. (Optional) If you do not see the drive in the list, select one of the following so that the system recognizes the drive.
I

Perform a reconguration boot.

# touch /reconfigure # init 6

Add the drive without rebooting the system

# drvconfig # disks

Then restart vold.

# /etc/init.d/vold stop # /etc/init.d/vold start

Creating a Data CD or DVD

Prepare the data rst by using the mkisofs command to convert the le and le information into the High Sierra format used on CDs or DVDs.

How to Create an ISO 9660 File System for a Data CD or DVD

1. Insert a blank CD or DVD into the drive. 2. Create the ISO 9660 le system on the new CD or DVD.
$ mkisofs -r /pathname > cd-le-system

Steps

-r /pathname > cd-le-system

Creates Rock Ridge information and resets le ownerships to zero. Identies the path name used to create the ISO 9660 le system. Identies the name of the le system to be put on the CD or DVD.

3. Copy the le system onto the CD or DVD.

$ cdrw -i cd-le-system

The -i cd-le-system species the image le for creating a data CD or DVD.

64

System Administration Guide: Devices and File Systems June 2005

Example 41

Creating an ISO 9660 File System for a Data CD or DVD

The following example shows how to create an ISO 9660 le system for a data CD or DVD.
$ mkisofs -r /home/dubs/ufs_dir > ufs_cd Total extents actually written = 56 Total translation table size: 0 Total rockridge attributes bytes: 329 Total directory bytes: 0 Path table size(bytes): 10 Max brk space used 8000 56 extents written (0 Mb)

Then, copy the le system onto the CD or DVD.

$ cdrw -i ufs_cd Initializing device...done. Writing track 1...done. Finalizing (Can take several minutes)...done.

How to Create a Multi-Session Data CD

This procedure describes how to put more than one session on a CD. This procedure includes an example of copying the infoA and infoB directories onto the CD.

Steps

1. Create the le system for the rst CD session.

$ mkisofs -o infoA -r -V my_infoA /data/infoA Total translation table size: 0 Total rockridge attributes bytes: 24507 Total directory bytes: 34816 Path table size(bytes): 98 Max brk space used 2e000 8929 extents written (17 Mb)

-o infoA -r -V my_infoA /data/infoA

Identies the name of the ISO le system. Creates Rock Ridge information and resets le ownerships to zero. Identies a volume label to be used as the mount point by vold. Identies the ISO image directory to create.

2. Copy the ISO le system for the rst session onto the CD.
$ cdrw -iO infoA Initializing device...done. Writing track 1...done. done. Finalizing (Can take several minutes)...done. Chapter 4 Writing CDs and DVDs (Tasks) 65

-i infoA -O

Identies the name of the image le to write to the CD. Keeps the CD open for writing.

3. Re-insert the CD after it is ejected. 4. Identify the path name of the CD media to include in the next write session.
$ eject -n . . . cdrom0 -> /vol/dev/rdsk/c2t4d0/my_infoA

Note the /vol/dev/... path name. 5. Identify the next writeable address on the CD to write the next session.
% cdrw -M /cdrom Device : YAMAHA CRW8424S Firmware : Rev. 1.0d (06/10/99) Track No. |Type |Start address ----------+--------+------------1 |Audio |0 2 |Audio |33057 3 |Data |60887 4 |Data |68087 5 |Data |75287 Leadout |Data |84218 Last session start address: 75287 Next writable address: 91118

Note the address in the Next writable address output so that you can provide this address when you write the next session. 6. Create the next ISO le system for the next CD session, and write it onto the CD.
$ mkisofs -o infoB -r -C 0,91118 -M /vol/dev/rdsk/c2t4d0/my_infoA /data/infoB Total translation table size: 0 Total rockridge attributes bytes: 16602 Total directory bytes: 22528 Path table size(bytes): 86 Max brk space used 20000 97196 extents written (189 Mb)

-o infoB -r

Identies the name of the ISO le system. Creates Rock Ridge information and resets le ownerships to zero.

66

System Administration Guide: Devices and File Systems June 2005

-C 0,91118

Identies the starting address of the rst session and the next writable address. Species the path of the existing ISO image to be merged. Identies the ISO image directory to create.

-M /vol/dev/rdsk/c2t4d0/my_infoA /data/infoB

Creating an Audio CD
You can use the cdrw command to create audio CDs from individual audio tracks or from .au and .wav les. The supported audio formats are describes in the following table:

Format

Description

sun wav cda

Sun .au le with data in Red Book CDDA format RIFF (.wav) le with data in Red Book CDDA format .cda le with raw CD audio data, which is 16-bit PCM stereo at 44.1 kHz sample rate in little-endian byte order .aur les with raw CD data in big-endian byte order

aur

If no audio format is specied, the cdrw command tries to determine the audio le format based on the le extension. The case of the characters in the extension is ignored.

How to Create an Audio CD

This procedure describes how to copy audio les onto a CD. 1. Insert a blank CD into the CD-RW drive. 2. Change to the directory that contains the audio les.
$ cd /myaudiodir

Steps

3. Copy the audio les onto the CD.

$ cdrw -a track1.wav track2.wav track3.wav

The -a option creates an audio CD.

Chapter 4 Writing CDs and DVDs (Tasks) 67

Example 42

Creating an Audio CD
The following example shows how to create an audio CD.
$ cdrw -a bark.wav chirp.au meow.wav Initializing device...done. Writing track 1...done. done. Writing track 2...done. Writing track 3...done. done. Finalizing (Can take several minutes)...done.

The following example shows how to create a multisession audio CD. The CD is ejected after the rst session is written. You would need to re-insert the CD before the next writing session.
$ cdrw -aO groucho.wav chico.au harpo.wav Initializing device...done. Writing track 1...done. done. Writing track 2...done. Writing track 3...done. done. Finalizing (Can take several minutes)...done. <Re-insert CD> $ cdrw -a zeppo.au Initializing device...done. Writing track 1...done. done. Finalizing (Can take several minutes)...done.

How to Extract an Audio Track on a CD

Use the following procedure to extract an audio track from a CD and copy the audio track to a new CD. If you dont use the cdrw -T option to specify the audio le type, the cdrw command uses the lename extension to determine the audio le type. For example, the cdrw command detects that this le is a .wav le.
$ cdrw -x 1 testme.wav

Steps

1. Insert an audio CD into the CD-RW drive. 2. Extract an audio track.

$ cdrw -x -T audio-type 1 audio-le

-x T audio-type

Extracts audio data from an audio CD. Identies the type of audio le to be extracted. Supported audio types are sun, wav, cda, or aur.

68

System Administration Guide: Devices and File Systems June 2005

audio-le

Identies the audio track to be extracted.

3. Copy the track to a new CD.

$ cdrw -a audio-le

Example 43

Extracting and Creating Audio CDs

The following example shows how to extract the rst track from an audio CD and name the le song1.wav.
$ cdrw -x -T wav 1 song1.wav Extracting audio from track 1...done.

This example shows how to copy a track to an audio CD.

$ cdrw -a song1.wav Initializing device...done. Writing track 1...done. Finalizing (Can take several minutes)...done.

How to Copy a CD
This procedure describes how to extract all the tracks from an audio CD into a directory and then copy all of them onto a blank CD.
Note By default, the cdrw command copies the CD into the /tmp directory. The copying might require up to 700 Mbytes of free space. If there is insufficient space in the /tmp directory for copying the CD, use the -m option to specify an alternate directory.

Steps

1. Insert an audio CD into a CD-RW drive. 2. Create a directory for the audio les.
$ mkdir /music_dir

3. Extract the tracks from the audio CD.

$ cdrw -c -m music_dir

An Extracting audio ... message is display for each track. The CD is ejected when all the tracks are extracted. 4. Insert a blank CD and press Return. After the tracks are extracted, the audio CD is ejected. You are prompted to insert a blank CD.
Chapter 4 Writing CDs and DVDs (Tasks) 69

Example 44

Copying a CD
This example shows how to copy one CD to another CD. You must have two CD-RW devices to do this task.
$ cdrw -c -s cdrom0 -d cdrom1

How to Erase CD-RW Media

You have to erase existing CD-RW data before the CD can be rewritten.
G Erase the entire media or just the last session on the CD by selecting one of the

Step

following:
I

Erase the last session only.

$ cdrw -d cdrom0 -b session

Erasing just the last session with the -b session option is faster than erasing the entire media with the -b all option. You can use the -b session option even if you used the cdrw command to create a data or audio CD in just one session.
I

Erase the entire media.

$ cdrw -d cdrom0 -b all

70

System Administration Guide: Devices and File Systems June 2005

CHAPTER

Managing Devices (Tasks)

This chapter provides overview information and step-by-step instructions for managing peripheral devices, such as disks, CD-ROMs, and tape devices, in the Solaris release. This is a list of the overview information in this chapter.
I I I I I

Whats New in Device Management? on page 71 Where to Find Device Management Tasks on page 74 About Device Drivers on page 74 Automatic Conguration of Devices on page 75 Displaying Device Conguration Information on page 77

This is a list of the step-by-step instructions in this chapter.

I I I

How to Display System Conguration Information on page 77 How to Add a Device Driver on page 82 How to Add a Peripheral Device on page 81

For information about accessing peripheral devices, see Chapter 10. Device management in the Solaris release usually involves adding and removing peripheral devices from systems, possibly adding a third-party device driver to support a device, and displaying system conguration information.

Whats New in Device Management?

This section provides information about new device management features in the Solaris 10 release.

71

USB Device Enhancements

For information on new USB device enhancements, see Whats New in USB Devices? on page 113

1394 (FireWire) and Mass Storage Support on x86 Systems

In this Solaris release, the 1394 OpenHCI host controller driver has been updated to include support for x86 systems. Previously, 1394 (Firewire) support was only available on SPARC systems. IEEE 1394 is also known by the Apple Computer trademark name, Firewire. Sonys trademark name for 1394 is i.LINK. 1394 is an industry standard serial bus which supports data rates of 100 Mbit/sec, 200 Mbit/sec, or 400 Mbit/sec. It is well suited to handle data from consumer electronics devices, such as video cameras, due to its high bandwidth and isochronous (on time) capabilities. For more information, see hci1394(7D). In this Solaris release, the scsa1394 driver has been added to support 1394 mass storage devices that are compliant with the Serial Bus Protocol 2 (SBP-2) specication. It supports both bus-powered and self-powered 1394 mass storage devices. Previously, only 1394 video cameras were supported. 1394 mass storage devices are treated as removable media devices. A 1394 mass storage device can be formatted by using the rmformat command. Using a 1394 mass storage devices is no different than using a USB mass storage device. This means you can mount, eject, hot-remove, and hot-insert a 1394 mass storage device. For more information on using these devices, see scsa1394(7D) and Chapter 8.

Device File System (devfs)

The devfs le system manages devices in this Solaris release. Continue to access all devices through entries in the /dev directory, which are symbolic links to entries in the /devices directory. The content of the /devices directory is now controlled by the devfs le system. The entries in the /devices directory dynamically represent the current state of accessible devices on the system and require no administration. The devfs le system provides the following enhancements:
I

Operations in the /devices directory result in attaching device entries. Unused device entries are detached.

72

System Administration Guide: Devices and File Systems June 2005

Increases system boot performance because only device entries that are needed to boot the system are attached. New device entries are added as the devices are accessed.

For more information, see the devfs(7FS) man page.

Power Management of Fibre Channel Devices

Power management of Sun systems has been provided in many previous Solaris releases. For example, the internal drives on the following systems are power managed by default:
I I I

SunBlade 1000 or 2000 SunBlade 100 or 150 SunBlade 2500 or 1500

The default settings in the /etc/power.conf le ensure Energy Star compliance and fully support power management of these systems. The following adapters connect external Fibre Channel storage devices:
I I

Sun StorEdge PCI Dual Fibre Channel Host Adapter Sun StorEdge PCI Single Fibre Channel Network Adapter

If a combination of the above adapters and Sun systems are used to attach external Fibre Channel storage devices, the external storage devices will also be power managed by default. Under the following conditions, power management should be disabled:
I

If the system has Fibre Channel attached disks that are connected to a storage area network (SAN) If the system has Fibre Channel attached disks that are used in a multi-initiator conguration, such as with the SunCluster software If the system is using IP over a Fibre Channel interface (see fcip(7D))

Power management should not be enabled when more than one Solaris system might share the same devices, as in the above conditions. You can disable power management for the system by changing the autopm keyword in the /etc/power.conf le as follows:
autopm disable

Then, recongure power management by running the pmconfig command or by rebooting the system. For more information, see power.conf(4) and pmconfig(1M).
Chapter 5 Managing Devices (Tasks) 73

Where to Find Device Management Tasks

The following table describes where to nd step-by-step instructions for hot-plugging devices and adding serial devices, such as printers and modems, and peripheral devices, such as a disk, CD-ROM, or tape device.
TABLE 51

Where to Find Instructions for Adding a Device

For More Information

Device Management Task

Add a disk that is not hot-pluggable. Chapter 13 or Chapter 14 Hot-plug a SCSI or PCI device. SCSI Hot-Plugging With the cfgadm Command on page 90 or PCI Hot-Plugging With the cfgadm Command on page 100 Using USB Mass Storage Devices (Task Map) on page 126 How to Add a Peripheral Device on page 81 Chapter 8, Managing Terminals and Modems (Overview), in System Administration Guide: Advanced Administration Chapter 1, Managing Printing Services (Overview), in System Administration Guide: Advanced Administration Chapter 4, Controlling Access to Devices (Tasks), in System Administration Guide: Security Services

Hot-plug a USB device. Add a CD-ROM or tape device. Add a modem.

Add a printer. Secure a device.

About Device Drivers

A computer typically uses a wide range of peripheral devices and mass-storage devices. Your system, for example, probably has a disk drive, a keyboard and a mouse, and some kind of magnetic backup medium. Other commonly used devices include the following:
I I I I I

CD-ROM drives Printers and plotters Light pens Touch-sensitive screens Digitizers

74

System Administration Guide: Devices and File Systems June 2005

Tablet-and-stylus pairs

The Solaris software does not directly communicate with all these devices. Each type of device requires different data formats, protocols, and transmission rates. A device driver is a low-level program that allows the operating system to communicate with a specic piece of hardware. The driver serves as the operating systems interpreter for that piece of hardware.

Automatic Conguration of Devices

The kernel consists of a small generic core with a platform-specic component and a set of modules. The kernel is congured automatically in the Solaris release. A kernel module is a hardware or software component that is used to perform a specic task on the system. An example of a loadable kernel module is a device driver that is loaded when the device is accessed. The platform-independent kernel is /kernel/genunix. The platform-specic component is /platform/uname -m/kernel/unix. The kernel modules are described in the following table.
TABLE 52 Location

Description of Solaris Kernel Modules

Directory Contents

/platform/uname -m/kernel /kernel

Platform-specic kernel components Kernel components common to all platforms that are needed for booting the system Kernel components common to all platforms within a particular instruction set

/usr/kernel

The system determines what devices are attached to it at boot time. Then, the kernel congures itself dynamically, loading needed modules into memory. At this time, device drivers are loaded when devices, such as disk devices and tape devices, are accessed. This process is called autoconguration because all kernel modules are loaded automatically when they are needed. You can customize the way in which kernel modules are loaded by modifying the /etc/system le. For instructions on modifying this le, see system(4).

Chapter 5 Managing Devices (Tasks)

75

Features and Benets of Autoconguration

The benets of autoconguration are as follows:
I I

Main memory is used more efficiently because modules are loaded when needed. There is no need to recongure the kernel when new devices are added to the system. Drivers can be loaded and tested without having to rebuild the kernel and reboot the system.

Autoconguration is used when you add a new device (and driver) to the system. At this time, you might need to perform reconguration boot so that the system recognizes the new device unless the device is hot-pluggable. For information about hot-plugging devices, see Chapter 6.

What You Need for Unsupported Devices

Device drivers needed to support a wide range of standard devices are included in the Solaris release. These drivers can be found in the /kernel/drv and /platform/uname -m/kernel/drv directories. However, if you have purchased an unsupported device, the manufacturer should provide the software that is needed for the device to be properly installed, maintained, and administered. At a minimum, this software includes a device driver and its associated conguration (.conf) le. The .conf les reside in the drv directories. This software might also include custom maintenance and administrative utilities because the device might be incompatible with Solaris utilities. For more information about what you need for unsupported devices, contact your device manufacturer.

76

System Administration Guide: Devices and File Systems June 2005

Displaying Device Conguration Information

Three commands are used to display system and device conguration information.
Command Description Man Page

prtconf

Displays system conguration information, including the total amount of memory and the device conguration as described by the systems device hierarchy. The output displayed by this command depends upon the type of system. Displays device conguration information, including system hardware, pseudo devices, loadable modules, and selected kernel parameters. Displays system diagnostic messages as well as a list of devices attached to the system since the last reboot.

prtconf(1M)

sysdef

sysdef(1M)

dmesg

dmesg(1M)

For information on the device names that are used to identify devices on the system, see Device Naming Conventions on page 170.

driver not attached Message

The following driver-related message might be displayed by the prtconf and sysdef commands:
device, instance #number (driver not attached)

This message does not always mean that a driver is unavailable for this device. This message means that no driver is currently attached to the device instance because no device exists at this node or the device is not in use. Drivers are loaded automatically when the device is accessed. They are unloaded when the device is not in use.

How to Display System Conguration Information

Use the output of the prtconf and sysdef commands to identify which disk, tape, and CD-ROM devices are connected to the system. The output of these commands displays the driver not attached messages next to the device instances. Because these devices are always being monitored by some system process, the driver not attached message is usually a good indication that no device exists at that device instance.
Chapter 5 Managing Devices (Tasks) 77

Use the sysdef command to display system conguration information that include pseudo devices, loadable modules, and selected kernel parameters.
Step
G Display system and device conguration information.
I

Display all the devices connected to a system. For example, the following prtconf -v output on a SunBlade 1000 identies the disk devices connected to the system. The detailed disk information is described in the Device Minor Nodes section within the ssd/fp driver section.
$ /usr/sbin/prtconf -v | more . . . Device Minor Nodes: dev=(118,8) dev_path=/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w210000 2037bde864,0:a spectype=blk type=minor dev_link=/dev/dsk/c0t1d0s0 dev_path=/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w210000 2037bde864,0:a,raw spectype=chr type=minor dev_link=/dev/rdsk/c0t1d0s0 dev=(118,9) dev_path=/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w210000 2037bde864,0:b spectype=blk type=minor dev_link=/dev/dsk/c0t1d0s1 dev_path=/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w210000 2037bde864,0:b,raw . . .

Display information about one specic device connected to the system. For example, the following prtconf output on a SunBlade 1000 displays the ssd instance number for /dev/dsk/c0t1d0s0.
# prtconf -v /dev/dsk/c0t1d0s0 ssd, instance #1

Display only the devices that are attached to the system.

# prtconf | grep -v not

Display device usage information. For example, the following fuser command displays which processes are accessing the /dev/console device.
# fuser -d /dev/console /dev/console: 346o # 323o

78

System Administration Guide: Devices and File Systems June 2005

Example 51

Displaying System Conguration Information

The following prtconf output is displayed on a SPARC based system.
# prtconf System Configuration: Sun Microsystems Memory size: 512 Megabytes System Peripherals (Software Nodes): sun4u

SUNW,Sun-Blade-1000 scsi_vhci, instance #0 packages (driver not attached) SUNW,builtin-drivers (driver not attached) deblocker (driver not attached) disk-label (driver not attached) terminal-emulator (driver not attached) obp-tftp (driver not attached) dropins (driver not attached) kbd-translator (driver not attached) ufs-file-system (driver not attached) chosen (driver not attached) openprom (driver not attached) client-services (driver not attached) options, instance #0 aliases (driver not attached) memory (driver not attached) virtual-memory (driver not attached) SUNW,UltraSPARC-III, instance #0 memory-controller, instance #0 SUNW,UltraSPARC-III, instance #1 memory-controller, instance #1 pci, instance #0 ebus, instance #0 flashprom (driver not attached) bbc (driver not attached) ppm, instance #0 i2c, instance #0 dimm-fru, instance #0 dimm-fru, instance #1 dimm-fru, instance #2 dimm-fru, instance #3 nvram, instance #4 idprom (driver not attached) i2c, instance #1 cpu-fru, instance #5 temperature, instance #0 cpu-fru, instance #6 temperature, instance #1 fan-control, instance #0 motherboard-fru, instance #7 i2c-bridge (driver not attached) beep, instance #0 rtc, instance #0 gpio (driver not attached) pmc (driver not attached) Chapter 5 Managing Devices (Tasks) 79

floppy (driver not attached) parallel (driver not attached) serial, instance #0 network, instance #0 firewire, instance #0 usb, instance #0 scsi (driver not attached) disk (driver not attached) tape (driver not attached) scsi (driver not attached) disk (driver not attached) tape (driver not attached) pci, instance #1 SUNW,qlc, instance #0 fp (driver not attached) disk (driver not attached) fp, instance #1 ssd, instance #1 ssd, instance #0 (driver not attached) ssd, instance #2 (driver not attached) ssd, instance #3 (driver not attached) ssd, instance #4 (driver not attached) ssd, instance #5 (driver not attached) ssd, instance #6 (driver not attached) upa, instance #0 SUNW,ffb, instance #0 (driver not attached) ppm, instance #0 pseudo, instance #0

The following sysdef output is displayed from an x86 based system.

# sysdef * Hostid * 29f10b4d * * i86pc Configuration * * * Devices * +boot (driver not attached) memory (driver not attached) aliases (driver not attached) chosen (driver not attached) i86pc-memory (driver not attached) i86pc-mmu (driver not attached) openprom (driver not attached) options, instance #0 packages (driver not attached) delayed-writes (driver not attached) itu-props (driver not attached) isa, instance #0 motherboard (driver not attached) pnpADP,1542, instance #0 80 System Administration Guide: Devices and File Systems June 2005

asy, instance #0 asy, instance #1 lp, instance #0 (driver not attached) fdc, instance #0 fd, instance #0 fd, instance #1 (driver not attached) kd (driver not attached) kdmouse (driver not attached) . . .

Adding a Peripheral Device to a System

Adding a new peripheral device that is not-pluggable usually involves the following:
I I I

Shutting down the system Connecting the device to the system Rebooting the system

Use How to Add a Peripheral Device on page 81 to add the following devices that are not hot-pluggable to a system:
I I I I

CD-ROM Secondary disk drive Tape drive SBUS card

In some cases, you might have to add a third-party device driver to support the new device. For information on hot-plugging devices, see Chapter 6.

M
Steps

How to Add a Peripheral Device

1. Become superuser. 2. (Optional) If you need to add a device driver to support the device, complete the procedure How to Add a Device Driver on page 82. 3. Create the /reconfigure le.
# touch /reconfigure

The /reconfigure le causes the Solaris software to check for the presence of any newly installed devices the next time you turn on or boot your system.
Chapter 5 Managing Devices (Tasks) 81

4. Shut down the system.

# shutdown -i0 -g30 -y

-i0 -g30 -y

Brings the system to the 0 init state, which is the appropriate state for turning the system power off for adding and removing devices. Shuts the system down in 30 seconds. The default is 60 seconds. Continues the system shutdown without user intervention. Otherwise, you are prompted to continue the shutdown process.

5. Select one of the following to turn off power to the system after it is shut down:
I I

For SPARC platforms, it is safe to turn off power if the ok prompt is displayed. For x86 platforms, it is safe to turn off power if the type any key to continue prompt is displayed.

6. Turn off power to all peripheral devices. For the location of power switches on any peripheral devices, refer to the hardware installation guides that accompany your peripheral devices. 7. Install the peripheral device, making sure that the device you are adding has a different target number than the other devices on the system. Often, a small switch is located at the back of the disk for selecting the target number. Refer to the hardware installation guide that accompanies the peripheral device for information on installing and connecting the device. 8. Turn on the power to the system. The system boots to multiuser mode, and the login prompt is displayed. 9. Verify that the peripheral device has been added by attempting to access the device. For information on accessing the device, see Chapter 10.

How to Add a Device Driver

This procedure assumes that the device has already been added to the system. If not, see What You Need for Unsupported Devices on page 76.

Steps

1. Become superuser. 2. Place the tape, diskette, or CD-ROM into the drive.

82

System Administration Guide: Devices and File Systems June 2005

3. Install the driver.

# pkgadd [-d] device package-name

-d device package-name

Identies the device path name that contains the package. Identies the package name that contains the device driver.

4. Verify that the package has been added correctly.

# pkgchk package-name #

The system prompt returns with no response if the package is installed correctly.
Example 52

Adding a Device Driver

The following example shows how to install and verify a package called XYZdrv.
# pkgadd XYZdrv (licensing messages displayed) . . . Installing XYZ Company driver as <XYZdrv> . . . Installation of <XYZdrv> was successful. # pkgchk XYZdrv #

Chapter 5 Managing Devices (Tasks)

83

84

System Administration Guide: Devices and File Systems June 2005

CHAPTER

Dynamically Conguring Devices (Tasks)

This chapter provides instructions for dynamically conguring devices in the Solaris OS. You can add, remove, or replace devices in the Solaris OS while the system is still running, if the system components support hot-plugging. If the system components do not support hot-plugging, you can reboot the system to recongure the devices. For information on the procedures associated with dynamically conguring devices, see the following:
I I I I

SCSI Hot-Plugging With the cfgadm Command (Task Map) on page 89 PCI Hot-Plugging With the cfgadm Command (Task Map) on page 99 Application Developer RCM Script (Task Map) on page 106 System Administrator RCM Script (Task Map) on page 106

For information on hot-plugging USB devices with the cfgadm command, see Hot-Plugging USB Devices With the cfgadm Command on page 146. For information about accessing devices, see Chapter 10.

Dynamic Reconguration and Hot-Plugging

Hot-plugging is the ability to physically add, remove, or replace system components while the system is running. Dynamic reconguration refers to the ability to hot-plug system components. This term also refers to the general ability to move system resources (both hardware and software) around in the system or to disable them in some way without physically removing them from the system. Generally, you can hot-plug the following bus types:
I

USB
85

I I I I

Fibre Channel 1394 ATA SCSI

In addition, you can hot-plug the following devices with the cfgadm command:
I I I

USB devices on SPARC and x86 platforms SCSI devices on SPARC and x86 platforms PCI devices on SPARC and x86 platforms

Features of the cfgadm command include the following:

I I I I

Displaying system component status Testing system components Changing component congurations Displaying conguration help messages

The benet of using the cfgadm command to recongure systems components is that you can add, remove, or replace components while the system is running. An added benet is that the cfgadm command guides you through the steps needed to add, remove, or replace system components. For step-by-step instructions on hot-plugging components, see the following:
I I I

SCSI Hot-Plugging With the cfgadm Command on page 90 PCI Hot-Plugging With the cfgadm Command on page 100 cfgadm(1M)

Note Not all SCSI and PCI controllers support hot-plugging with the cfgadm command.

As part of Suns high availability strategy, dynamic reconguration is expected to be used in conjunction with additional layered products, such as alternate pathing or fail over software. Both products provide fault tolerance in the event of a device failure. Without any high availability software, you can replace a failed device by manually stopping the appropriate applications, unmounting noncritical le systems, and then proceeding with the add or remove operations.
Note Some systems have slots that hot-pluggable and slots that are not hot-pluggable. For information about hot-plugging devices on your specic hardware conguration, such as on enterprise-level systems, refer to your hardware conguration documentation.

86

System Administration Guide: Devices and File Systems June 2005

Attachment Points
The cfgadm command displays information about attachment points, which are locations in the system where dynamic reconguration operations can occur. An attachment point consists of the following:
I

An occupant, which represents a hardware component that can be congured into the system A receptacle, which is the location that accepts the occupant

Attachment points are represented by logical and physical attachment point IDs (Ap_Ids). The physical Ap_Id is the physical path name of the attachment point. The logical Ap_Id is a user-friendly alternative for the physical Ap_Id. For more information on Ap_Ids, refer to cfgadm(1M). The logical Ap_Id for a SCSI Host Bus Adapter (HBA), or SCSI controller, is usually represented by the controller number, such as c0. In cases where no controller number has been assigned to a SCSI HBA, then an internally generated unique identier is provided. An example of a unique identier for a SCSI controller is the following: fas1:scsi The logical Ap_Id for a SCSI device usually has this format: HBA-logical-apid::device-identier In the following example, c0 is the logical Ap_Id for the SCSI HBA: c0::dsk/c0t3d0 The device identier is typically derived from the logical device name for the device in the /dev directory. For example, a tape device with logical device name, /dev/rmt/1, has the following logical Ap_Id: c0::rmt/1 If a logical Ap_Id of a SCSI device cannot be derived from the logical name in the /dev directory, then an internally generated unique identier is provided. An example of an identier for the /dev/rmt/1 tape device is the following: c0::st4 For more information on SCSI Ap_Ids, refer to cfgadm_scsi(1M). The cfgadm command represents all resources and dynamic reconguration operations in terms of a common set of states (such as congured and uncongured) and operations (such as connect, congure, uncongure, and so on). For more information on these common states and operations, see cfgadm(1M).
Chapter 6 Dynamically Conguring Devices (Tasks) 87

The following table shows the receptacle and occupant states for the SCSI HBA attachment points.

Receptacle State

Description

Occupant State

Description

empty

N/A for SCSI HBA

configured

One or more devices is congured on the bus No devices are congured

disconnected

Bus quiesced Bus active

unconfigured

connected

The following table shows the receptacle and occupant states for SCSI device attachment points.

Receptacle State

Description

Occupant State

Description

empty disconnected

N/A for SCSI devices Bus quiesced Bus active

configured unconfigured

Device is congured Device is not congured

connected

The state of SCSI attachment points is unknown unless special hardware indicates otherwise. For instructions on displaying SCSI component information, see How to Display Information About SCSI Devices on page 90.

x86: Detaching PCI Adapter Cards

A PCI adapter card that is hosting nonvital system resources can be removed if the device driver supports hot-plugging. A PCI adapter card is not detachable if it is a vital system resource. For a PCI adapter card to be detachable, the following conditions must be met:
I I

The device driver must support hot-plugging. Critical resources must be accessible through an alternate pathway.

For example, if a system has only one Ethernet card installed in it, the Ethernet card cannot be detached without losing the network connection. This detachment requires additional layered software support to keep the network connection active.

88

System Administration Guide: Devices and File Systems June 2005

x86: Attaching PCI Adapter Cards

A PCI adapter card can be added to the system as long as the following conditions are met:
I I

There are slots available. The device driver supports hot-plugging for this adapter card.

For step-by-step instructions on adding or removing a PCI adapter card, see PCI Hot-Plugging With the cfgadm Command on page 100.

SCSI Hot-Plugging With the cfgadm Command (Task Map)

Task Description For Instructions

Display information about SCSI devices.

Display information about SCSI controllers and devices.

How to Display Information About SCSI Devices on page 90

Uncongure a SCSI controller. Uncongure a SCSI controller. How to Uncongure a SCSI Controller on page 91 Congure a SCSI controller. Congure a SCSI controller that was previously uncongured. Congure a specic SCSI device. Disconnect a specic SCSI controller. Connect a specic SCSI controller that was previously disconnected. Add a specic SCSI device to a SCSI bus. How to Congure a SCSI Controller on page 91 How to Congure a SCSI Device on page 92 How to Disconnect a SCSI Controller on page 93 SPARC: How to Connect a SCSI Controller on page 94 SPARC: How to Add a SCSI Device to a SCSI Bus on page 94

Congure a SCSI device. Disconnect a SCSI controller. Connect a SCSI controller.

Add a SCSI device to a SCSI bus.

Replace an identical device on Replace a device on the SCSI SPARC: How to Replace an a SCSI controller. bus with another device of the Identical Device on a SCSI Controller on page 95 same type.

Chapter 6 Dynamically Conguring Devices (Tasks)

89

Task

Description

For Instructions

Remove a SCSI device. Troubleshoot SCSI conguration problems.

Remove a SCSI device from the system. Resolve a failed SCSI uncongure operation.

SPARC: How to Remove a SCSI Device on page 96 How to Resolve a Failed SCSI Uncongure Operation on page 99

SCSI Hot-Plugging With the cfgadm Command

This section describes various SCSI hot-plugging procedures that you can perform with the cfgadm command.
Note The SCSI framework generally supports hot-plugging of SCSI devices. However, you should consult your hardware documentation to conrm whether hot-plugging is supported for your SCSI devices.

These procedures use specic devices as examples to illustrate how to use the cfgadm command to hot-plug SCSI components. The device information that you supply, and that the cfgadm command displays, depends on your system conguration.

How to Display Information About SCSI Devices

The following procedure uses SCSI controllers c0 and c1 and the devices that are attached to them in the examples of the type of device conguration information that you can display with the cfgadm command.
Note If the SCSI device is not supported by the cfgadm command, the device does

not display in the cfgadm command output.

Steps

1. Become superuser. 2. Display information about attachment points on the system.

# cfgadm -l Ap_Id c0 Type scsi-bus Receptacle connected Occupant configured Condition unknown

90

System Administration Guide: Devices and File Systems June 2005

c1

scsi-bus

connected

configured

unknown

In this example, c0 and c1 represent two SCSI controllers. 3. Display information about a systems SCSI controllers and their attached devices.
# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0 Type scsi-bus disk tape scsi-bus disk unavailable Receptacle connected connected connected connected connected connected Occupant configured configured configured configured configured unconfigured Condition unknown unknown unknown unknown unknown unknown

Note The cfgadm -l commands displays information about SCSI HBAs but not

SCSI devices. Use the cfgadm -al command to display information about SCSI devices such as disk and tapes.

How to Uncongure a SCSI Controller

The following procedure uses SCSI controller c1 in the example of unconguring a SCSI controller.

Steps

1. Become superuser. 2. Uncongure a SCSI controller.

# cfgadm -c unconfigure c1

3. Verify that the SCSI controller is uncongured.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 Type scsi-bus disk tape scsi-bus Receptacle connected connected connected connected Occupant configured configured configured unconfigured Condition unknown unknown unknown unknown

Notice that the Occupant column for c1 species unconfigured, indicating that the SCSI bus has no congured occupants. If the uncongure operation fails, see How to Resolve a Failed SCSI Uncongure Operation on page 99.

How to Congure a SCSI Controller

The following procedure uses SCSI controller c1 in the example of conguring a SCSI controller.
Chapter 6 Dynamically Conguring Devices (Tasks) 91

Steps

1. Become superuser. 2. Congure a SCSI controller.

# cfgadm -c configure c1

3. Verify that the SCSI controller is congured.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0 Type scsi-bus disk tape scsi-bus disk unavailable Receptacle connected connected connected connected connected connected Occupant configured configured configured configured configured unconfigured Condition unknown unknown unknown unknown unknown unknown

The previous uncongure procedure removed all devices on the SCSI bus. Now all the devices are congured back into the system.

How to Congure a SCSI Device

The following procedure uses SCSI disk c1t4d0 in the example of conguring a SCSI device.

Steps

1. Become superuser. 2. Identify the device to be congured.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0 Type scsi-bus disk tape scsi-bus disk unavailable Receptacle connected connected connected connected connected connected Occupant configured configured configured configured configured unconfigured Condition unknown unknown unknown unknown unknown unknown

3. Congure the SCSI device.

# cfgadm -c configure c1::dsk/c1t4d0

4. Verify that the SCSI device is congured.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0 92 Type scsi-bus disk tape scsi-bus disk disk Receptacle connected connected connected connected connected connected Occupant configured configured configured configured configured configured Condition unknown unknown unknown unknown unknown unknown

System Administration Guide: Devices and File Systems June 2005

How to Disconnect a SCSI Controller

Caution Disconnecting a SCSI device must be done with caution, particularly when you are dealing with controllers for disks that contain critical le systems such as root (/), usr, var, and the swap partition. The dynamic reconguration software cannot detect all cases where a system hang might result. Use this procedure with caution.

The following procedure uses SCSI controller c1 in the example of disconnecting a SCSI device.
Steps

1. Become superuser. 2. Verify that the device is connected before you disconnect it.
# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0 Type scsi-bus disk tape scsi-bus disk disk Receptacle connected connected connected connected connected connected Occupant configured configured configured configured configured configured Condition unknown unknown unknown unknown unknown unknown

3. Disconnect the SCSI controller.

# cfgadm -c disconnect c1 WARNING: Disconnecting critical partitions may cause system hang. Continue (yes/no)? y

Caution This command suspends all I/O activity on the SCSI bus until the cfgadm -c connect command is used. The cfgadm command does some basic checking to prevent critical partitions from being disconnected, but it cannot detect all cases. Inappropriate use of this command can result in a system hang and could require a system reboot.

4. Verify that the SCSI bus is disconnected.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t10d0 c1::dsk/c1t4d0 Type scsi-bus disk tape unavailable unavailable unavailable Receptacle connected connected connected disconnected disconnected disconnected Occupant configured configured configured configured configured configured Condition unknown unknown unknown unknown unknown unknown

The controller and all the devices that are attached to it are disconnected from the system.
Chapter 6 Dynamically Conguring Devices (Tasks) 93

SPARC: How to Connect a SCSI Controller

The following procedure uses SCSI controller c1 in the example of connecting a SCSI controller.

Steps

1. Become superuser. 2. Verify that the device is disconnected before you connect it.
# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t10d0 c1::dsk/c1t4d0 Type scsi-bus disk tape unavailable unavailable unavailable Receptacle connected connected connected disconnected disconnected disconnected Occupant configured configured configured configured configured configured Condition unknown unknown unknown unknown unknown unknown

3. Connect the SCSI controller.

# cfgadm -c connect c1

4. Verify that the SCSI controller is connected.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0 Type scsi-bus disk tape scsi-bus disk disk Receptacle connected connected connected connected connected connected Occupant configured configured configured configured configured configured Condition unknown unknown unknown unknown unknown unknown

SPARC: How to Add a SCSI Device to a SCSI Bus

SCSI controller c1 is used in the example of how to add a SCSI device to a SCSI bus.
Note When you add devices, you specify the Ap_Id of the SCSI HBA (controller) to

which the device is attached, not the Ap_Id of the device itself.

Steps

1. Become superuser. 2. Identify the current SCSI conguration.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 Type scsi-bus disk Receptacle connected connected Occupant configured configured Condition unknown unknown

94

System Administration Guide: Devices and File Systems June 2005

c0::rmt/0 c1 c1::dsk/c1t3d0

tape scsi-bus disk

connected connected connected

configured configured configured

unknown unknown unknown

3. Add the SCSI device to the SCSI bus. a. Type the following cfgadm command. For example:
# cfgadm -x insert_device c1 Adding device to SCSI HBA: /devices/sbus@1f,0/SUNW,fas@1,8800000 This operation will suspend activity on SCSI bus: c1

b. Type y at the Continue (yes/no)? prompt to proceed.

Continue (yes/no)? y SCSI bus quiesced successfully. It is now safe to proceed with hotplug operation.

I/O activity on the SCSI bus is suspended while the hot-plug operation is in progress. c. Connect the device and then power it on. d. Type y at the Enter y if operation is complete or n to abort (yes/no)? prompt.
Enter y if operation is complete or n to abort (yes/no)? y

4. Verify that the device has been added.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0 Type scsi-bus disk tape scsi-bus disk disk Receptacle connected connected connected connected connected connected Occupant configured configured configured configured configured configured Condition unknown unknown unknown unknown unknown unknown

A new disk has been added to controller c1.

SPARC: How to Replace an Identical Device on a SCSI Controller

The following procedure uses SCSI disk c1t4d0 in the example of replacing an identical device on a SCSI controller.

Steps

1. Become superuser. 2. Identify the current SCSI conguration.

# cfgadm -al Ap_Id Type Receptacle Occupant Condition 95

Chapter 6 Dynamically Conguring Devices (Tasks)

c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0

scsi-bus disk tape scsi-bus disk disk

connected connected connected connected connected connected

configured configured configured configured configured configured

unknown unknown unknown unknown unknown unknown

3. Replace a device on the SCSI bus with another device of the same type. a. Type the following cfgadm command. For example:
# cfgadm -x replace_device c1::dsk/c1t4d0 Replacing SCSI device: /devices/sbus@1f,0/SUNW,fas@1,8800000/sd@4,0 This operation will suspend activity on SCSI bus: c1

b. Type y at the Continue (yes/no)? prompt to proceed. I/O activity on the SCSI bus is suspended while the hot-plug operation is in progress.
Continue (yes/no)? y SCSI bus quiesced successfully. It is now safe to proceed with hotplug operation.

c. Power off the device to be removed and remove it. d. Add the replacement device. Then, power it on. The replacement device should be of the same type and at the same address (target and lun) as the device to be removed. e. Type y at the Enter y if operation is complete or n to abort (yes/no)? prompt.
Enter y if operation is complete or n to abort (yes/no)? y

4. Verify that the device has been replaced.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0 Type scsi-bus disk tape scsi-bus disk disk Receptacle connected connected connected connected connected connected Occupant configured configured configured configured configured configured Condition unknown unknown unknown unknown unknown unknown

SPARC: How to Remove a SCSI Device

The following procedure uses SCSI disk c1t4d0 in the example of removing a device on a SCSI controller.

Steps
96

1. Become superuser.

System Administration Guide: Devices and File Systems June 2005

2. Identify the current SCSI conguration.

# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 c1::dsk/c1t4d0 Type scsi-bus disk tape scsi-bus disk disk Receptacle connected connected connected connected connected connected Occupant configured configured configured configured configured configured Condition unknown unknown unknown unknown unknown unknown

3. Remove the SCSI device from the system. a. Type the following cfgadm command. For example:
# cfgadm -x remove_device c1::dsk/c1t4d0 Removing SCSI device: /devices/sbus@1f,0/SUNW,fas@1,8800000/sd@4,0 This operation will suspend activity on SCSI bus: c1

b. Type y at the Continue (yes/no)? prompt to proceed.

Continue (yes/no)? y SCSI bus quiesced successfully. It is now safe to proceed with hotplug operation.

I/O activity on the SCSI bus is suspended while the hot-plug operation is in progress. c. Power off the device to be removed and remove it. d. Type y at the Enter y if operation is complete or n to abort (yes/no)? prompt.
Enter y if operation is complete or n to abort (yes/no)? y

4. Verify that the device has been removed from the system.
# cfgadm -al Ap_Id c0 c0::dsk/c0t0d0 c0::rmt/0 c1 c1::dsk/c1t3d0 Type scsi-bus disk tape scsi-bus disk Receptacle connected connected connected connected connected Occupant configured configured configured configured configured Condition unknown unknown unknown unknown unknown

Troubleshooting SCSI Conguration Problems

This section provides error messages and possible solutions for troubleshooting SCSI conguration problems. For more information on troubleshooting SCSI conguration problems, see cfgadm(1M).

Chapter 6 Dynamically Conguring Devices (Tasks)

97

Error Message
cfgadm: Component system is busy, try again: failed to offline: device-path Resource Information ------------------ -------------------------/dev/dsk/c1t0d0s0 mounted filesystem "/le-system"

Cause You attempted to remove or replace a device with a mounted le system. Solution Unmount the le system that is listed in the error message and retry the cfgadm operation. Error Message
cfgadm: Component system is busy, try again: failed to offline: device-path Resource Information ------------------ -------------------------/dev/dsk/device-name swap area

Cause If you use the cfgadm command to remove a system resource, such as a swap device or a dedicated dump device, a similar error message is displayed if the system resource is still active. Solution Uncongure the swap areas on the device that is specied and retry the cfgadm operation. Error Message
cfgadm: Component system is busy, try again: failed to offline: device-path Resource Information ------------------ -------------------------/dev/dsk/device-name dump device (swap)

Cause You attempted to remove or replace a dump device that is congured on a swap area. Solution Uncongure the dump device that is congured on the swap area and retry the cfgadm operation. Error Message
cfgadm: Component system is busy, try again: failed to offline: device-path Resource Information ------------------ -------------------------/dev/dsk/device-name dump device (dedicated)

Cause You attempted to remove or replace a dedicated dump device.

98 System Administration Guide: Devices and File Systems June 2005

Solution Uncongure the dedicate dump device and retry the cfgadm operation.

How to Resolve a Failed SCSI Uncongure Operation

Use this procedure if one or more target devices are busy and the SCSI uncongure operation fails. Otherwise, future dynamic reconguration operations on this controller and target devices will fail with a dr in progress message.

Steps

1. Become superuser. 2. Recongure the controller.

# cfgadm -c configure device-name

PCI Hot-Plugging With the cfgadm Command (Task Map)

Task Description For Instructions

Display PCI slot conguration Display the status of PCI information. hot-pluggable devices and slots on the system. Remove a PCI adapter card. Uncongure the card, disconnect power from the slot, and remove the card from the system. Insert the adapter card into a hot-pluggable slot, connect power to the slot, and congure the card. Identify error message and possible solutions to resolve PCI conguration problems.

How to Display PCI Slot Conguration Information on page 100 How to Remove a PCI Adapter Card on page 101

Add a PCI adapter card.

How to Add a PCI Adapter Card on page 102

Troubleshoot PCI conguration problems.

Troubleshooting PCI Conguration Problems on page 103

Chapter 6 Dynamically Conguring Devices (Tasks)

99

PCI Hot-Plugging With the cfgadm Command

This section provides step-by-step instructions for hot-plugging PCI adapter cards on x86 based systems. In the examples, only PCI attachment points are listed, for brevity. The attachment points that are displayed on your system depend on your system conguration.

How to Display PCI Slot Conguration Information

The cfgadm command displays the status of PCI hot-pluggable devices and slots on a system. For more information, see cfgadm(1M).

Steps

1. Become superuser. 2. Display PCI conguration information.

I

Display PCI slot conguration information. For example:

# cfgadm Ap_Id pci1:hpc0_slot0 pci1:hpc0_slot1 pci1:hpc0_slot2 pci1:hpc0_slot3 pci1:hpc0_slot4 Type unknown unknown unknown ethernet/hp unknown Receptacle empty empty empty connected empty Occupant unconfigured unconfigured unconfigured configured unconfigured Condition unknown unknown unknown ok unknown

Display specic PCI device information. For example:

# cfgadm -s "cols=ap_id:type:info" pci Ap_Id Type Information pci1:hpc0_slot0 unknown Slot 7 pci1:hpc0_slot1 unknown Slot 8 pci1:hpc0_slot2 unknown Slot 9 pci1:hpc0_slot3 ethernet/hp Slot 10 pci1:hpc0_slot4 unknown Slot 11

The logical Ap_Id, pci1:hpc0_slot0, is the logical Ap_Id for hot-pluggable slot, Slot 7. The component hpc0 indicates the hot-pluggable adapter card for this slot, and pci1 indicates the PCI bus instance. The Type eld indicates the type of PCI adapter card that is present in the slot.
100 System Administration Guide: Devices and File Systems June 2005

M
Steps

How to Remove a PCI Adapter Card

1. Become superuser. 2. Determine which slot the PCI adapter card is in.
# cfgadm Ap_Id pci1:hpc0_slot0 pci1:hpc0_slot1 pci1:hpc0_slot2 pci1:hpc0_slot3 pci1:hpc0_slot4 Type unknown unknown unknown ethernet/hp unknown Receptacle empty empty empty connected empty Occupant unconfigured unconfigured unconfigured configured unconfigured Condition unknown unknown unknown ok unknown

3. Stop the application that has the device open. For example, if the device is an Ethernet card, use the ifconfig command to bring down the interface and unplumb the interface. 4. Uncongure the device.
# cfgadm -c unconfigure pci1:hpc0_slot3

5. Conrm that the device has been uncongured.

# cfgadm Ap_Id pci1:hpc0_slot0 pci1:hpc0_slot1 pci1:hpc0_slot2 pci1:hpc0_slot3 pci1:hpc0_slot4 Type unknown unknown unknown ethernet/hp unknown Receptacle empty empty empty connected empty Occupant unconfigured unconfigured unconfigured unconfigured unconfigured Condition unknown unknown unknown unknown unknown

6. Disconnect the power to the slot.

# cfgadm -c disconnect pci1:hpc0_slot3

7. Conrm that the device has been disconnected.

# cfgadm Ap_Id pci1:hpc0_slot0 pci1:hpc0_slot1 pci1:hpc0_slot2 pci1:hpc0_slot3 pci1:hpc0_slot4 Type unknown unknown unknown ethernet/hp unknown Receptacle empty empty empty disconnected empty Occupant unconfigured unconfigured unconfigured unconfigured unconfigured Condition unknown unknown unknown unknown unknown

8. Open the slot latches and remove the PCI adapter card.

Chapter 6 Dynamically Conguring Devices (Tasks)

101

M
Steps

How to Add a PCI Adapter Card

1. Become superuser. 2. Identify the hot-pluggable slot and open latches. 3. Insert the PCI adapter card into a hot-pluggable slot. 4. Determine which slot the PCI adapter card is in once it is inserted. Close the latches.
# cfgadm Ap_Id pci1:hpc0_slot0 pci1:hpc0_slot1 pci1:hpc0_slot2 pci1:hpc0_slot3 pci1:hpc0_slot4 Type unknown unknown unknown ethernet/hp unknown Receptacle empty empty empty disconnected empty Occupant unconfigured unconfigured unconfigured unconfigured unconfigured Condition unknown unknown unknown unknown unknown

5. Connect the power to the slot.

# cfgadm -c connect pci1:hpc0_slot3

6. Conrm that the slot is connected.

# cfgadm Ap_Id pci1:hpc0_slot0 pci1:hpc0_slot1 pci1:hpc0_slot2 pci1:hpc0_slot3 pci1:hpc0_slot4 Type unknown unknown unknown ethernet/hp unknown Receptacle empty empty empty connected empty Occupant unconfigured unconfigured unconfigured unconfigured unconfigured Condition unknown unknown unknown unknown unknown

7. Congure the PCI adapter card.

# cfgadm -c configure pci1:hpc0_slot3

8. Verify the conguration of the PCI adapter card in the slot.

# cfgadm Ap_Id pci1:hpc0_slot0 pci1:hpc0_slot1 pci1:hpc0_slot2 pci1:hpc0_slot3 pci1:hpc0_slot4 Type unknown unknown unknown ethernet/hp unknown Receptacle empty empty empty connected empty Occupant unconfigured unconfigured unconfigured configured unconfigured Condition unknown unknown unknown unknown unknown

9. Congure any supporting software if this device is a new device. For example, if this device is an Ethernet card, use the ifconfig command to set up the interface.

102

System Administration Guide: Devices and File Systems June 2005

Troubleshooting PCI Conguration Problems

Error Message
cfgadm: Configuration operation invalid: invalid transition

Cause An invalid transition was attempted. Solution Check whether the cfgadm -c command was issued appropriately. Use the cfgadm command to check the current receptacle and occupant state and to make sure that the Ap_Id is correct. Error Message
cfgadm: Attachment point not found

Cause The specied attachment point was not found. Solution Check whether the attachment point is correct. Use the cfgadm command to display a list of available attachment points. Also check the physical path to see if the attachment point is still there.
Note In addition to the cfgadm command, the prtconf command is helpful during hot-pluggable operations. The prtconf command displays whether the Solaris software recognizes the hardware. After adding hardware, use the prtconf command to verify that the hardware is recognized. After a congure operation, use the prtconf -D command to verify that the driver is attached to the newly installed hardware device.

Reconguration Coordination Manager (RCM) Script Overview

The Reconguration Coordination Manager (RCM) is the framework that manages the dynamic removal of system components. By using RCM, you can register and release system resources in an orderly manner. You can use the new RCM script feature to write your own scripts to shut down your applications, or to cleanly release the devices from your applications during dynamic reconguration. The RCM framework launches a script automatically in response to a reconguration request, if the request impacts the resources that are registered by the script.
Chapter 6 Dynamically Conguring Devices (Tasks) 103

You can also release resources from applications manually before you dynamically remove the resource. Or, you can use the cfgadm command with the -f option to force a reconguration operation. However, this option might leave your applications in an unknown state. Also, the manual release of resources from applications commonly causes errors. The RCM script feature simplies and better controls the dynamic reconguration process. By creating an RCM script, you can do the following:
I

Automatically release a device when you dynamically remove a device. This process also closes the device if the device is opened by an application. Run site-specic tasks when you dynamically remove a device from the system.

What Is an RCM Script?

I

An executable shell script (Perl, sh, csh, or ksh) or binary program that the RCM daemon runs. Perl is the recommended language. A script that runs in its own address space by using the user ID of the script le owner. A script that is run by the RCM daemon when you use the cfgadm command to dynamically recongure a system resource.

What Can an RCM Script Do?

You can use an RCM script to release a device from an application when you dynamically remove a device. If the device is currently open, the RCM script also closes the device. For example, an RCM script for a tape backup application can inform the tape backup application to close the tape drive or shut down the tape backup application.

How Does the RCM Script Process Work?

You can invoke an RCM script as follows:
$ script-name command [args ...]

An RCM script performs the following basic steps: 1. 2. 3. 4.

104

Takes the RCM command from command-line arguments. Executes the command. Writes the results to stdout as name-value pairs. Exits with the appropriate exit status.

System Administration Guide: Devices and File Systems June 2005

The RCM daemon runs one instance of a script at a time. For example, if a script is running, the RCM daemon does not run the same script until the rst script exits.

RCM Script Commands

You must include the following RCM commands in an RCM script:
I I I

scriptinfo Gathers script information register Registers interest in resources resourceinfo Gathers resource information

You might include some or all of the following RCM commands:

I I I I

queryremove Queries whether the resource can be released preremove Releases the resource postremove Provides post-resource removal notication undoremove Undoes the actions done in preremove

For a complete description of these RCM commands, see rcmscript(4).

RCM Script Processing Environment

When you dynamically remove a device, the RCM daemon runs the following:
I

The scripts register command to gather the list of resources (device names) that are identied in the script. The scripts queryremove and preremove commands prior to removing the resource if the scripts registered resources are affected by the dynamic remove operation. The scripts postremove command if the remove operation succeeds. However, if the remove operation fails, the RCM daemon runs the scripts undoremove command.

RCM Script Tasks

The following sections describe the RCM script tasks for application developers and system administrators.

Chapter 6 Dynamically Conguring Devices (Tasks)

105

Application Developer RCM Script (Task Map)

The following task map describes the tasks for an application developer who is creating an RCM script.

Task

Description

For Instructions

1. Identify the resources your application uses.

Identify the resources (device names) your application uses that you could potentially dynamically remove. Identify the commands for notifying the application to cleanly release the resource from the application. Include the commands for notifying the application of the resource removal. Include the commands for notifying the application of the available resource. Write the RCM script based on the information identied in tasks 1-4. Add the script to the appropriate script directory. Test the script by running the script commands manually and by initiating a dynamic reconguration operation.

cfgadm(1M)

2. Identify the commands to release the resource.

Application documentation

3. Identify the commands for post-removal of the resource. 4. Identify the commands if the resource removal fails. 5. Write the RCM script.

rcmscript(4)

rcmscript(4)

Tape Backup RCM Script Example on page 109 How to Install an RCM Script on page 108 How to Test an RCM Script on page 108

6. Install the RCM script. 7. Test the RCM script

System Administrator RCM Script (Task Map)

The following task map describes the tasks for a system administrator who is creating an RCM script to do site customization.

Task

Description

For Instructions

1. Identify the resources to be dynamically removed.

Identify the resources (device cfgadm(1M) names) to be potentially removed by using the cfgadm -l command.

106

System Administration Guide: Devices and File Systems June 2005

Task

Description

For Instructions

2. Identify the applications to be stopped.

Identify the commands for stopping the applications cleanly.

Application documentation

3. Identify the commands for Identify the actions to be pre-removal and post-removal taken before and after the of the resource. resource is removed. 4. Write the RCM script. Write the RCM script based on the information identied in tasks 1-3. Add the script to the appropriate script directory. Test the script by running the script commands manually and by initiating a dynamic reconguration operation.

rcmscript(4)

Tape Backup RCM Script Example on page 109 How to Install an RCM Script on page 108 How to Test an RCM Script on page 108

5. Install the RCM script. 6. Test the RCM script.

Naming an RCM Script

A script must be named as vendor,service where the following applies: vendor service Is the stock symbol of the vendor that provides the script, or any distinct name that identies the vendor. Is the name of the service that the script represents.

Installing or Removing an RCM Script

You must be superuser (root) to install or remove an RCM script. Use this table to determine where you should install your RCM script.
TABLE 61

RCM Script Directories

Script Type

Directory Location

/etc/rcm/scripts /usr/platform/uname -i/lib/rcm/scripts /usr/platform/uname -m/lib/rcm/scripts /usr/lib/rcm/scripts

Scripts for specic systems Scripts for a specic hardware implementation Scripts for a specic hardware class Scripts for any hardware

Chapter 6 Dynamically Conguring Devices (Tasks)

107

M
Steps

How to Install an RCM Script

1. Become superuser. 2. Copy the script to the appropriate directory. See Table 61. For example:
# cp SUNW,sample.pl /usr/lib/rcm/scripts

3. Change the user ID and the group ID of the script to the desired values.
# chown user:group /usr/lib/rcm/scripts/SUNW,sample.pl

4. Send SIGHUP to the RCM daemon.

# pkill -HUP -x -u root rcm_daemon

M
Steps

How to Remove an RCM Script

1. Become superuser. 2. Remove the script from the RCM script directory. For example:
# rm /usr/lib/rcm/scripts/SUNW,sample.pl

3. Send SIGHUP to the RCM daemon.

# pkill -HUP -x -u root rcm_daemon

M
Steps

How to Test an RCM Script

1. Set environment variables, such as RCM_ENV_FORCE, in the command-line shell before running your script. For example, in the Korn shell, use the following:
$ export RCM_ENV_FORCE=TRUE

2. Test the script by running the script commands manually from the command line. For example:
$ script-name scriptinfo $ script-name register $ script-name preremove resource-name 108 System Administration Guide: Devices and File Systems June 2005

$ script-name postremove resource-name

3. Make sure that each RCM script command in your script prints appropriate output to stdout. 4. Install the script in the appropriate script directory. For more information, see How to Install an RCM Script on page 108. 5. Test the script by initiating a dynamic remove operation. For example, assume your script registers the device, /dev/dsk/c1t0d0s0. Try these commands.
$ cfgadm -c unconfigure c1::dsk/c1t0d0 $ cfgadm -f -c unconfigure c1::dsk/c1t0d0 $ cfgadm -c configure c1::dsk/c1t0d0

Caution Make sure that you are familiar with these commands because they can alter the state of the system and cause system failures.

Tape Backup RCM Script Example

This example illustrates how to use an RCM script for tape backups.

What the Tape Backup RCM Script Does

The tape backup RCM script performs the following steps: 1. Sets up a dispatch table of RCM commands. 2. Calls the dispatch routine that corresponds to the specied RCM command and exits with status 2 for unimplemented RCM commands. 3. Sets up the scriptinfo section.
rcm_script_func_info=Tape backup appl script for DR

4. Registers all tape drives in the system by printing all tape drive device names to stdout.
rcm_resource_name=/dev/rmt/$f

If an error occurs, the script prints the error information to stdout.

rcm_failure_reason=$errmsg

5. Sets up the resource information for the tape device.

rcm_resource_usage_info=Backup Tape Unit Number $unit Chapter 6 Dynamically Conguring Devices (Tasks) 109

6. Sets up the preremove information by checking if the backup application is using the device. If the backup application is not using the device, the dynamic reconguration operation continues. If the backup application is using the device, the script checks RCM_ENV_FORCE. If RCM_ENV_FORCE is set to FALSE, the script denies the dynamic reconguration operation and prints the following message:
rcm_failure_reason=tape backup in progress pid=...

If RCM_ENV_FORCE is set to TRUE, the backup application is stopped, and the reconguration operation proceeds.

Outcomes of the Tape Backup Reconguration Scenarios

Here are the various outcomes if you use the cfgadm command to remove a tape device without the RCM script.
I

If you use the cfgadm command and the backup application is not using the tape device, the operation succeeds. If you use the cfgadm command and the backup application is using the tape device, the operation fails.

Here are the various outcomes if you use the cfgadm command to remove a tape device with the RCM script.
I

If you use the cfgadm command and the backup application is not using the tape device, the operation succeeds. If you use the cfgadm command without the -f option and the backup application is using the tape device, the operation fails with an error message similar to the following:
tape backup in progress pid=...

If you use the cfgadm -f command and the backup application is using the tape device, the script stops the backup application and the cfgadm operation succeeds.

ExampleTape Backup RCM Script

#! /usr/bin/perl -w # # A sample site customization RCM script. # # When RCM_ENV_FORCE is FALSE this script indicates to RCM that it cannot # release the tape drive when the tape drive is being used for backup. # # When RCM_ENV_FORCE is TRUE this script allows DR removing a tape drive # when the tape drive is being used for backup by killing the tape # backup application. #

110

System Administration Guide: Devices and File Systems June 2005

use strict; my ($cmd, %dispatch); $cmd = shift(@ARGV); # dispatch table for RCM commands %dispatch = ( "scriptinfo" => "register" => "resourceinfo" => "queryremove" => "preremove" => );

\&do_scriptinfo, \&do_register, \&do_resourceinfo, \&do_preremove, \&do_preremove

if (defined($dispatch{$cmd})) { &{$dispatch{$cmd}}; } else { exit (2); } sub do_scriptinfo { print "rcm_script_version=1\n"; print "rcm_script_func_info=Tape backup appl script for DR\n"; exit (0); } sub do_register { my ($dir, $f, $errmsg); $dir = opendir(RMT, "/dev/rmt"); if (!$dir) { $errmsg = "Unable to open /dev/rmt directory: $!"; print "rcm_failure_reason=$errmsg\n"; exit (1); } while ($f = readdir(RMT)) { # ignore hidden files and multiple names for the same device if (($f !~ /^\./) && ($f =~ /^[0-9]+$/)) { print "rcm_resource_name=/dev/rmt/$f\n"; } } closedir(RMT); exit (0); } sub do_resourceinfo { my ($rsrc, $unit); $rsrc = shift(@ARGV); if ($rsrc =~ /^\/dev\/rmt\/([0-9]+)$/) { Chapter 6 Dynamically Conguring Devices (Tasks) 111

$unit = $1; print "rcm_resource_usage_info=Backup Tape Unit Number $unit\n"; exit (0); } else { print "rcm_failure_reason=Unknown tape device!\n"; exit (1); } } sub do_preremove { my ($rsrc); $rsrc = shift(@ARGV); # check if backup application is using this resource #if (the backup application is not running on $rsrc) { # allow the DR to continue # exit (0); #} # # If RCM_ENV_FORCE is FALSE deny the operation. # If RCM_ENV_FORCE is TRUE kill the backup application in order # to allow the DR operation to proceed # if ($ENV{RCM_ENV_FORCE} eq TRUE) { if ($cmd eq preremove) { # kill the tape backup application } exit (0); } else { # # indicate that the tape drive can not be released # since the device is being used for backup by the # tape backup application # print "rcm_failure_reason=tape backup in progress pid=...\n" ; exit (3); } }

112

System Administration Guide: Devices and File Systems June 2005

CHAPTER

Using USB Devices (Overview)

This chapter provides an overview of Universal Serial Bus (USB) devices in the Solaris OS. This is a list of the overview information in this chapter.
I I I

Whats New in USB Devices? on page 113 Overview of USB Devices on page 114 About USB in the Solaris OS on page 119

For recent information about USB devices, go the following site: http://www.sun.com/io_technologies/USB-Faq.html For step-by-step instructions on using USB devices in the Solaris OS, see Chapter 8. For general information about dynamic reconguration and hot-plugging, see Chapter 6. For information on conguring USB printers, see Whats New in Printing? in System Administration Guide: Advanced Administration.

Whats New in USB Devices?

For information about new USB features in the Solaris 10 release, see the following references:
I I I

USB Wheel Mouse Support on page 122 USB 2.0 Features on page 119 USB Driver Enhancements on page 117

For a complete listing of new Solaris features and a description of Solaris releases, see Solaris 10 Whats New.
113

Solaris Support for USB Devices

Solaris 10: All USB 1.1 devices are supported on a USB 2.0 hub, including audio devices. Use the following table to identify Solaris support information for specic USB 1.1 and USB 2.0 devices.

Solaris 8 HW 5/03 and later releases

Solaris 9 releases

Solaris 10

USB 1.1

SPARC and x86

SPARC and x86 Not supported on a USB 2.0 hub SPARC and x86 (Solaris 9 4/04) Not supported Supported on a USB 2.0 hub (Solaris 9 4/04)

SPARC and x86 Supported on a USB 2.0 hub SPARC and x86 Not supported Supported on a USB 2.0 hub

USB 1.1 audio devices Not supported on a USB 2.0 hub USB 2.0 SPARC

USB 2.0 audio devices Not supported USB 2.0 storage devices Supported on a USB 2.0 hub

For task information associated with mass storage devices, see Chapter 8.

Overview of USB Devices

Universal Serial Bus (USB) was developed by the PC industry to provide a low-cost solution for attaching peripheral devices, such as keyboards, mouse devices, and printers, to a system. USB connectors are designed to t only one type of cable, in one way. The primary design motivation for USB was to alleviate the need for multiple connector types for different devices. This design reduces the clutter on the back panel of a system. Devices connect to USB ports on external USB hubs, or on a root hub that is located on the computer itself. Since hubs have several ports, several branches of a device tree can stem from a hub.

Commonly Used USB Acronyms

The following table describes the USB acronyms that are used in the Solaris OS. For a complete description of USB components and acronyms, go to: http://www.usb.org
114 System Administration Guide: Devices and File Systems June 2005

Acronym

Denition

UGEN USB USBA USBAI HCD EHCI OHCI UHCI

USB generic driver Universal Serial Bus Universal Serial Bus Architecture (Solaris) USBA Client Driver Interface (Solaris) USB host controller driver Enhanced Open Controller Interface Open Host Controller Interface Universal Host Controller Interface

USB Bus Description

The USB specication is openly available and free of royalties. The specication denes the electrical and mechanical interfaces of the bus and the connectors. USB employs a topology in which hubs provide attachment points for USB devices. The host controller contains the root hub, which is the origin of all USB ports in the system. For more information about hubs, see USB Host Controller and Hubs on page 123.
System

Zip drive

Hub

Hub Printer

cdrw

Keyboard

Mouse

USB Host Controller and Root Hub Compound Device Composite Device

FIGURE 71

USB Physical Device Hierarchy

Chapter 7 Using USB Devices (Overview)

115

Figure 71 shows a system with three active USB ports. The rst USB port connects a Zip drive. The second USB port connects an external hub, which in turn, connects a cdrw device and a composite keyboard/mouse device. As a composite device, this keyboard contains a USB controller, which operates both the keyboard and an attached mouse. The keyboard and the mouse share a common USB bus address because they are directed by the same USB controller. Figure 71 also shows an example of a hub and a printer as a compound device. The hub is an external hub that is enclosed in the same casing as the printer. The printer is permanently connected to the hub. The hub and printer have separate USB bus addresses. The device tree path name for some of the devices that are displayed in Figure 71 are listed here. Zip drive Keyboard Mouse cdrw device Printer /pci@1f,4000/usb@5/storage@1 /pci@1f,4000/usb@5/hub@2/device@1/keyboard@0 /pci@1f,4000/usb@5/hub@2/device@1/mouse@1 /pci@1f,4000/usb@5/hub@2/storage@3 /pci@1f,4000/usb@5/hub@3/printer@1

USB Devices and Drivers

USB devices with similar attributes and services are grouped into device classes. Each device class has a corresponding driver. Devices within a class are managed by the same device driver pair. However, the USB specication also allows for vendor-specic devices that are not part of a specic class. The Human Interface Device (HID) class contains devices that are user-controlled such as the following devices:
I I I

Keyboards Mouse devices Joysticks

The Communication Device class contains devices that connect to a telephone, such as the following devices:
I I

Modems ISDN interface

Other device classes include the following classes:

I I I I

Audio Monitor Printer Storage Device

116

System Administration Guide: Devices and File Systems June 2005

Each USB device contains descriptors that reect the class of the device. A device class species how its members should behave in conguration and data transfer. You can obtain additional class information from: http://www.usb.org

USB Driver Enhancements

Starting in the Solaris 9 4/04 release, the following USB driver enhancements are included.
I

Generic USB driver USB devices can now be accessed and manipulated by applications using standard UNIX read(2) and write(2) system calls, and without writing a special kernel driver. Additional features include:
I I

Applications have access to raw device data and device status. The driver supports control, bulk, and interrupt (in and out) transfers.

For more information, refer to ugen(7D) and the USB DDK at: http://developers.sun.com/solaris/developer/support/driver/usb.html
I

Digi Edgeport USB support The driver provides support for several Digi Edgeport USB to serial port converter devices.
I I

New devices are accessed as /dev/term/[0-9]* and /dev/cua/[0-9]*. USB serial ports are usable as any other serial port would be, except that they cannot serve as a local serial console. The fact that their data is run through a USB port is transparent to the user.

For more information, see usbser_edge(7D), or go to the following sites:

I I I

http://www.digi.com http://www.sun.com/io

Documentation and binary support for user-written kernel and userland drivers A Solaris USB Driver Development Kit (DDK) is available. For up-to-date information on USB driver development, including information on the DDK, go to: http://developers.sun.com/solaris/developer/support/driver/usb.html

The EHCI, OHCI, and UHCI Drivers

Features of the EHCI driver include:
I I I

Complies with enhanced host controller interface that supports USB 2.0. Supports high-speed control, bulk, and interrupt transfers. Currently, there is no support for high-speed isochronous.

Chapter 7 Using USB Devices (Overview)

117

If there are USB 2.0 and USB 1.x devices on the system, the EHCI and OHCI or UHCI drivers hand-off device control depending upon the type of device that is connected to the system.
I

The USB 2.0 PCI card has one EHCI controller and one or more OHCI or UHCI controllers. A USB 1.1 device is dynamically assigned to the OHCI or UHCI controller when it is plugged in. A USB 2.0 device is dynamically assigned to the EHCI controller when it is plugged in.

Use the prtconf command output to identify whether your system supports USB 1.0 or USB 2.0 devices. For example:
# prtconf -D | egrep "ehci|ohci|uhci"

If your prtconf output identies an EHCI controller, your system supports USB 2.0 devices. If your prtconf output identies an OHCI or UHCI controller, your system supports USB 1.1 devices.

Solaris USB Architecture (USBA)

USB devices can be represented as two levels of device tree nodes. A device node represents the entire USB device. One or more child interface nodes represent the individual USB interfaces on the device. Driver binding is achieved by using the compatible name properties. For more information, refer to 3.2.2.1 of the IEEE 1275 USB binding and Writing Device Drivers. A driver can either bind to the entire device and control all the interfaces, or can bind to just one interface. If no vendor or class driver claims the entire device, a generic USB multi-interface driver is bound to the device-level node. This driver attempts to bind drivers to each interface by using compatible names properties, as dened in section 3.3.2.1 of the IEEE 1275 binding specication. The Solaris USB Architecture (USBA) adheres to the USB 1.1 and USB 2.0 specications plus Solaris driver requirements. The USBA model is similar to Sun Common SCSI Architecture (SCSA). As the following gure shows, the USBA is a thin layer that provides a generic USB transport-layer abstraction to client drivers, providing them with services that implement core generic USB functionality.

118

System Administration Guide: Devices and File Systems June 2005

client drivers USBA Host controller drivers Bus with devices

Solaris USB Architecture (USBA)

FIGURE 72

About USB in the Solaris OS

This section describes information you should know about USB in the Solaris OS.

USB 2.0 Features

Starting in the Solaris 9 4/04 release, the following USB 2.0 features are included:
I

Better performance Increased data throughput for devices connected to USB 2.0 controllers, up to 40 times faster than USB 1.1 devices. You can take advantage of the high-speed USB protocol when accessing high-speed mass storage devices, such as DVDs and hard disks.

Backward Compatibility Compatibility with 1.0 and 1.1 devices and drivers so that you can use the same cables, connectors, and software interfaces.

For a description of USB devices and terminology, see Overview of USB Devices on page 114.

USB 2.0 Device Features and Compatibility Issues

USB 2.0 devices are dened as high-speed devices that follow the USB 2.0 specication. You can refer to the USB 2.0 specication at http://www.usb.org. To identify the speed of your USB device, check the /var/adm/messages le for messages similar to the following:
Chapter 7 Using USB Devices (Overview) 119

Dec 13 17:05:57 mysystem usba: [ID 912658 kern.info] USB 2.0 device (usb50d,249) operating at hi speed (USB 2.x) on USB 2.0 external hub: storage@4, scsa2usb0 at bus address 4

Here are some of the USB devices that are supported in this Solaris release:
I

Mass storage devices CD-RWs, hard disks, DVDs, digital cameras, diskettes, tape drives, memory sticks, and multi-format card readers Keyboards, mouse devices, speakers and microphones Audio devices

I I

For a full listing of USB devices that have been veried on the Solaris release, go to: http://www.sun.com/io_technologies/USB.html Additional storage devices might work by modifying the scsa2usb.conf le. For more information, see scsa2usb(7D). Solaris USB 2.0 device support includes the following features:
I

Increased USB bus speed from 12 Mbyte/sec to 480 Mbyte/sec. This increase means devices that support the USB 2.0 specication can run signicantly faster than their USB 1.1 counterparts, when they are connected to a USB 2.0 port. A USB 2.0 port is dened as follows:
I I

A port on a USB 2.0 PCI card A port on a USB 2.0 hub that is connected to USB 2.0 port

USB 2.0 is Solaris Ready on all PCI-based platforms. A USB 2.0 PCI card is needed to provide USB 2.0 ports. For a list of USB 2.0 PCI cards that have been veried for the Solaris release, go to: http://www.sun.com/io_technologies/USB.html USB 1.1 devices work as they have in the past, even if you have both USB 1.1 and USB 2.0 devices on the same system. While USB 2.0 devices operate on a USB 1.x port, their performance is signicantly better when they are connected to a USB 2.0 port. A USB 2.0 host controller has one high-speed Enhanced Host Controller Interface (EHCI) and one or more low-speed or full-speed OpenHCI Host Controller Interface (OHCI) embedded controllers. Devices connected to a USB 2.0 port are dynamically assigned to either an EHCI or OHCI controller, depending on whether they support USB 2.0.

120

System Administration Guide: Devices and File Systems June 2005

Note USB 2.0 storage devices connected to a port on a USB 2.0 PCI card, and that were used with a prior Solaris release in the same hardware conguration, can change device names after upgrading to this release. This change occurs because these devices are now seen as USB 2.0 devices and are taken over by the EHCI controller. The controller number, w in /dev/[r]dsk/cwtxdysz, is changed for these devices.

For more information on USB 2.0 device support, see ehci(7D) and usba(7D).

USB 2.0 Cables

I I

The maximum cable length supported is 5 meters. Do not use passive cable extenders. For best results, use a self-powered hub to extend cable length. For more information, go to: http://www.usb.org/channel/faq/ans5

Bus-Powered Devices
Bus-powered hubs use power from the USB bus to which they are connected, to power devices connected to them. Special care must be taken to not overload these hubs, because the power these hubs offer to their downstream devices is limited.
I

Do not cascade bus-powered hubs. For example, do not connect one bus-powered hub to another bus-powered hub. Avoid connecting bus-powered devices to bus-powered hubs, except for low-speed, low-power devices, such as keyboards or mouse devices. Connecting high-powered devices such as disks, speakers, or microphones to a bus-powered hub could cause power shortages for all devices connected to that hub. This scenario could cause these devices to behave unpredictably.

USB Keyboards and Mouse Devices

System congurations with multiple USB keyboards and mouse devices might work, but are not supported in the Solaris release. See the following for details.
I

A USB keyboard and mouse can be connected anywhere on the bus and can be congured as the console keyboard and mouse. Booting the system is slower if the keyboard and mouse are connected to an external hub.

Chapter 7 Using USB Devices (Overview)

121

Do not move the console keyboard and mouse during a reboot or at the ok prompt. You can move the console keyboard and mouse to another hub at any time after a system reboot. After you plug in a keyboard and mouse, they are fully functional again. SPARC The power key on a USB keyboard behaves differently than the power key on the Sun type 5 keyboard. On a USB keyboard, you can suspend or shut down the system by using the SUSPEND/SHUTDOWN key. However, you cannot use that key to power up the system. The keys just to the left of the keypad do not function on third-party USB keyboards. Multiple keyboards are not supported:
I

Multiple keyboards enumerate and are usable, but they are not plumbed as console keyboards. The rst keyboard that is probed at boot time becomes the console keyboard. The result of this probing might cause confusion if multiple keyboards are plugged in at boot time. If you unplug the console keyboard, the next available USB keyboard does not become the console keyboard. The next hot-plugged keyboard becomes the console keyboard. Multiple mouse devices enumerate and are usable, but they are not plumbed as console mouse devices. The rst mouse that is probed at boot time becomes the console mouse. The result of this probing might cause confusion if you have multiple mouse devices plugged in at boot time. If you unplug the console mouse, the next available USB mouse does not become the console mouse. The next hot-plugged mouse becomes the console mouse.

Multiple mouse devices are not supported:

I

If you have a third-party composite keyboard with a PS/2 mouse, and the composite keyboard/mouse is the rst one to be probed, it becomes the console keyboard/mouse even if the PS/2 mouse is not plugged in. Thus, another USB mouse plugged into the system cannot work because it is not congured as the console mouse.

USB Wheel Mouse Support

Starting in the Solaris 9 9/04 release, the following wheel mouse features are supported:
I

Support for more than 3 buttons is available on USB or PS/2 mouse devices.

122

System Administration Guide: Devices and File Systems June 2005

Wheel mouse scrolling is available on a USB or PS/2 mouse device. This support means that rolling the wheel on a USB or a PS/2 mouse results in a scroll in the application or window under mouse focus. StarOffice, Mozilla, and GNOME applications support wheel mouse scrolling. However, other applications might not support wheel mouse scrolling.

USB Host Controller and Hubs

A USB hub is responsible for the following:
I I I

Monitoring the insertion or removal of a device on its ports Power managing individual devices on its ports Controlling power to its ports

The USB host controller has an embedded hub called the root hub. The ports that are visible at the systems back panel are the ports of the root hub. The USB host controller is responsible for the following:
I I

Directing the USB bus. Individual devices cannot arbitrate for the bus. Polling the devices by using a polling interval that is determined by the device. The device is assumed to have sufficient buffering to account for the time between the polls. Sending data between the USB host controller and its attached devices. Peer-to-peer communication is not supported.

USB Hub Devices

I

Do not cascade hubs beyond four levels on either SPARC based systems or x86 based systems. On SPARC systems, the OpenBoot PROM cannot reliably probe beyond four levels of devices. Do not plug a bus-powered hub into another bus-powered hub in a cascading style. A bus-powered hub does not have its own power supply. Do not connect a device that requires a large amount of power to a bus-powered hub. These devices might not work well on bus-powered hubs or might drain the hub of power for other devices. An example of such a device is a USB diskette device.

SPARC: USB Power Management

Suspending and resuming USB devices is fully supported on SPARC systems. However, do not suspend a device that is busy and never remove a device when the system is powered off under a suspend shutdown.
Chapter 7 Using USB Devices (Overview) 123

The USB framework makes a best effort to power manage all devices on SPARC based systems with power management enabled. Power managing a USB device means that the hub driver suspends the port to which the device is connected. Devices that support remote wake up can notify the system to wake up everything in the devices path so that the device can be used. The host system could also wake up the device if an application sends an I/O to the device. All HID devices (keyboard, mouse, hub, and storage devices), hub devices, and storage devices are power managed by default if they support remote wake-up capability. A USB printer is power managed only between two print jobs. Devices that are directed by the generic USB driver (UGEN) are power managed only when they are closed. When power management is running to reduce power consumption, USB leaf devices are powered down rst. After all devices that are connected to a hubs ports are powered down, the hub is powered down after some delay. To achieve the most efficient power management, do not cascade many hubs.

Guidelines for USB Cables

Keep the following guidelines in mind when connecting USB cables:
I

Always use USB 2.0 compliant, fully rated (480 Mbit/sec) 20/28 AWG cables for connecting USB 2.0 devices. The maximum cable length that is supported is 5 meters. Do not use cable extenders. For best results, use a self-powered hub to extend cable length.

I I

For more information, go to: http://www.usb.org/channel/faq/ans5

124

System Administration Guide: Devices and File Systems June 2005

CHAPTER

Using USB Devices (Tasks)

This chapter provides step-by-step instructions for using USB devices in the Solaris OS. For information on the procedures associated with using USB devices, see the following:
I I I I

Managing USB Devices in the Solaris OS (Roadmap) on page 125 Using USB Mass Storage Devices (Task Map) on page 126 Using USB Audio Devices (Task Map) on page 141 Hot-Plugging USB Devices With the cfgadm Command (Task Map) on page 145

For overview information about using USB devices, see Chapter 7.

Managing USB Devices in the Solaris OS (Roadmap)

Use this road map to identify all the tasks for managing USB devices in the Solaris OS. Each task points to a series of additional tasks such as using USB devices, hot-plugging USB devices, and adding USB audio devices. For information about using USB components in the Solaris OS, see About USB in the Solaris OS on page 119.

125

Task

Description

For Instructions

Use USB mass storage devices.

USB mass storage devices must be formatted before le systems can be created and mounted on them. This section also describes how to physically add or remove USB devices from your system.

Using USB Mass Storage Devices (Task Map) on page 126

Add USB audio devices.

Use this task map to identify tasks associated with adding USB audio devices. You can logically add or remove USB devices to and from your system with the cfgadm command.

Using USB Audio Devices (Task Map) on page 141 Hot-Plugging USB Devices With the cfgadm Command (Task Map) on page 145

Add or remove USB devices to and from your system with the cfgadm command.

Using USB Mass Storage Devices (Task Map)

Task Description For Instructions

Add or remove a USB mass storage device.

Select one of the following to add a USB mass storage device: Add a USB mass storage device with vold running. How to Add a USB Mass Storage Device With vold Running on page 130

Add a USB mass storage How to Add a USB Mass device without vold running. Storage Device Without vold Running on page 131 Add a USB camera to access digital images. Select one of the following to remove a USB mass storage device: How to Add a USB Camera on page 131

126

System Administration Guide: Devices and File Systems June 2005

Task

Description

For Instructions

Remove a USB mass storage device with vold running.

How to Remove a USB Mass Storage Device With vold Running on page 132

Remove a USB mass storage How to Remove a USB Mass device without vold running. Storage Device Without vold Running on page 133 Prepare to use a USB mass storage device. Prepare to use a USB mass storage device with vold running. Prepare to use a USB mass storage device without vold running. Display USB device information. Format a USB mass storage device. Mount a USB mass storage device. Use the prtconf command to display information about USB devices. Format a USB mass storage device so that you can put data on it. Mount a USB mass storage device with vold running. Preparing to Use a USB Mass Storage Device With vold Running on page 133 How to Prepare to Use USB Mass Storage Devices Without vold Running on page 134 How to Display USB Device Information (prtconf) on page 135 How to Format a USB Mass Storage Device Without vold Running on page 135 How to Mount or Unmount a USB Mass Storage Device With vold Running on page 137

Mount a USB mass storage How to Mount or Unmount device without vold running. a USB Mass Storage Device Without vold Running on page 138 (Optional) Disable USB device Disable USB device drivers if drivers. you do not want USB support on your system. (Optional) Remove unused USB device links. Remove USB device links with the devfsadm command. How to Disable Specic USB Drivers on page 140 How to Remove Unused USB Device Links on page 140

Using USB Mass Storage Devices

Starting in the Solaris 9 release, the following USB removable mass storage devices are supported:
I

CD-RWs
Chapter 8 Using USB Devices (Tasks) 127

I I I I I

Hards disks DVDs Digital cameras Diskette devices Zip, Peerless, SmartMedia, CompactFlash, and ORB devices

For a complete list of USB devices that are supported in the Solaris OS, see: http://www.sun.com/io_technologies/USB.html All USB storage devices are accessed as removable media devices, which provides the following advantages:
I

USB storage devices with standard MS-DOS or Windows (FAT) le systems are supported. You can use the user-friendly rmformat command instead of the format command to format and partition all USB storage devices. If the functionality of the format command is needed, use the format -e command. You can use the fdisk command if you need to do fdisk-style partitioning. Non-root users can now access USB storage devices, since the root-privileged mount command is no longer needed. The device is automatically mounted by vold and is available under the /rmdisk directory. If a new device is connected while the system is down, do a reconguration boot with the boot -r command so that vold recognizes the device. Note that vold does not automatically recognize a hot-plugged device. If a new device is connected while the system is up, restart vold. For more information, refer to the vold(1M) and scsa2usb(7D) man pages. These devices can be managed with or without volume management. Disks with FAT le systems can be mounted and accessed. For example:
mount -F pcfs /dev/dsk/c2t0d0s0:c /mnt

I I

I I

All USB storage devices are now power managed, except for those that support LOG SENSE pages. Devices with LOG SENSE pages are usually SCSI drives connected through a USB-to-SCSI bridge device. In previous Solaris releases, some USB storage devices were not power managed because they were not seen as removable media. Applications might work differently with USB mass storage devices. Keep the following issues in mind when using applications with USB storage devices:
I

Applications might make incorrect assumptions about the size of the media since only smaller devices like diskettes and Zip drives were removable previously. Requests by applications to eject media on devices where this would be inapplicable, such as a hard drive, will succeed and do nothing. If you prefer the behavior in previous Solaris releases where not all USB mass storage were treated as removable media devices, then you can force the old behavior by updating the /kernel/drv/scsa2usb.conf le.

128

System Administration Guide: Devices and File Systems June 2005

For more information on using USB mass storage devices, see scsa2usb(7D).

Using USB Diskette Devices

USB diskette devices appear as removable media devices just as other USB devices. USB diskette devices are not managed by the fd (oppy) driver. Applications that issue ioctl(2) calls intended for the fd (native oppy) driver will fail. Applications that issue only read(2) and write(2) calls will succeed. Other applications, such as SunPCI and rmformat, will also succeed.
Note Solaris Common Desktop Environments (CDE) File Manager does not fully support USB diskettes at this time. However, you can open, rename, and format diskettes that contain a UFS le system from File Managers Removable Media Manager. You can only open diskettes that contain a PCFS le system from Removable Media Manager. If a diskette contains either type of le system, you can successfully drag and drop les between the diskette and File Manager.

Volume management (vold) sees the USB diskette device as a SCSI removable media device. Volume management makes the device available for access under the /rmdisk directory. For more information on how to use USB diskette devices, see Chapter 1.

Using Non-Compliant USB Mass Storage Devices

Some devices might be supported by the USB mass storage driver even though they do not identify themselves as compliant with the USB mass storage class or identify themselves incorrectly. The scsa2usb.conf le contains an attribute-override list that lists the vendor ID, product ID, and revision for matching mass storage devices, as well as elds for overriding the default device attributes. The entries in this list are commented out by default. These entries can be copied and uncommented to enable support of particular devices. If you connect a USB mass storage device to a system running this Solaris release and the system is unable to use it, you can check the /kernel/drv/scsa2usb.conf le to see if there is a matching, commented entry for this device. Follow the information given in the scsa2usb.conf le to see if a particular device can be supported by using the override information. For a listing of recommended USB mass storage devices, go to: http://www.sun.com/io_technologies/USB.html For more information, see scsa2usb(7D).
Chapter 8 Using USB Devices (Tasks) 129

Hot-Plugging USB Mass Storage Devices

Hot-plugging a device means the device is added or removed without shutting down the operating system or powering off the system. All USB devices are hot-pluggable. When you hot-plug a USB device, the device is immediately seen in the systems device hierarchy, as displayed in the prtconf command output. When you remove a USB device, the device is removed from the systems device hierarchy, unless the device is in use. If the USB device is in use when it is removed, the hot-plug behavior is a little different. If a device is in use when it is unplugged, the device node remains, but the driver controlling this device stops all activity on the device. Any new I/O activity issued to this device is returned with an error. In this situation, the system prompts you to plug in the original device. If the device is no longer available, stop the applications. After a few seconds, the port becomes available again.
Note Data integrity might be impaired if you remove an active or open device. Always close the device before removing, except the console keyboard and mouse, which can be moved while active.

M How to Add a USB Mass Storage Device With vold

Running
This procedure describes how to add a USB device with vold running. Or, instead of using this procedure, you can stop and restart vold.
Steps

1. Connect the USB mass storage device. 2. Instruct vold to scan for new devices.
# touch /etc/vold.conf

3. Restart vold.
# pkill -HUP vold

4. Verify that the device has been added.

$ ls device-alias

For more information on volume management device names, see Chapter 1.

130

System Administration Guide: Devices and File Systems June 2005

M How to Add a USB Mass Storage Device Without vold

Running
Steps

1. If needed, see How to Prepare to Use USB Mass Storage Devices Without vold Running on page 134 for information on disabling vold. 2. Connect the USB mass storage device. 3. Verify that the USB device has been added. Locate the USB disk device links, which might be among device links of non-USB storage devices, as follows:
$ cd /dev/rdsk $ ls -l c*0 | grep usb lrwxrwxrwx 1 root root 67 Apr 30 15:12 c1t0d0s0 -> ../../devices/pci@1f,0/pci@5/pci@0/usb@8,2/storage@1/disk@0,0:a,raw

M How to Add a USB Camera

Steps

1. Become superuser. 2. Plug in and turn on the USB camera. The system creates a logical device for the camera. After the camera is plugged in, output is written to the /var/adm/messages le to acknowledge the devices connection. The system treats the camera as a storage device. 3. Examine the output that is written to the /var/adm/messages le.
# more /var/adm/messages

Examining this output enables you to determine which logical device was created so that you can then use that device to access your images. The output looks similar to the following:
Jul 15 09:53:35 buffy usba: [ID 349649 kern.info] OLYMPUS, C-3040ZOOM, 000153719068 Jul 15 09:53:35 buffy genunix: [ID 936769 kern.info] scsa2usb1 is /pci@0,0/pci925,1234@7,2/storage@2 Jul 15 09:53:36 buffy scsi: [ID 193665 kern.info] sd3 at scsa2usb1: target 0 lun 0

Match the device with a mountable /dev/dsk link entry, by doing the following:
# ls -l /dev/dsk/c*0 | grep /pci@0,0/pci925,1234@7,2/storage@2 lrwxrwxrwx 1 root root 58 Jun 30 2004 c3t0d0p0 -> ../../devices/pci@0,0/pci925,1234@7,2/storage@2/disk@0,0:a

4. Mount the USB camera le system. The cameras le system is most likely a PCFS le system. To mount the le system on the device created, the slice that represents the disk must be specied. The slice
Chapter 8 Using USB Devices (Tasks) 131

is normally s0 for a SPARC system, and p0 for an x86 system. For example, to mount the le system on an x86 system, you would execute the following command:
# mount -F pcfs /dev/dsk/c3t0d0p0:c /mnt

To mount the le system on a SPARC system, you would execute the following command:
# mount -F pcfs /dev/dsk/c3t0d0s0:c /mnt

For information on mounting le systems, see Chapter 19. For information on mounting different PCFS le systems, see mount_pcfs(1M). 5. Verify that the image les are available. For example:
# ls /mnt/DCIM/100OLYMP/ P7220001.JPG* P7220003.JPG* P7220002.JPG* P7220004.JPG* P7220005.JPG* P7220006.JPG*

6. View and manipulate the image les created by the USB camera. For example:
# /usr/dt/bin/sdtimage P7220001.JPG &

7. (Optional) Unmount the le system before disconnecting the camera. For example:
# umount /mnt

8. (Optional) Turn off and disconnect the camera.

M How to Remove a USB Mass Storage Device With vold

Running
The following procedure uses a Zip drive in the example of removing a USB device with vold running.
Steps

1. Stop any active applications that are using the device. 2. Unmount the device. For example:
$ volrmmount -e zip0

3. Eject the device. For example:

$ eject zip0 132 System Administration Guide: Devices and File Systems June 2005

4. Become superuser and stop vold.

# /etc/init.d/volmgt stop

5. Remove the USB mass storage device. 6. Start vold.

# /etc/init.d/volmgt start

M How to Remove a USB Mass Storage Device Without

vold Running
Steps

1. If needed, see How to Prepare to Use USB Mass Storage Devices Without vold Running on page 134 for information on disabling vold. 2. Become superuser. 3. Stop any active applications that are using the device. 4. Unmount the device. 5. Remove the device.

Preparing to Use a USB Mass Storage Device With vold Running

If you are running CDE, USB removable mass storage devices are managed by the Removable Media Manager component of the CDE File Manager. For more information on the CDE File Manager, see dtfile(1).
Note You must include the /usr/dt/man directory in your MANPATH variable to display the man pages that are listed in this section. You must also have the /usr/dt/bin directory in your path and have CDE running to use these commands. Or, you must have a DISPLAY variable set to use these commands remotely.

The following table identies the commands that Removable Media Manager uses to manage storage devices in the CDE environment.

Chapter 8 Using USB Devices (Tasks)

133

Command

Task

Man Page

sdtmedia_format sdtmedia_prop sdtmedia_prot sdtmedia_slice

Format and label a device Display properties of a device Change device protection Create or modify slices on a device

sdtmedia_format(1) sdtmedia_prop(1) sdtmedia_prot(1) sdtmedia_slice(1)

After the USB device is formatted, it is usually mounted under the /rmdisk/label directory. For more information on conguring removable storage devices, see rmmount.conf(4) or vold.conf(4). The device nodes are created under the /vol/dev directory. For more information, see scsa2usb(7D). The following procedures describe how to manage USB mass storage devices without vold running. The device nodes are created under the /dev/rdsk directory for character devices and under the /dev/dsk directory for block devices. Device links are created when the devices are hot-plugged. For more information, see scsa2usb(7D).

How to Prepare to Use USB Mass Storage Devices Without vold Running
You can use USB mass storage devices without volume management (vold) running. Stop vold by issuing the following command:
# /etc/init.d/volmgt stop

Or, use the following procedure to keep vold running, but do not register the USB mass storage devices with vold.
Steps

1. Become superuser. 2. Remove volume management registration of USB mass storage devices by commenting the following line in the /etc/vold.conf le.
# use rmdisk drive /dev/rdsk/c*s2 dev_rmdisk.so rmdisk%d

3. After this line is commented, restart vold.

# /etc/init.d/volmgt start

134

System Administration Guide: Devices and File Systems June 2005

Caution If you comment out this line and other removable devices, such as SCSI, ATAPI Zip, or Peerless, are on the system, vold registration for these devices is disabled as well.

For more information, see vold.conf(4).

How to Display USB Device Information (prtconf)

G Display information about USB devices.

Step

For example:
$ prtconf usb, instance #0 hub, instance #2 device, instance #8 interface (driver not attached) printer (driver not attached) mouse, instance #14 device, instance #9 keyboard, instance #15 mouse, instance #16 storage, instance #7 disk (driver not attached) communications, instance #10 modem (driver not attached) data (driver not attached) storage, instance #0 disk (driver not attached) storage, instance #1 disk (driver not attached)

How to Format a USB Mass Storage Device Without vold Running

USB mass storage devices, as with all other devices used by the Solaris OS, must be formatted and contain a le system before they can be used. USB mass storage devices, including diskettes, support both PCFS and UFS le systems. Be sure the diskette is formatted before putting either a PCFS or UFS le system on it.

Steps

1. See How to Prepare to Use USB Mass Storage Devices Without vold Running on page 134 for information on disabling vold.
Chapter 8 Using USB Devices (Tasks) 135

2. (Optional) Add the USB diskette device to your system. For information on hot-plugging USB devices, see:
I I

Using USB Audio Devices (Task Map) on page 141 Hot-Plugging USB Devices With the cfgadm Command (Task Map) on page 145

3. (Optional) Identify the diskette device. For example:

# cd /dev/rdsk # ls -l c*0 | grep usb lrwxrwxrwx 1 root root 55 Mar 5 10:35 c2t0d0s0 -> ../../devices/pci@1f,0/usb@c,3/storage@3/disk@0,0:a,raw

In this example, the diskette device is c2t0d0s0. 4. Insert a diskette into the diskette drive. 5. Format the diskette.
% rmformat -Flong raw-device

For example:
% rmformat -Flong /dev/rdsk/c2t0d0s0

6. Determine the le system type and select one of the following:

I

Create a PCFS le system.

# mkfs -F pcfs -o nofdisk,size=size raw-device

Specify the -size option in 512-byte blocks. The following example shows how to create a PCFS le system on a 1.4-Mbyte diskette:
# mkfs -F pcfs -o nofdisk,size=2880 /dev/rdsk/c4t0d0s0

The following example shows how to create a PCFS le system on a 100-Mbyte Zip drive:
# mkfs -F pcfs -o nofdisk,size=204800 /dev/rdsk/c5t0d0s0

This command can take several minutes to complete.

I

Create a UFS le system.

# newfs raw-device

For example:
# newfs /dev/rdsk/c4t0d0s0

136

System Administration Guide: Devices and File Systems June 2005

Note UFS le system overhead consumes a signicant portion of space on a diskette, due to a diskettes limited storage capacity.

How to Mount or Unmount a USB Mass Storage Device With vold Running
1. Display device aliases for all removable mass storage devices, including USB mass storage devices.
$ eject -n . . . cdrom0 -> /vol/dev/rdsk/c0t6d0/audio_cd (Generic CD device) zip0 -> /vol/dev/rdsk/c1t0d0/zip100 (USB Zip device) zip1 -> /vol/dev/rdsk/c2t0d0/fat32 (USB Zip device) rmdisk0 -> /vol/dev/rdsk/c5t0d0/unnamed_rmdisk (Peerless, HD or floppy) rmdisk1 -> /vol/dev/rdsk/c4t0d0/clik40 (Generic USB storage)

Steps

2. Select one of the following to mount or unmount a USB mass storage device.
I

Mount a USB mass storage device by using the device aliases listed previously.
$ volrmmount -i device-alias

This example shows how to mount a USB Zip drive (/rmdisk/zip0).

$ volrmmount -i zip0
I

Unmount a USB mass storage device.

$ volrmmount -e device-alias

This example shows how to unmount a USB Zip drive (/rmdisk/zip0).

$ volrmmount -e zip0

3. Eject a USB device from a generic USB drive.

$ eject device-alias

For example:
$ eject rmdisk0

Chapter 8 Using USB Devices (Tasks)

137

Note The eject command also unmounts the device if the device is not

unmounted already. The command also terminates any active applications that access the device.

How to Mount or Unmount a USB Mass Storage Device Without vold Running
1. See How to Prepare to Use USB Mass Storage Devices Without vold Running on page 134 for information on disabling vold. 2. Become superuser. 3. (Optional) Identify the diskette device. For example:
# cd /dev/rdsk # ls -l c*0 | grep usb lrwxrwxrwx 1 root root 55 Mar 5 10:35 c2t0d0s0 -> ../../devices/pci@1f,0/usb@c,3/storage@3/disk@0,0:a,raw

Steps

In this example, the diskette device is c2t0d0s0. 4. Select one of the following to mount or unmount a USB mass storage device:
I

Mount a USB mass storage device.

# mount [ -F fstype ] block-device mount-point

This example shows how to mount a device with a UFS le system:

# mount /dev/dsk/c1t0d0s2 /mnt

This example shows how to mount a device with a PCFS le system:

# mount -F pcfs /dev/dsk/c1t0d0s0:c /mnt

This example shows how to mount a CD with a read-only HSFS le system:

# mount -F hsfs -o ro /dev/dsk/c1t0d0s2 /mnt
I

Unmount a USB mass storage device. First, be sure no one is using the le system on the device. For example:
# fuser -c -u /mnt # umount /mnt

5. Eject the device.

# eject /dev/[r]dsk/cntndnsn 138 System Administration Guide: Devices and File Systems June 2005

For example:
# eject /dev/rdsk/c1t0d0s2

Troubleshooting Tips for USB Mass Storage Devices

Keep the following tips in mind if you have problems adding or removing a USB mass storage device.
I

If USB devices are added or removed when the system is down, you must perform a reconguration boot.
ok boot -r

If you have problems accessing a device that was connected while the system is running, try the following command:
# devfsadm
I

Do not move devices around if the system has been powered down by a suspend operation. For more information, see SPARC: USB Power Management on page 123. If a device has been hot removed while in use by applications and is no longer available, then stop the applications. Use the prtconf command to see whether the device node has been removed.

Disabling Specic USB Drivers

You can disable specic types of USB devices by disabling their client driver. For example, USB printers can be disabled by disabling the usbprn driver that directs them. Disabling usbprn does not affect other kinds of devices, such as USB storage devices. The following table identies some USB device types and their corresponding drivers.

Device Type

Driver to Disable

Audio HID (usually keyboard and mouse) Storage Printer

usb_ac and usb_as hid scsa2usb usbprn

Chapter 8 Using USB Devices (Tasks)

139

Device Type

Driver to Disable

Serial

usbser_edge

If you disable a driver for a USB device that is still connected to the system, you see a console message similar to the following:
usba10: WARNING: usba: no driver found for device name

M
Steps

How to Disable Specic USB Drivers

1. Become superuser. 2. Record the driver aliases that you are about to remove.
# cp /etc/driver_aliases /etc/driver_aliases.orig

3. Identify the specic USB driver alias name. For example:

# grep usbprn /etc/driver_aliases usbprn "usbif,class7.1.1" usbprn "usbif,class7.1.2"

4. Remove the driver alias entry. For example:

# update_drv -d -i "usbif,class7.1.1" usbprn # update_drv -d -i "usbif,class7.1.2" usbprn

5. Reboot the system.

# init 6

How to Remove Unused USB Device Links

Use this procedure if a USB device is removed while the system is powered off. Removing the USB device while the system is powered off can leave device links for devices that do not exist.

Steps

1. Become superuser. 2. Close all applications that might be accessing the device. 3. Remove the unused links for a specic USB class. For example:
# devfsadm -C -c audio

140

System Administration Guide: Devices and File Systems June 2005

Or, just remove the dangling links:

# devfsadm -C

Using USB Audio Devices (Task Map)

Task Description For Instructions

Add USB audio devices. Identify your systems primary audio device. Change the primary USB audio device.

Add a USB microphone and speakers.

How to Add USB Audio Devices on page 143

Identify which audio device is How to Identify Your your primary audio device. Systems Primary Audio Device on page 143 You might want to make one How to Change the Primary audio device the primary USB Audio Device on page audio device if you remove or 144 change your USB audio devices. How to Remove Unused If you remove a USB audio device while the system is USB Device Links on page powered off, the /dev/audio 140 device might be pointing to a /dev/sound/* device that doesnt exist. Use this section if no sound Troubleshooting USB Audio comes from the USB speakers. Device Problems on page 144

Remove unused USB device links.

Solve USB audio problems.

Using USB Audio Devices

For information about USB audio support in specic Solaris releases, see Solaris Support for USB Devices on page 114. This Solaris release provides USB audio support that is implemented by a pair of cooperating drivers, usb_ac and usb_as. The audio control driver, usb_ac, is a Solaris USB Architecture compliant client driver that provides the controlling interface to user applications. The audio streaming driver, usb_as, processes audio data messages during play and record. It sets sample frequency and precision, and encodes requests from the usb_ac driver. Both drivers comply with the USB audio class 1.0 specication.
Chapter 8 Using USB Devices (Tasks) 141

Some audio devices can set volume under software control. A STREAMS module, usb_ah, is pushed on top of the HID driver for managing this function. Solaris supports USB audio devices that are play-only, record-only, or record and play. Hot-plugging of USB audio devices is supported.
I

USB audio devices are supported on SPARC Ultra and x86 platforms that have USB connectors. USB audio devices that are supported in the Solaris 8 10/01, Solaris 8 2/02, or Solaris 9 releases must support a xed 44100 or 48000 Hz sampling frequency to play or record. The 44100 Hz or 48000 Hz sampling frequency is no longer required in the Solaris 10 release. For fully supported audio data format information, see usb_ac(7D).

The primary audio device is /dev/audio. You can verify that /dev/audio is pointing to USB audio by using the following command:
% mixerctl Device /dev/audioctl: Name = USB Audio Version = 1.0 Config = external Audio mixer for /dev/audioctl is enabled

After you connect your USB audio devices, you access them with the audioplay and audiorecord command through the /dev/sound/N device links. Note that the /dev/audio and /dev/sound/N devices can refer to speakers, microphones, or combination devices. If you refer to the incorrect device type, the command fails. For example, the audioplay command fails if you try to use it with a microphone. You can select a specic default audio device for most Sun audio applications, such as audioplay and audiorecord, by setting the AUDIODEV shell variable or by specifying the -d option for these commands. However, setting AUDIODEV does not work for third-party applications that have /dev/audio hardcoded as the audio le. When you plug in a USB audio device, it automatically becomes the primary audio device, /dev/audio, unless /dev/audio is in use. For instructions on changing /dev/audio from on-board audio to USB audio and vice versa, refer to How to Change the Primary USB Audio Device on page 144, and usb_ac(7D).

Hot-Plugging Multiple USB Audio Devices

If a USB audio device is plugged into a system, it becomes the primary audio device, /dev/audio. It remains the primary audio device even after the system is rebooted. If additional USB audio devices are plugged in, the last one becomes the primary audio device.
142 System Administration Guide: Devices and File Systems June 2005

For additional information on troubleshooting USB audio device problems, see usb_ac(7D).

M
Steps

How to Add USB Audio Devices

1. Plug in the USB speaker. The primary audio device, /dev/audio, points to the USB speaker.
% ls -l /dev/audio lrwxrwxrwx 1 root root 10 Feb 13 08:46 /dev/audio -> usb/audio0

2. (Optional) Remove the speaker. Then, plug it back in. If you remove the speaker, the /dev/audio device reverts back to on-board audio.
% ls -l /dev/audio lrwxrwxrwx 1 root root 7 Feb 13 08:47 /dev/audio -> sound/0

3. Add a USB microphone.

% ls -l /dev/audio lrwxrwxrwx 1 root root 10 Feb 13 08:54 /dev/audio -> usb/audio1

How to Identify Your Systems Primary Audio Device

This procedure assumes that you have already connected the USB audio devices.
G Examine your systems new audio links.
I

Step

Display your systems new audio links with the ls command. For example:
% ls -lt /dev/audio* lrwxrwxrwx 1 root root 7 Jul 23 15:46 /dev/audio -> usb/audio0 lrwxrwxrwx 1 root root 10 Jul 23 15:46 /dev/audioctl -> usb/audioctl0/ % ls -lt /dev/sound/* lrwxrwxrwx 1 root root 74 Jul 23 15:46 /dev/sound/1 -> ../../devices/pci@1f,4000/usb@5/hub@1/device@3/sound-control@0:... lrwxrwxrwx 1 root root 77 Jul 23 15:46 /dev/sound/1ctl -> ../../devices/pci@1f,4000/usb@5/hub@1/device@3/sound-control@0:... lrwxrwxrwx 1 root other 66 Jul 23 14:21 /dev/sound/0 -> ../../devices/pci@1f,4000/ebus@1/SUNW,CS4231@14,200000:sound,audio lrwxrwxrwx 1 root other 69 Jul 23 14:21 /dev/sound/0ctl -> ../../devices/pci@1f,4000/ebus@1/SUNW,CS4231@14,200000:sound,audioctl % Chapter 8 Using USB Devices (Tasks) 143

Notice that the primary audio device, /dev/audio, is pointing to the newly plugged in USB audio device, /dev/usb/audio0.
I

You can also examine your systems USB audio devices with the prtconf command and look for the USB device information.
% prtconf . . . usb, instance #0 hub, instance #0 mouse, instance #0 keyboard, instance #1 device, instance #0 sound-control, instance #0 sound, instance #0 input, instance #0 . . .

M
Step

How to Change the Primary USB Audio Device

G Select one of the following to change the primary USB audio device.
I

If you want the on-board audio device to become the primary audio device, remove the USB audio devices. The /dev/audio link then points to the /dev/sound/0 entry. If the /dev/sound/0 entry is not the primary audio device, then either shut down the system and use the boot -r command, or run the devfsadm -i command as root. If you want the USB audio device to become primary audio device, just plug it in and check the device links.

Troubleshooting USB Audio Device Problems

Sometimes, USB speakers do not produce any sound, even though the driver is attached and the volume is set to high. Hot-plugging the device might not change this behavior. The workaround is to power cycle the USB speakers.

Key Points of Audio Device Ownership

Keep the following key points of audio device ownership in mind when working with audio devices:
144 System Administration Guide: Devices and File Systems June 2005

When you plug in a USB audio device and you are logged in on the console, the console is the owner of the /dev/* entries. This situation means you can use the audio device as long as you are logged in to the console. If you are not logged in to the console when you plug in a USB audio device, root becomes the owner of the device. However, if you log in to the console and attempt to access the USB audio device, device ownership changes to the console. For more information, see logindevperm(4). When you remotely log in with the rlogin command and attempt to access the USB audio device, the ownership does not change. This means that, for example, unauthorized users cannot listen to conversations over a microphone owned by someone else.

Hot-Plugging USB Devices With the cfgadm Command (Task Map)

Task Description For Instructions

Display USB bus information.

Display information about USB devices and buses. Logically uncongure a USB device that is still physically connected to the system. Congure a USB device that was previously uncongured.

How to Display USB Bus Information (cfgadm) on page 147 How to Uncongure a USB Device on page 148 How to Congure a USB Device on page 148

Uncongure a USB device.

Congure a USB device. Logically disconnect a USB device. Logically connect a USB device.

You can logically disconnect a How to Logically Disconnect USB device if you are not a USB Device on page 149 physically near the system. Logically connect a USB device that was previously logically disconnected or uncongured. How to Logically Connect a USB Device on page 149

Disconnect a USB device subtree.

How to Logically Disconnect Disconnect a USB device subtree, which is the hierarchy a USB Device Subtree (or tree) of devices below a on page 150 hub.

Chapter 8 Using USB Devices (Tasks)

145

Task

Description

For Instructions

Reset a USB device.

Reset a USB device to How to Reset a USB Device logically remove and re-create on page 150 the device. Change the default conguration of a multi-conguration USB device. How to Change the Default Conguration of a Multi-Conguration USB Device on page 150

Change the default conguration of a multi-conguration USB device.

Hot-Plugging USB Devices With the cfgadm Command

You can add and remove a USB device from a running system without using the cfgadm command. However, a USB device can also be logically hot-plugged without physically removing the device. This scenario is convenient when you are working remotely and you need to disable or reset a non functioning USB device. The cfgadm command also provides a way to display the USB device tree, including manufacturer and product information. The cfgadm command displays information about attachment points, which are locations in the system where dynamic reconguration operations can occur. An attachment point consists of the following:
I

An occupant, which represents a hardware resource, such as a USB device, that might be congured into the system A receptacle, which is the location that accepts the occupant, such as a USB port

Attachment points are represented by logical and physical attachment point IDs (Ap_Ids). The physical Ap_Id is the physical path name of the attachment point. The logical Ap_Id is a user-friendly alternative for the physical Ap_Id. For more information on Ap_Ids, see cfgadm_usb(1M). The cfgadm command provides the following USB device status information.

Receptacle State

Description

empty/unconfigured disconnected/unconfigured

The device is not physically connected. The device is logically disconnected and unavailable, even though the device could still be physically connected. The device is logically connected, but unavailable. The device is visible in prtconf output.

connected/unconfigured

146

System Administration Guide: Devices and File Systems June 2005

Receptacle State

Description

connected/configured

The device is connected and available.

The following sections describe how to hot-plug a USB device through the software with the cfgadm command. All of the sample USB device information in these sections has been truncated to focus on relevant information.

How to Display USB Bus Information (cfgadm)

For examples of using the prtconf command to display USB conguration information, see How to Display USB Device Information (prtconf) on page 135.

Steps

1. Display USB bus information. For example:

% cfgadm Ap_Id usb0/4.5 usb0/4.5.1 usb0/4.5.2 usb0/4.5.3 usb0/4.5.4 usb0/4.5.5 usb0/4.5.6 usb0/4.5.7 usb0/4.6 usb0/4.7 Type usb-hub usb-device usb-printer usb-mouse usb-device usb-storage usb-communi unknown usb-storage usb-storage Receptacle connected connected connected connected connected connected connected empty connected connected Occupant configured configured configured configured configured configured configured unconfigured configured configured Condition ok ok ok ok ok ok ok ok ok ok

In the preceding example, usb0/4.5.1 identies a device connected to port 1 of the second-level external hub, which is connected to port 5 of rst-level external hub, which is connected to the rst USB controllers root hub, port 4. 2. Display specic USB device information. For example:
% cfgadm -l -s "cols=ap_id:info" Ap_Id Information usb0/4.5.1 Mfg: Inside Out Networks Product: Edgeport/421 NConfigs: 1 Config: 0 : ... usb0/4.5.2 Mfg: <undef> Product: <undef> NConfigs: 1 Config: 0 ... usb0/4.5.3 Mfg: Mitsumi Product: Apple USB Mouse NConfigs: 1 Config: 0 ... usb0/4.5.4 Mfg: NMB Product: NMB USB KB/PS2 M NConfigs: 1 Config: 0 usb0/4.5.5 Mfg: Hagiwara Sys-Com Product: SmartMedia R/W NConfigs: 1 Config: 0 : ... usb0/4.5.6 Mfg: 3Com Inc. Product: U.S.Robotics 56000 Voice USB Modem NConfigs: 2 ... usb0/4.5.7 usb0/4.6 Mfg: Iomega Product: USB Zip 250 NConfigs: 1 Config: 0 Chapter 8 Using USB Devices (Tasks) 147

: Default usb0/4.7 : Default

Mfg: Iomega

Product: USB Zip 100

NConfigs: 1

Config: 0

How to Uncongure a USB Device

You can uncongure a USB device that is still physically connected to the system. However, a driver will never attach to the device. Note that a USB device remains in the prtconf output even after that device is uncongured.

Steps

1. Become superuser. 2. Uncongure the USB device. For example:

# cfgadm -c unconfigure usb0/4.7 Unconfigure the device: /devices/pci@8,700000/usb@5,3/hub@4:4.7 This operation will suspend activity on the USB device Continue (yes/no)? y

3. Verify that the device is uncongured. For example:

# cfgadm Ap_Id usb0/4.5 usb0/4.5.1 usb0/4.5.2 usb0/4.5.3 usb0/4.5.4 usb0/4.5.5 usb0/4.5.6 usb0/4.5.7 usb0/4.6 usb0/4.7 Type usb-hub usb-device usb-printer usb-mouse usb-device usb-storage usb-communi unknown usb-storage usb-storage Receptacle connected connected connected connected connected connected connected empty connected connected Occupant configured configured configured configured configured configured configured unconfigured configured unconfigured Condition ok ok ok ok ok ok ok ok ok ok

M
Steps

How to Congure a USB Device

1. Become superuser. 2. Congure a USB device. For example:
# cfgadm -c configure usb0/4.7

3. Verify that the USB device is congured.

148 System Administration Guide: Devices and File Systems June 2005

For example:
# cfgadm usb0/4.7 Ap_Id usb0/4.7 Type usb-storage Receptacle connected Occupant configured Condition ok

How to Logically Disconnect a USB Device

If you want to remove a USB device from the system and the prtconf output, but you are not physically near the system, just logically disconnect the USB device. The device is still physically connected. However, the device is logically disconnected, unusable, and not visible to the system.

Steps

1. Become superuser. 2. Disconnect a USB device. For example:

# cfgadm -c disconnect -y usb0/4.7

3. Verify that the device is disconnected. For example:

# cfgadm usb0/4.7 Ap_Id usb0/4.7 Type unknown Receptacle disconnected Occupant unconfigured Condition ok

How to Logically Connect a USB Device

Use this procedure to logically connect a USB device that was previously logically disconnected or uncongured.

Steps

1. Become superuser. 2. Connect a USB device. For example:

# cfgadm -c configure usb0/4.7

3. Verify that the device is connected. For example:

# cfgadm usb0/4.7 Ap_Id usb0/4.7 Type usb-storage Receptacle connected Occupant configured Condition ok

The device is now available and visible to the system.

Chapter 8 Using USB Devices (Tasks) 149

How to Logically Disconnect a USB Device Subtree

Use this procedure to disconnect a USB device subtree, which is the hierarchy (or tree) of devices below a hub.

Steps

1. Become superuser. 2. Remove a USB device subtree. For example:

# cfgadm -c disconnect -y usb0/4

3. Verify that the USB device subtree is disconnected. For example:

# cfgadm usb0/4 Ap_Id usb0/4 Type unknown Receptacle Occupant Condition disconnected unconfigured ok

How to Reset a USB Device

If a USB device behaves erratically, use the cfgadm command to reset the device, which logically removes and re-creates the device.

Steps

1. Become superuser. 2. Make sure that the device is not in use. 3. Reset the device. For example:
# cfgadm -x usb_reset -y usb0/4.7

4. Verify that the device is connected. For example:

# cfgadm usb0/4.7 Ap_Id usb0/4.7 Type usb-storage Receptacle connected Occupant configured Condition ok

How to Change the Default Conguration of a Multi-Conguration USB Device

Keep the following in mind when working with multi-conguration USB devices:
I

A USB device conguration denes how a device presents itself to the operating system. This method is different from system device congurations discussed in other cfgadm sections.

150

System Administration Guide: Devices and File Systems June 2005

Some USB devices support multiple congurations, but only one conguration can be active at a time. Multi-conguration devices can be identied by examining the cfgadm -lv output. Nconfigs will be greater than 1. The default USB conguration is conguration 1. The current conguration is reected in cfgadm -lv output as Config. Changes to the default conguration persist across reboots, hot-removes, and the reconguration of the device, as long as the device is reconnected to the same port.

Steps

1. Make sure that the device is not in use. 2. Change the default USB conguration. For example:
# cfgadm -x usb_config -o config=2 usb0/4 Setting the device: /devices/pci@1f,0/usb@c,3:4 to USB configuration 2 This operation will suspend activity on the USB device Continue (yes/no)? yes

3. Verify that the device changed. For example:

# cfgadm -lv usb0/4 Ap_Id Receptacle Occupant Condition Information When Type Busy Phys_Id usb0/4 connected unconfigured ok Mfg: Sun 2000 Product: USB-B0B0 aka Robotech With 6 EPPS High Clk Mode NConfigs: 7 Config: 2 : EVAL Board Setup unavailable usb-device n /devices/pci@1f,0/usb@c,3:4

Note that Config: now shows 2.

Chapter 8 Using USB Devices (Tasks)

151

152

System Administration Guide: Devices and File Systems June 2005

CHAPTER

Using InniBand Devices (Overview/Tasks)

InniBand (IB) is a new I/O technology based on switch fabrics introduced in the Solaris 10 release. It provides high bandwidth, low latency interconnect for attaching I/O devices to hosts and for host-to-host communication. This is a list of the overview information in this chapter.
I I

Overview of InniBand Devices on page 153 Dynamically Reconguring IB Devices (cfgadm) on page 156

For information on the procedures associated with using IB devices, see the following:
I I

Dynamically Reconguring IB Devices (Task Map) on page 155 Using the uDAPL Application Interface With InniBand Devices on page 165

For general information about dynamic reconguration and hot-plugging, see Chapter 6.

Overview of InniBand Devices

IB devices are managed by the Solaris IB nexus driver. This driver supports 5 types of devices:
I I I I I

IB Port devices IB virtual physical point of attachment (VPPA) devices IB HCA service (HCA_SVC) devices Pseudo devices I/O controller (IOC) devices

The IB nexus driver queries the Solaris IB Device Manager (IBDM) for services, referred in this guide as communication services, to enumerate the IB Port, HCA_SVC, and IB VPPA devices.
153

The Port devices bind a communication service to a given port# of a Host Channel Adapter (HCA). The VPPA devices bind a communication service to a port#, p_key# combination instead. The HCA_SVC devices bind a communication service to a given HCA. Note that the Port devices and the HCA_SVC devices always use a p_key (partition key) whose value is zero. The Port, HCA_SVC, and VPPA devices are children of the HCA and are enumerated through the ib.conf le. For more information, see ib(7D). The IOC devices are children of the IB nexus driver and are part of an I/O unit. The pseudo devices are also children of the IB nexus driver and refer to all other devices that provide their own conguration les to enumerate. For more information, see ib(4). The possible IB device tree path name(s) are listed in the following table.
IOC device IB pseudo device IB VPPA device

/ib/ioc@1730000007F510C,173000007F50 /ib/<driver>@<unit-address> /pci@1f,2000/pci@1/pci15b3,5a44@0/ibport@<port#>, <p_key>,<service>

IB HCA_SVC device /pci@1f,2000/pci@1/pci15bc,5a44@0/ibport@0,0,<service> IB Port device HCA /pci@1f,2000/pci@1/pci15b3,5a44@0/ibport@<port#>,0, <service> /pci@1f,2000/pci@1/pci15b3,5a44@0

Note that the IB HCA_SVC devices have zero as the port# and the p_key. The IB components in the preceding table are described as follows: <services> <p_key> <port> <unit-address> Is a communication service. For example, ipib is the communication service used by the ibd kernel client driver. Is the partition key value being used. Is the port number. Refers to IB kernel client drivers property by this name specied in its driver.conf le. For more information, see driver.conf(4).

154

System Administration Guide: Devices and File Systems June 2005

Dynamically Reconguring IB Devices (Task Map)

Task Description For Instructions

Display IB device information. Display information about the How to Display IB Device IB devices on your system. Information on page 157 Congure or uncongure an IOC device. Select one of the following: Uncongure an IOC device. Congure an IOC device. Congure or uncongure a port or VPPA device. Select one of the following: Uncongure a port or a VPPA How to Uncongure an IB device. Port, HCA_SVC, or a VPPA Device on page 159 Congure a port or a VPPA device. Congure or uncongure an IB pseudo device. Select one of the following: Uncongure an IB pseudo device. Congure an IB pseudo device. Display kernel IB clients of an HCA. You might need to display information about kernel IP clients of an HCA, particularly if youre going to uncongure an HCA. Select one of the following: Uncongure IB devices that are connected to an HCA. How to Uncongure IB Devices Connected to an HCA on page 162 How to Uncongure an IB Pseudo Device on page 161 How to Congure an IB Pseudo Device on page 161 How to Display Kernel IB Clients of an HCA on page 161 How to Congure a IB Port, HCA_SVC, or a VPPA Device on page 160 How to Uncongure an IOC Device on page 159 How to Congure an IOC Device on page 159

Congure or uncongure an IB HCA.

Chapter 9 Using InniBand Devices (Overview/Tasks)

155

Task

Description

For Instructions

Congure IB devices that are connected to an HCA. Update the IB p_key tables.

Conguring an IB HCA on page 163

If the p_key table information How to Update the IB p_key of a HCA port(s) changes, Tables on page 163 IBTF and IBDM need to be notied so that their internal p_key databases are updated. Display the IB communication How to Display IB services that are currently in Communication Services use by the IBTF. on page 163 Select one of the following: Add a VPPA communication service. Remove a VPPA communication service. How to Add a VPPA Communication Service on page 164 How to Remove an Existing IB Port, HCA_SVC, or a VPPA Communication Service on page 164

Display IB communication services Add or remove a VPPA communication service.

Update an IOC conguration.

You can update the properties How to Update an IOC of all the IOC device nodes or Conguration on page 165 update a particular IOC Ap_Id.

Dynamically Reconguring IB Devices (cfgadm)

One can congure or uncongure an IB device from a running system by using the cfgadm CLI only. This command also provides a way to display the IB fabric, manage communication services, and update p_key table database(s). For more information, see cfgadm_ib(1M). The cfgadm CLI manages dynamic reconguration, referred to in this guide as DR, of the entire IB fabric as seen by a host. The cfgadm operations are supported on all the IB devices, such as Port, VPPA, HCA_SVC, IOC, and pseudo devices. The cfgadm command displays information about attachment points (Ap_Ids), which are locations in the system where DR operations can occur. For details on the Ap_Ids that cfgadm supports, see cfgadm_ib.1M. Note that all IB Ap_Ids are shown as connected.
156 System Administration Guide: Devices and File Systems June 2005

The cfgadm command provides the following IB device status information.

Receptacle State

Description

connected/configured/ok

The device is connected and available. The devinfo node is present. The device is unavailable and no devinfo node or device driver exists for this device. Or, the device was never congured for use by ib nexus driver. The device might be known to the IB Device Manager.

connected/unconfigured/unknown

The following sections describe how to dynamically recongure (DR) IB devices with the cfgadm command. All of the sample IB device information in these sections has been truncated to focus on relevant information.

How to Display IB Device Information

You can use the prtconf command to display general information about IB devices. For example:
$ prtconf pci, instance #0 pci15b3,5a44, instance #0 ibport, instance #253 ibport, instance #254 ibport, instance #255 . . . ib, instance #0 ioc, instance #243 ioc, instance #244 ioc, instance #245 ioc, instance #246 ioc, instance #247 ioc, instance #248 ibgen, instance #249

In the preceding example, pci15b3,5a44 refers to an IB HCA. Use the following steps to display specic IB device information.
Steps

1. Become superuser. 2. Display IB fabric information.

Chapter 9 Using InniBand Devices (Overview/Tasks)

157

For example:
# cfgadm -a Ap_Id ib hca:1730000008070 ib::1730000007F5198 ib::1730000007F5199 ib::1730000008070,0,hnfs ib::1730000008071,0,sdp ib::1730000008072,0,sdp ib::1730000008071,8001,ipib ib::1730000008072,8001,ipib ib::ibgen,0 # Type IB-Fabric IB-HCA IB-IOC IB-IOC IB-HCA_SVC IB-PORT IB-PORT IB-VPPA IB-VPPA IB-PSEUDO Receptacle connected connected connected connected connected connected connected connected connected connected Occupant Condition configured ok configured ok configured ok configured ok configured ok configured ok configured ok configured ok configured ok configured ok

In the preceding example output, the components are described as follows: Ap_Id ib::1730000008072,0,sdp Identies an IB Port device that is connected to port 2 and is bound to the sdp service. Ap_Id ib::1730000008072,8001,ipib Identies an IB VPPA device that is connected to port 2, using a p_key value of 0x8001, and is bound to the ibd service. Ap_Id ib:: 1730000008070,0,hnfs Identies an IB HCA_SVC device bound to the hnfs service. Ap_Id ib::1730000007F5198 Identies an IOC device. Ap_Id ib::ibgen,0 Identies a pseudo device. 3. Display specic IB device information. For example, for an IB VPPA device:
# cfgadm -al -s "cols=ap_id:info" ib::1730000008072,8001,ipib Ap_Id Information ib::1730000008072,8001,ipib ipib #

For example, for an IB HCA device:

# cfgadm -al -s "cols=ap_id:info" hca::1730000008070 Ap_Id Information hca::1730000008070 VID: 0x15b3, PID: 0x5a44, #ports: 0x2, port1 GUID: 0x1730000008071, port2 GUID: 0x1730000008072 #

The preceding output displays the number of ports and their GUIDs.

158

System Administration Guide: Devices and File Systems June 2005

How to Uncongure an IOC Device

You can uncongure an IB device that is still physically connected to the system, but a driver will never attach to it.

Steps

1. Become superuser. 2. Uncongure the IB device. For example:

# cfgadm -c unconfigure ib::1730000007F5198 Unconfigure the device: /devices/ib:fabric::1730000007F5198 This operation will suspend activity on the IB device Continue (yes/no)? y #

3. Verify that the device is uncongured. For example:

# cfgadm -a ib::1730000007F5198 ib::1730000007F5198 IB-IOC # connected unconfigured unknown

M
Steps

How to Congure an IOC Device

1. Become superuser. 2. Congure a IB device. For example:
# cfgadm -yc configure ib::1730000007F5198

3. Verify that the IB device is congured. For example:

# cfgadm -al ib::1730000007F5198 Ap_Id Type Receptacle ib::1730000007F5198 IB-IOC connected Occupant configured Condition ok

How to Uncongure an IB Port, HCA_SVC, or a VPPA Device

Use the following steps if you want to remove an IB Port, HCA_SVC, or a VPPA device from the system. The example below illustrates how to uncongure a VPPA device, but the same procedure applies to Port and HCA_SVC devices as well.
Chapter 9 Using InniBand Devices (Overview/Tasks) 159

Steps

1. Become superuser. 2. Uncongure the IB VPPA device. For example:

# cfgadm -c unconfigure ib::1730000007F51,8001,ipib Unconfigure the device: /devices/ib:fabric::1730000007F51,8001,ipib This operation will suspend activity on the IB device Continue (yes/no)? Y #

3. Verify that the device is disconnected. For example:

# cfgadm -a ib::1730000007F51,8001,ipib Ap_Id Type Receptacle Occupant Condition ib::1730000007F51,8001,ipib IB-VPPA connected unconfigured unknown #

How to Congure a IB Port, HCA_SVC, or a VPPA Device

Use the following steps if you want to congure an IB Port, HCA_SVC, or a VPPA device on the system. The example below illustrates how to congure a VPPA device, but similar steps can be used to congure Port and HCA_SVC devices as well.

Steps

1. Become superuser. 2. Congure the IB VPPA device. For example:

# cfgadm -c configure ib::1730000007F51,8001,ipib

3. Verify that the device is connected. For example:

# cfgadm -a ib::1730000007F51,8001,ipib Ap_Id Type Receptacle Occupant Condition ib::1730000007F51,8001,ipib IB-VPPA connected configured ok

Note A cfgadm based congure or uncongure operation of IB Port and

HCA_SVC devices is similar to the preceding examples for an IB VPPA device.

160

System Administration Guide: Devices and File Systems June 2005

How to Uncongure an IB Pseudo Device

Use the following steps if you want to remove an IB pseudo device from the system. 1. Become superuser. 2. Uncongure the IB pseudo device. For example:
# cfgadm -c unconfigure ib::ibgen,0 Unconfigure the device: /devices/ib:fabric::ibgen,0 This operation will suspend activity on the IB device Continue (yes/no)? Y #

Steps

3. Verify that the device is disconnected.

# cfgadm -a ib::ibgen,0 Ap_Id Type Receptacle Occupant Condition ib::ibgen,0 IB-PSEUDO connected unconfigured unknown

How to Congure an IB Pseudo Device

Use the following steps to congure an IB pseudo device. 1. Become superuser. 2. Congure the IB pseudo device. For example:
# cfgadm -yc configure ib::ibgen,0

Steps

3. Verify that the device is connected. For example:

# cfgadm -a ib::ibgen,0 Ap_Id Type ib::ibgen,0 IB-PSEUDO Receptacle Occupant Condition connected configured ok

How to Display Kernel IB Clients of an HCA

The following IB cfgadm plugin command can be invoked to list kernel IB clients using this HCA. Note that the last column would show a yes if a kernel IB client uses another HCA. IB Managers and kernel clients that do not use the HCA are shown with an Ap_Id of -.

Step

G Display kernel IB clients of an HCA. Chapter 9 Using InniBand Devices (Overview/Tasks) 161

For example:
$ cfgadm -x list_clients hca:173000007F50 Ap_Id IB Client ib::1730000007F51D0 ibgen ib::1730000007F51D1 ibgen ib::1730000007F51,8001,ipib ibd ib::ibgen,0 ibgen ibdm ibmf nfs/ib $ Alternate HCA no no no no no no no

How to Uncongure IB Devices Connected to an HCA

An actual DR of an HCA is beyond the scope of the IB cfgadm plugin. Although DR of an HCA can be achieved by using the plugin of the underlying bus. For example, a PCI based HCA can use the cfgadm_pci command. For more information, see cfgadm_pci(1M). However, the IB cfgadm plugin assists in the HCA DR by listing its kernel IB clients as illustrated in steps below.

Steps

1. Become superuser. 2. List the kernel IB clients of the HCA. For example:
# cfgadm -x list_clients hca:173000007F50 Ap_Id IB Client ib::1730000007F51D0 ibgen ib::1730000007F51D1 ibgen ib::1730000007F51,8001,ipib ibd ib::ibgen,0 ibgen ibdm ibmf nfs/ib Alternate HCA no no no no no no no

3. Uncongure kernel IB clients, such as Port, VPPA, HCA_SVC, or IOC devices, that do not have alternate HCA(s) present. For example:
# cfgadm -x unconfig_clients hca:1730000008070 Unconfigure Clients of HCA /devices/ib:1730000008070 This operation will unconfigure IB clients of this HCA Continue (yes/no)? y

4. Verify that the kernel IB clients of the HCA are uncongured.

# cfgadm -x list_clients hca:173000007F50 Ap_Id IB Client 162 System Administration Guide: Devices and File Systems June 2005 Alternate HCA

ibdm ibmf nfs/ib

no no no

Conguring an IB HCA
Invoke the bus-specic cfgadm plugin to congure the HCA. The exact details are beyond the scope of this chapter.

How to Update the IB p_key Tables

If the p_key table information of an HCA port(s) changes, for example, additional p_keys are enabled or disabled, InniBand Transport Framework (IBTF) and IBDM need to be notied so that their internal p_key databases are updated. The cfgadm command helps update the p_key databases of IBTF and IBDM. For more information, see ibtl(7D) and ibdm(7D).

Steps

1. Become superuser. 2. Update the p_key tables. For example:

# cfgadm -x update_pkey_tbls -y ib

How to Display IB Communication Services

Use the following steps to display the communication services that are currently in use by the IBTF.

Steps

1. Become superuser. 2. Display IB communication services. For example:

# cfgadm -x list_services ib Port communication services: srp VPPA communication services: ibd HCA_SVC communication services: hnfs

Chapter 9 Using InniBand Devices (Overview/Tasks)

163

How to Add a VPPA Communication Service

Use the following steps to add a new VPPA communication service. Similar steps can be used to add a new HCA_SVC or a port communication service.

Steps

1. Become superuser. 2. Add a new VPPA communication service. For example:

# cfgadm -o comm=vppa,service=new -x add_service ib

3. Verify that the new service has been added. For example:
# cfgadm -x list_services ib Port communication services: srp VPPA communication services: ibd new HCA_SVC communication services: nfs_service #

How to Remove an Existing IB Port, HCA_SVC, or a VPPA Communication Service

Use the following steps to delete an existing IB Port, HCA_SVC, or a VPPA communication service.

Steps

1. Become superuser. 2. Remove a VPPA communication service. For example:

# cfgadm -o comm=vppa,service=new -x delete_service ib

3. Verify that the communication service has been removed. For example:
# cfgadm -x list_services ib Port communication services: srp VPPA communication services: ibd HCA_SVC communication services: 164 System Administration Guide: Devices and File Systems June 2005

hnfs #

How to Update an IOC Conguration

Use the following steps to update properties of all the IOC device nodes or for a particular IOC Ap_Id. The properties that can get updated are as follows:
I I I I

port-list port-entries service-id service-name

For more information on these properties, see ib(7D). Note that these properties may not get updated if there is no conguration change. The following example describes how to update a particular IOCs conguration. If you need to update the conguration of all the IOCs, then specify the static ib Ap_Id instead of the particular IOC Ap_Id.
Steps

1. Become superuser. 2. Update the conguration of an IOC. For example:

# cfgadm -x update_ioc_conf ib::1730000007F5198 This operation can update properties of IOC devices. Continue (yes/no)? y #

3. Verify that the properties have been updated by running prtconf -v.

Using the uDAPL Application Interface With InniBand Devices

User Direct Access Programming Library (uDAPL) is a standard API that promotes data center application data messaging performance, scalability, and reliability over Remote Direct Memory Access (RDMA) capable interconnects such as InniBand. The uDAPL interface is dened by the DAT collaborative. For more information about the DAT collaborative, go to the following site: http://www.datcollaborative.org
Chapter 9 Using InniBand Devices (Overview/Tasks) 165

The Solaris release provides the following uDAPL features:

I I

A standard DAT registry library, libdat. For more information, see libdat(3LIB). A standard service provider registration le, dat.conf. For more information, see dat.conf(4). Support for multiple service providers so that each provider species their own uDAPL library path, version number, etc. in their own service_provider.conf le. For more information, see, service_provider.conf(4). An administrative tool, the datadm command, to congure dat.conf. For more information, see datadm(1M). A new resource control property, project.max-device-locked-memory, to regulate the amount of locked down physical memory. A naming scheme that uses either IPv4 or IPv6 addresses that leverage the IP infrastructure, such as ARP in IPv4 and neighbor discovery in IPv6, for address resolution. The Solaris uDAPL Interface Adapter directly maps to an IPoIB device instance. Support for the standard Address Translation Scheme that is used by the DAT collaborative community. A uDAPL service provider library to support the Mellanox Tavor Host Channel Adapter with automatic registration to the dat.conf registration le. Supports both SPARC platform and x86 platforms.

M
Steps

How to Enable uDAPL

1. Become superuser. 2. Conrm that the following packages are installed. Or, install them, if needed.
I I I I I I I

SUNWib Sun InniBand Framework SUNWtavor Sun Tavor HCA Driver SUNWipoib Sun IP over InniBand SUNWudaplr Direct Access Transport (DAT) registry package (root) SUNWudaplu Direct Access Transport (DAT) registry packages (usr) SUNWudapltr Service Provider for Tavor packages (root) SUNWudapltu Service Provider for Tavor packages (usr)

3. Select one of the following to plumb the IPoIB interfaces.

I

Manually plumb the interfaces with the ifconfig and datadm commands. For example:
# ifconfig ibd1 plumb # ifconfig ibd1 192.168.0.1/24 up # datadm -a /usr/share/dat/SUNWudaplt.conf

166

System Administration Guide: Devices and File Systems June 2005

Automatically plumb the interfaces by doing the following:

I

Create the following le with the appropriate IP address.

/etc/hostname.ibd1

Reboot the system.

Updating the DAT Static Registry

You can use the datadm command to maintain the DAT static registry, the dat.conf le. For more information about this le, see dat.conf(4). The datadm command can also be used to register or unregister a service provider to the dat.conf le. For more information, see datadm(1M). When IPoIB interface adapters are added or removed, run the datadm command to update the dat.conf le to reect the current state of the system. A new set of interface adapters for all the service providers that are currently installed will be regenerated.

M How to Update the DAT Static Registry

Steps

1. Become superuser. 2. Update the DAT static registry after you add or remove IPoIP interface adapters from the system.
# datadm -u

3. Display the updated DAT static registry.

# datadm

M How to Register a Service Provider in the DAT Static

Registry
Steps

1. Become superuser. 2. Update the DAT static registry after you add Suns service provider for the Mellanox Tavor Host Channel Adapter.
# datadm -a /usr/share/dat/SUNWudaplt.conf

3. Display the updated DAT static registry.

# datadm -v Chapter 9 Using InniBand Devices (Overview/Tasks) 167

M How to Unregister a Service Provider from the DAT Static

Registry
Steps

1. Become superuser. 2. Update the DAT static registry after you remove Suns service provider for the Mellanox Tavor Host Channel Adapter from the system.
# datadm -r /usr/share/dat/SUNWudaplt.conf

3. Display the updated DAT static registry.

# datadm -v

168

System Administration Guide: Devices and File Systems June 2005

CHAPTER

10

Accessing Devices (Overview)

This chapter provides information about how to access the devices on a system. This is a list of the overview information in this chapter.
I I I I

Accessing Devices on page 169 Logical Disk Device Names on page 171 Logical Tape Device Names on page 174 Logical Removable Media Device Names on page 174

For overview information about conguring devices, see Chapter 5.

Accessing Devices
You need to know how to specify device names when using commands to manage disks, le systems, and other devices. In most cases, you can use logical device names to represent devices that are connected to the system. Both logical and physical device names are represented on the system by logical and physical device les.

How Device Information Is Created

When a system is booted for the rst time, a device hierarchy is created to represent all the devices connected to the system. The kernel uses the device hierarchy information to associate drivers with their appropriate devices. The kernel also provides a set of pointers to the drivers that perform specic operations. For more information on device hierarchy, see OpenBoot 3.x Command Reference Manual.

169

How Devices Are Managed

The devfsadm command manages the special device les in the /dev and /devices directories. By default, the devfsadm command attempts to load every driver in the system and attach to all possible device instances. Then, devfsadm creates the device les in the /devices directory and the logical links in the /dev directory. In addition to managing the /dev and /devices directories, the devfsadm command also maintains the path_to_inst instance database. For more information, see path_to_inst(4). Both reconguration boot processing and updates to the /dev and /devices directories in response to dynamic reconguration events are handled by devfsadmd, the daemon version of the devfsadm command. This daemon is started from the /etc/rc* scripts when a system is booted. Because the devfsadmd daemon automatically detects device conguration changes generated by any reconguration event, there is no need to run this command interactively. For more information, see devfsadm(1M).

Device Naming Conventions

Devices are referenced in three ways in the Solaris OS.
I

Physical device name Represents the full device path name in the device information hierarchy. The physical device name is created by when the device is rst added to the system. Physical device les are found in the /devices directory. Instance name Represents the kernels abbreviation name for every possible device on the system. For example, sd0 and sd1 represent the instance names of two disk devices. Instance names are mapped in the /etc/path_to_inst le. Logical device name The logical device name is created by when the device is rst added to the system. Logical device names are used with most le system commands to refer to devices. For a list of le commands that use logical device names, see Table 101. Logical device les in the /dev directory are symbolically linked to physical device les in the /devices directory.

The preceding device name information is displayed with the following commands:
I I I I

dmesg format sysdef prtconf

170

System Administration Guide: Devices and File Systems June 2005

Logical Disk Device Names

Logical device names are used to access disk devices when you perform the following tasks:
I I I I

Add a new disk to the system. Move a disk from one system to another system. Access or mount a le system residing on a local disk. Back up a local le system.

Many administration commands take arguments that refer to a disk slice or le system. Refer to a disk device by specifying the subdirectory to which it is symbolically linked, either /dev/dsk or /dev/rdsk, followed by a string identifying the particular controller, disk, and slice.
/dev/[r]dsk/cvtwdx[sy,pz] Slice number (s0 to s7) or fdisk partition number (p0 to p4) Drive number Physical bus target number Logical controller number Raw disk device subdirectory Devices directory
FIGURE 101

Description of Logical Device Names

Specifying the Disk Subdirectory

Disk and le administration commands require the use of either a raw (or character) device interface, or a block device interface. The distinction is made by how data is read from the device. Raw device interfaces transfer only small amounts of data at a time. Block device interfaces include a buffer from which large blocks of data are read at once. Different commands require different interfaces:
I

When a command requires the raw device interface, specify the /dev/rdsk subdirectory. (The r in rdsk stands for raw.) When a command requires the block device interface, specify the /dev/dsk subdirectory.
Chapter 10 Accessing Devices (Overview) 171

When you are not sure whether a command requires use of /dev/dsk or /dev/rdsk, check the man page for that command.

The following table shows which interface is required for some commonly used disk and le system commands.
TABLE 101 Command Reference

Device Interface Type Required by Some Frequently Used Commands

Interface Type Example of Use

df(1M) fsck(1M) mount(1M) newfs(1M) prtvtoc(1M)

Block Raw Block Raw Raw

df /dev/dsk/c0t3d0s6 fsck -p /dev/rdsk/c0t0d0s0 mount /dev/dsk/c1t0d0s7 /export/home newfs /dev/rdsk/c0t0d1s1 prtvtoc /dev/rdsk/c0t0d0s2

Direct and Bus-Oriented Controllers

You might access disk partitions or slices differently depending upon whether the disk device is connected to a direct or bus-oriented controller. Generally, direct controllers do not include a target identier in the logical device name. The conventions for both types of controllers are explained in the following subsections.
Note Controller numbers are assigned automatically during system initialization. The numbers are strictly logical and imply no direct mapping to physical controllers.

x86: Disks With Direct Controllers

To specify a slice on a disk with an IDE controller on an x86 based system, follow the naming convention shown in the following gure.

172

System Administration Guide: Devices and File Systems June 2005

cwdx [sy, pz] Slice number (s0 to s7) or fdisk partition number (p0 to p4) Drive number Logical controller number
FIGURE 102

x86: Disks With Direct Controllers

To indicate the entire Solaris fdisk partition, specify slice 2 (s2). If you have only one controller on your system, w is usually 0.

Disks With Bus-Oriented Controllers

To specify a slice on a disk with a bus-oriented controller, SCSI for instance, follow the naming convention shown in the following gure.
cvtwdx[sy,pz] Slice number (s0 to s7) or fdisk partition number (p0 to p4) Drive number Physical bus target number Logical controller number
FIGURE 103

Disks With Bus-Oriented Controllers

On a SPARC based system with directly connected disks such as the IDE disks on an UltraSPARC system, the naming convention is the same as that for systems with bus-oriented controllers. If you have only one controller on your system, w is usually 0. For SCSI controllers, x is the target address set by the switch on the back of the unit, and y is the logical unit number (LUN) of the drive attached to the target. If the disk has an embedded controller, y is usually 0. For more information about SCSI addressing on SPARC based systems, see the SunSolveSM Info Doc 48041 and scsi_address(9S). To indicate the whole disk, specify slice 2 (s2).

Chapter 10 Accessing Devices (Overview)

173

Logical Tape Device Names

Logical tape device les are found in the /dev/rmt/* directory as symbolic links from the /devices directory.
/dev/rmt/xy Optional density l low m medium h high u ultra c compressed Drive number (0-n) Raw magnetic tape device directory Devices directory
FIGURE 104

Logical Tape Device Names

The rst tape device connected to the system is 0 (/dev/rmt/0). Tape density values (l, m, h, c, and u) are described in Chapter 30.

Logical Removable Media Device Names

Since removable media is managed by volume management (vold), the logical device name is usually not used unless you want to mount the media manually. The logical device name that represents the removable media devices on a system are described in Chapter 2.

174

System Administration Guide: Devices and File Systems June 2005

CHAPTER

11

Managing Disks (Overview)

This chapter provides overview information about Solaris disk slices and introduces the format utility. This is a list of overview information in this chapter.
I I I I I I I I

Whats New in Disk Management? on page 175 Where to Find Disk Management Tasks on page 180 Overview of Disk Management on page 181 Disk Terminology on page 181 About Disk Slices on page 182 format Utility on page 185 About Disk Labels on page 189 Partitioning a Disk on page 191

For instructions on how to add a disk to your system, see Chapter 13 or Chapter 14.

Whats New in Disk Management?

This section describes new disk management features in the Solaris 10 release.

Multiterabyte Disk Support With EFI Disk Label

Solaris 10 Provides support for disks that are larger than 1 terabyte on systems that run a 64-bit Solaris kernel. The Extensible Firmware Interface GUID Partition Table (EFI GPT) disk label is also available for disks less than 1 terabyte that are connected to a system that runs a 32-bit Solaris kernel. You can download the EFI specication at:
175

http://www.intel.com/technology/efi/main_specification.htm You can use the format -e command to apply an EFI label to a disk if the system is running the appropriate Solaris release. However, you should review the important information in Restrictions of the EFI Disk Label on page 177 before attempting to apply an EFI label. The EFI label provides support for physical disks and virtual disk volumes. This release also includes updated disk utilities for managing disks greater than 1 terabyte. The UFS le system is compatible with the EFI disk label, and you can create a UFS le system greater than 1 terabyte. For information on creating a multiterabyte UFS le system, see 64-bit: Support of Multiterabyte UFS File Systems on page 281. The unbundled Sun QFS le system is also available if you need to create le systems greater than 1 terabyte. For information on the Sun QFS le system, see http://docs.sun.com/db/doc/816-2542-10. The Solaris Volume Manager software can also be used to manage disks greater than 1 terabyte in this Solaris release. For information on using Solaris Volume Manager, see Solaris Volume Manager Administration Guide. The VTOC label is still available for disks less than 1 terabyte in size. If you are only using disks smaller than 1 terabyte on your systems, managing disks will be the same as in previous Solaris releases. In addition, you can use the format-e command to label a disk less than 1 terabyte with an EFI label. For more information, see Example 126.

Comparison of the EFI Label and the VTOC Label

The EFI disk label differs from the VTOC disk label in the following ways:
I I I

Provides support for disks greater than 1 terabyte in size. Provides usable slices 06, where slice 2 is just another slice. Partitions (or slices) cannot overlap with the primary or backup label, nor with any other partitions. The size of the EFI label is usually 34 sectors, so partitions start at sector 34. This feature means that no partition can start at sector zero (0). No cylinder, head, or sector information is stored in the EFI label. Sizes are reported in blocks. Information that was stored in the alternate cylinders area, the last two cylinders of the disk, is now stored in slice 8. If you use the format utility to change partition sizes, the unassigned partition tag is assigned to partitions with sizes equal to zero. By default, the format utility assigns the usr partition tag to any partition with a size greater than zero. You can use the partition change menu to reassign partition tags after the partitions are changed. However, you cannot change a partition with a non-zero size to the unassigned partition tag.

176

System Administration Guide: Devices and File Systems June 2005

Restrictions of the EFI Disk Label

Keep the following restrictions in mind when determining whether using disks greater than 1 terabyte is appropriate for your environment:
I

The SCSI driver, ssd or sd, currently supports only up to 2 terabytes. If you need greater disk capacity than 2 terabytes, use a disk and storage management product such as Solaris Volume Manager to create a larger device. Layered software products intended for systems with EFI-labeled disks might be incapable of accessing a disk with an EFI disk label. A disk with an EFI label is not recognized on systems running previous Solaris releases. The EFI disk label is not supported on IDE disks. You cannot boot from a disk with an EFI disk label. You cannot use the fdisk command on a disk with an EFI label. You cannot use the Solaris Management Consoles Disk Manager tool to manage disks with EFI labels. Use the format utility to partition disks with EFI labels. Then, you can use the Solaris Management Consoles Enhanced Storage Tool to manage volumes and disk sets with EFI-labeled disks. The EFI specication prohibits overlapping slices. The entire disk is represented by cxtydz. The EFI disk label provides information about disk or partition sizes in sectors and blocks, but not in cylinders and heads. The following format options are either not supported or are not applicable on disks with EFI labels:
I

I I I I

The save option is not supported because disks with EFI labels do not need an entry in the format.dat le. The backup option is not applicable because the disk driver nds the primary label and writes it back to the disk.

Support for EFI-Labeled Disks on x86 Systems

Solaris support for the EFI disk label is available on x86 systems. Use the following command to add an EFI label on an x86 system:
# > > > > > format -e [0] SMI Label [1] EFI Label Specify Label type[0]: 1 WARNING: converting this device to EFI labels will erase all current fdisk partition information. Continue? yes

Previous label information is not converted to the EFI disk label.

Chapter 11 Managing Disks (Overview)

177

You will have to recreate the labels partition information manually with the format command. You cannot use the fdisk command on a disk with an EFI label. The fdisk command is not intended for disks that are larger than 1 terabyte. For more information about EFI disk labels, see the preceding section.

Installing a System With an EFI-Labeled Disk

The Solaris installation utilities automatically recognize disks with EFI labels. However, you cannot use the Solaris installation program to repartition these disks. You must use the format utility to repartition an EFI-labeled disk before or after installation. The Solaris Upgrade and Live Upgrade utilities also recognize a disk with an EFI label. However, you cannot boot a system from an EFI-labeled disk. After the Solaris release is installed on a system with an EFI-labeled disk, the partition table appears similar to the following:
Current partition table (original): Total disk sectors available: 2576924638 + 16384 (reserved sectors) Part Tag 0 root 1 unassigned 2 unassigned 3 unassigned 4 unassigned 5 unassigned 6 unassigned 8 reserved Flag wm wm wm wm wm wm wm wm First Sector 34 0 0 0 0 0 0 2576924638 Size 1.20TB 0 0 0 0 0 0 8.00MB Last Sector 2576924636 0 0 0 0 0 0 2576941021

Managing Disks With EFI Disks Labels

Use the following table to locate information on managing disks with EFI disk labels.

Task

For More Information

If the system is already installed, connect the disk to the system and perform a reconguration boot. Repartition the disk by using the format utility, if necessary.

SPARC: Adding a System Disk or a Secondary Disk (Task Map) on page 217 SPARC: How to Create Disk Slices and Label a Disk on page 220

Create disk volumes, and if needed, create soft Chapter 2, Storage Management Concepts, partitions by using Solaris Volume Manager. in Solaris Volume Manager Administration Guide Create UFS le systems for the new disk by using the newfs command. SPARC: How to Create a UFS File System on page 225

178

System Administration Guide: Devices and File Systems June 2005

Task

For More Information

Or, create a QFS le system. Clone a disk with an EFI label

http://docs.sun.com/db/coll/20445.2 Example 292

Troubleshooting Problems With EFI Disk Labels

Use the following error messages and solutions to troubleshoot problems with EFI-labeled disks. Error Message
The capacity of this LUN is too large. Reconfigure this LUN so that it is < 2TB.

Cause You attempted to create a partition on a SCSI device that is larger than 2 terabytes. Solution Create a partition on a SCSI device that is less than 2 terabytes. Error Message
Dec 3 09:26:48 holoship scsi: WARNING: /sbus@a,0/SUNW,socal@d,10000/ sf@1,0/ssd@w50020f23000002a4,0 (ssd1): Dec 3 09:26:48 holoship disk has 2576941056 blocks, which is too large for a 32-bit kernel

Cause You attempted to boot a system running a 32-bit SPARC or x86 kernel with a disk greater than 1 terabyte. Solution Boot a system running a 64-bit SPARC or x86 kernel with a disk greater than 1 terabyte. Error Message
Dec 3 09:12:17 holoship scsi: WARNING: /sbus@a,0/SUNW,socal@d,10000/ sf@1,0/ssd@w50020f23000002a4,0 (ssd1): Dec 3 09:12:17 holoship corrupt label - wrong magic number

Cause You attempted to add a disk to a system running an older Solaris release. Solution Add the disk to a system running the Solaris release that supports the EFI disk label.

Chapter 11 Managing Disks (Overview)

179

Common SCSI Drivers for SPARC and x86 Systems

In this Solaris release, the disk drivers for the SPARC and the x86 platforms are merged into a single driver. This change creates one source le for the following 3 drivers:
I I I

SPARC sd for SCSI devices x86 sd for Fibre Channel and SCSI devices SPARC ssd for Fibre Channel devices

In previous Solaris releases, 3 separate drivers were needed to provide support of SCSI and Fibre Channel disk devices on the SPARC and x86 platforms. All of the disk utilities, such as the format, fmthard, and fdisk commands, have been updated to support these changes. For more information, see sd.7D and ssd.7D.

New fdisk Partition Identier

The Solaris fdisk partition identier on x86 systems has been changed from 130 (0x82) to 191 (0xbf). All Solaris commands, utilities, and drivers have been updated to work with either fdisk identier. There is no change in fdisk functionality. A new fdisk menu option enables you to switch back and forth between the new and old identier. The fdisk identier can be changed even when the le system that is contained in the partition is mounted. Two type values in the fdisk menu reect the old and new identiers as follows:
I I

Solaris identies 0x82 Solaris2 identies 0xbf

For step-by-step instructions on changing the Solaris fdisk partition identier, see How to Change the Solaris fdisk Identier on page 230.

Where to Find Disk Management Tasks

Use these references to nd step-by-step instructions for managing disks.
Disk Management Task For More Information

Format a disk and examine a disk label.

Chapter 12

180

System Administration Guide: Devices and File Systems June 2005

Disk Management Task

For More Information

Add a new disk to a SPARC system. Add a new disk to an x86 system. Hot-plug a SCSI or PCI disk.

Chapter 13 Chapter 14 Chapter 6

Overview of Disk Management

Managing disks in the Solaris OS usually involves setting up the system and running the Solaris installation program to create the appropriate disk slices and le systems and to install the Solaris OS. Occasionally, you might need to use the format utility to add a new disk drive or replace a defective disk drive.

Disk Terminology
Before you can effectively use the information in this section, you should be familiar with basic disk architecture. In particular, you should be familiar with the following terms:

Disk Term

Description

Track Cylinder Sector Block Disk controller Disk label Device driver

A concentric ring on a disk that passes under a single stationary disk head as the disk rotates. The set of tracks with the same nominal distance from the axis about which the disk rotates. Section of each disk platter. A sector holds 512 bytes. A data storage area on a disk. A disk block is 512 bytes. A chip and its associated circuitry that controls the disk drive. The rst sector of a disk that contains disk geometry and partition information. A kernel module that controls a hardware or virtual device.

For additional information, see the product information from your disks manufacturer.
Chapter 11 Managing Disks (Overview) 181

About Disk Slices

Files stored on a disk are contained in le systems. Each le system on a disk is assigned to a slice, which is a group of sectors set aside for use by that le system. Each disk slice appears to the Solaris OS (and to the system administrator) as though it were a separate disk drive. For information about le systems, see Chapter 17.
Note Slices are sometimes referred to as partitions. Certain interfaces, such as the format utility, refer to slices as partitions.

When setting up slices, remember these rules:

I I

Each disk slice holds only one le system. No le system can span multiple slices.

Slices are set up slightly differently on SPARC and x86 platforms. The following table summarizes the differences.
TABLE 111

Slice Differences on SPARC and x86 Platforms

x86 Platform

SPARC Platform

The entire disk is devoted to Solaris OS. VTOC The disk is divided into 8 slices, numbered 07. EFI The disk is divided into 7 slices, numbered 06.

Disk is divided into fdisk partitions, one fdisk partition per operating system. VTOC The Solaris fdisk partition is divided into 10 slices, numbered 09. EFI The disk is divided into 7 slices, numbered 06

Solaris Volume Manager, previously the Solstice DiskSuite, has a partitioning feature, soft partitions. Soft partitions enable more than eight partitions per disk. For general information about Solaris Volume Manager, see Chapter 2, Storage Management Concepts, in Solaris Volume Manager Administration Guide. For information on soft partitions, see Chapter 12, Soft Partitions (Overview), in Solaris Volume Manager Administration Guide.

Disk Slices
The following table describes the slices that might be found on a system that runs the Solaris OS. On x86 systems:
182 System Administration Guide: Devices and File Systems June 2005

Disks are divided into fdisk partitions. An fdisk partition is a section of the disk that is reserved for a particular operating system, such as the Solaris OS. The Solaris OS places ten slices, numbered 09, on a Solaris fdisk partition.
Customary Disk Slices
File System Usually Found on Client or Server Systems? Comments

TABLE 112

Slice

root (/)

Both

Holds les and directories that make up the OS. EFI You cannot boot from a disk with an EFI label.

1 2

swap

Both Both

Provides virtual memory, or swap space. VTOC Refers to the entire disk, by convention. The size of this slice should not be changed. EFI Optional slice to be dened based on your sites needs.

/export, for example

Both

Optional slice that can be dened based on your sites needs. Can be used on a server to hold alternative versions of operating systems that are required by client systems.

4 5 /opt, for example

Both Both

Optional slice to be dened based on your sites needs. Optional slice to be dened based on your sites needs. Can be used to hold application software added to a system. If a slice is not allocated for the /opt le system during installation, the /opt directory is put in slice 0.

/usr

Both

Holds OS commands (also known as executables). This slice also holds documentation, system programs (init and syslogd, for example), and library routines. Holds les that are created by users.

/home or /export/home

Both

Chapter 11 Managing Disks (Overview)

183

TABLE 112 Slice

Customary Disk Slices

File System

(Continued)
Comments

Usually Found on Client or Server Systems?

N/A

N/A

VTOC Not applicable. EFI A reserved slice created by default. This area is similar to the VTOCs alternate cylinders. Do not modify or delete this slice.

9 (x86 only)

Both

EFI Not applicable. VTOC Provides an area that is reserved for alternate disk blocks. Slice 9 is known as the alternate sector slice.

Using Raw Data Slices

The disk label is stored in block 0 of each disk. So, third-party database applications that create raw data slices must not start at block 0. Otherwise, the disk label will be overwritten, and the data on the disk will be inaccessible. Do not use the following areas of the disk for raw data slices, which are sometimes created by third-party database applications:
I I

Block 0 where the disk label is stored Slice 2, which represents the entire disk with a VTOC label

Slice Arrangements on Multiple Disks

Although a single large disk can hold all slices and their corresponding le systems, two or more disks are often used to hold a systems slices and le systems.
Note A slice cannot be split between two or more disks. However, multiple swap slices on separate disks are allowed.

For instance, a single disk might hold the root (/) le system, a swap area, and the /usr le system, while another disk holds the /export/home le system and other le systems that contain user data. In a multiple disk arrangement, the disk that contains the OS and swap space (that is, the disk that holds the root (/) and /usr le systems and the slice for swap space) is called the system disk. Other disks are called secondary disks or non-system disks. When you arrange a systems le systems on multiple disks, you can modify le systems and slices on the secondary disks without having to shut down the system or reload the OS.
184 System Administration Guide: Devices and File Systems June 2005

When you have more than one disk, you also increase input-output (I/O) volume. By distributing disk load across multiple disks, you can avoid I/O bottlenecks.

Determining Which Slices to Use

When you set up a disks le systems, you choose not only the size of each slice, but also which slices to use. Your decisions about these matters depend on the conguration of the system to which the disk is attached and the software you want to install on the disk. System congurations that need disk space are as follows:
I I

Servers Stand-alone systems

Each system conguration can use slices in a different way. The following table lists some examples.
TABLE 113 Slice

System Congurations and Slices

Servers Stand-alone Systems

0 1 2 3 6 7

root swap /export /usr /export/home

root swap /usr /home

For more information about system congurations, see Overview of System Types in System Administration Guide: Basic Administration.
Note The Solaris installation utility provides default slice sizes based on the software you select for installation.

format Utility
Read the following overview of the format utility and its uses before proceeding to the how-to or reference sections. The format utility is a system administration tool that is used to prepare hard disk drives for use on your Solaris system.
Chapter 11 Managing Disks (Overview) 185

The following table shows the features and associated benets that the format utility provides.
TABLE 114 Feature

Features and Benets of the format Utility

Benet

Searches your system for all attached disk drives

Reports on the following: Target location I Disk geometry I Whether the disk is formatted I If the disk has mounted partitions
I

Retrieves disk labels Repairs defective sectors Formats and analyzes a disk Partitions a disk Labels a disk

Convenient for repair operations Allows administrators to repair disk drives with recoverable errors instead of sending the drive back to the manufacturer Creates sectors on the disk and veries each sector Divides a disk into slices so that individual le systems can be created on separate slices Writes disk name and conguration information to the disk for future retrieval (usually for repair operations)

The format utility options are described in Chapter 16.

When to Use the format Utility

Disk drives are partitioned and labeled by the Solaris installation utility when you install the Solaris release. You can use the format utility to do the following:
I I I I I I I

Display slice information Partition a disk Add a disk drive to an existing system Format a disk drive Label a disk Repair a disk drive Analyze a disk for errors

The main reason a system administrator uses the format utility is to partition a disk. These steps are covered in Chapter 13 and Chapter 14. See the following section for guidelines on using the format utility.

186

System Administration Guide: Devices and File Systems June 2005

Guidelines for Using the format Utility

TABLE 115 Task

format Utility Guidelines

Guidelines
I

For More Information

Format a disk.

Any existing data is destroyed when you How to Format a Disk on page 197 reformat a disk. The need for formatting a disk drive has decreased as more and more manufacturers ship their disk drives formatted and partitioned. You might not need to use the format utility when you add a disk drive to an existing system. If a disk has been relocated and is displaying many disk errors, you can attempt to reformat it. Reformatting automatically remaps any bad sectors. Data from the damaged system disk must be restored from a backup medium. Otherwise, the system will have to be reinstalled by using the installation utility. SPARC: How to Connect a System Disk and Boot on page 218, x86: How to Connect a System Disk and Boot on page 229, or, if the system must be reinstalled, Solaris 10 Installation Guide: Basic Installations SPARC: How to Create Disk Slices and Label a Disk on page 220 or x86: How to Create Disk Slices and Label a Disk on page 238 SPARC: How to Connect a Secondary Disk and Boot on page 219 or x86: How to Connect a Secondary Disk and Boot on page 231

Replace a system disk.

Divide a disk into slices.

Any existing data is destroyed when you repartition and relabel a disk with existing slices. Existing data must be copied to backup media before the disk is repartitioned and restored. Any existing data must be restored from backup media if the secondary disk is reformatted or repartitioned.

Add a secondary disk to an existing system.

Chapter 11 Managing Disks (Overview)

187

TABLE 115 Task

format Utility Guidelines

Guidelines
I

(Continued)
For More Information

Repair a disk drive.

Some customer sites prefer to replace rather than repair defective drives. If your site has a repair contract with the disk drive manufacturer, you might not need to use the format utility to repair disk drives. The repair of a disk drive usually means that a bad sector is added to a defect list. New controllers remap bad sectors with no system interruption. If the system has an older controller, you might need to remap a bad sector and restore any lost data.

Repairing a Defective Sector on page 211

Formatting a Disk
In most cases, disks are formatted by the manufacturer or reseller. So, they do not need to be reformatted when you install the drive. To determine if a disk is formatted, use the format utility. For more information, see How to Determine if a Disk Is Formatted on page 196. If you determine that a disk is not formatted, use the format utility to format the disk. When you format a disk, you accomplish two steps:
I I

The disk media is prepared for use. A list of disk defects based on a surface analysis is compiled.

Caution Formatting a disk is a destructive process because it overwrites data on the disk. For this reason, disks are usually formatted only by the manufacturer or reseller. If you think disk defects are the cause of recurring problems, you can use the format utility to do a surface analysis. However, be careful to use only the commands that do not destroy data. For details, see How to Format a Disk on page 197.

A small percentage of total disk space that is available for data is used to store defect and formatting information. This percentage varies according to disk geometry, and decreases as the disk ages and develops more defects. Formatting a disk might take anywhere from a few minutes to several hours, depending on the type and size of the disk.

188

System Administration Guide: Devices and File Systems June 2005

About Disk Labels

A special area of every disk is set aside for storing information about the disks controller, geometry, and slices. That information is called the disks label. Another term that is used to described the disk label is the VTOC (Volume Table of Contents) on a disk with a VTOC label. To label a disk means to write slice information onto the disk. You usually label a disk after you change its slices. If you fail to label a disk after you create slices, the slices will be unavailable because the OS has no way of knowing about the slices.

Partition Table Terminology

An important part of the disk label is the partition table. The partition table identies a disks slices, the slice boundaries (in cylinders), and the total size of the slices. You can display a disks partition table by using the format utility. The following describes partition table terminology.
TABLE 116

Partition Table Terminology

Value Description

Partition Term

Number

07

VTOC Partitions or slices, numbered 07. EFI Partitions or slices, numbered 06.

Tag

0=UNASSIGNED 1=BOOT 2=ROOT 3=SWAP 4=USR 5=BACKUP 7=VAR 8=HOME 11=RESERVED wm wu rm

A numeric value that usually describes the le system mounted on this partition.

Flags

The partition is writable and mountable. The partition is writable and unmountable. This state is the default for partitions that are dedicated for swap areas. (However, the mount command does not check the not mountable ag.) The partition is read only and mountable.

rm

Partition ags and tags are assigned by convention and require no maintenance. For more information on displaying the partition table, see the following references:
I I

Displaying Partition Table Information on page 190 How to Display Disk Slice Information on page 199
Chapter 11 Managing Disks (Overview) 189

How to Examine a Disk Label on page 203

Displaying Partition Table Information

The following format utility output shows an example of a partition table from a 74-Gbyte disk with a VTOC label displayed:
Total disk cylinders available: 38756 + 2 (reserved cylinders) Part Tag 0 root 1 swap 2 backup 3 unassigned 4 unassigned 5 unassigned 6 unassigned 7 home 8 boot 9 alternates partition> Flag wm wu wm wm wm wm wm wm wu wu Cylinders 3 - 2083 2084 - 3124 0 - 38755 0 0 0 0 3125 - 38755 0 0 1 2 Size 4.00GB 2.00GB 74.51GB 0 0 0 0 68.50GB 1.97MB 3.94MB Blocks (2081/0/0) 8390592 (1041/0/0) 4197312 (38756/0/0) 156264192 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0 (35631/0/0) 143664192 (1/0/0) 4032 (2/0/0) 8064

The partition table displayed by the format utility contains the following information.

Column Name

Description

Part

Partition or slice number. See Table 116 for a description of this column. Partition tag. See Table 116 for a description of this column. Partition ag. See Table 116 for a description of this column. The starting and ending cylinder number for the slice. Not displayed on EFI-labeled disks. The slice size in Mbytes. The total number of cylinders and the total number of sectors per slice. Not displayed on EFI-labeled disks. EFI The starting block number. Not displayed on VTOC-labeled disks. EFI The ending block number. Not displayed on VTOC-labeled disks.

Tag Flag Cylinders

Size Blocks

First Sector

Last Sector

The following is an example of an EFI disk label displayed by using the prtvtoc command.
190 System Administration Guide: Devices and File Systems June 2005

# * * * * * * * * * * * * *

prtvtoc /dev/rdsk/c4t1d0s0 /dev/rdsk/c4t1d0s0 partition map Dimensions: 512 bytes/sector 2576941056 sectors 2576940989 accessible sectors Flags: 1: unmountable 10: read-only First Sector Last Flags Sector Count Sector Mount Directory 00 34 629145600 629145633 00 629145634 629145600 1258291233 00 1258291234 1318633404 2576924637 00 2576924638 16384 2576941021

Partition 0 1 6 8

Tag 2 4 4 11

The output of the prtvtoc command provides information in the following three sections:
I I I

Dimensions Flags Partition Table

prtvtoc Column Name

Description

Partition

Partition or slice number. For a description of this column, see Table 116. Partition tag. For a description of this column, see Table 116. Partition ag. For a description of this column, see Table 116. The rst sector of the slice. The total number of sectors in the slice. The last sector of the slice. The last mount point directory for the le system.

Tag Flags First Sector Sector Count Last Sector Mount Directory

Partitioning a Disk
The format utility is most often used by system administrators to partitioning a Disk. The steps are as follows:
I

Determining which slices are needed

Chapter 11 Managing Disks (Overview) 191

I I I I

Determining the size of each slice or partition Using the format utility to partition the disk Labeling the disk with new partition information Creating the le system for each partition

The easiest way to partition a disk is to use the modify command from the partition menu of the format utility. The modify command allows you to create partitions by specifying the size of each partition without having to keep track of the starting cylinder boundaries. The modify command also keeps tracks of any disk space that remains in the free hog slice.

Using the Free Hog Slice

When you use the format utility to change the size of one or more disk slices, you designate a temporary slice that will expand and shrink to accommodate the resizing operations. This temporary slice donates, or frees, space when you expand a slice, and receives, or hogs, the discarded space when you shrink a slice. For this reason, the donor slice is sometimes called the free hog. The free hog slice exists only during installation or when you run the format utility. There is no permanent free hog slice during day-to-day operations. For information on using the free hog slice, see SPARC: How to Create Disk Slices and Label a Disk on page 220 or x86: How to Create Disk Slices and Label a Disk on page 238.

192

System Administration Guide: Devices and File Systems June 2005

CHAPTER

12

Administering Disks (Tasks)

This chapter contains disk administration procedures. Many procedures described in this chapter are optional if you are already familiar with how disks are managed on systems running the Solaris OS. For information on the procedures associated with administering disks, see Administering Disks (Task Map) on page 193. For overview information about disk management, see Chapter 11.

Administering Disks (Task Map)

Task Description For Instructions

Identify the disks on a system. If you are not sure of the types How to Identify the Disks on of disks on a system, use the a System on page 194 format utility to identify the disk types. Format the disk. Determine whether a disk is already formatted by using the format utility. In most cases, disks are already formatted. Use the format utility if you need to format a disk. Display slice information. Display slice information by using the format utility. How to Determine if a Disk Is Formatted on page 196 How to Format a Disk on page 197

How to Display Disk Slice Information on page 199

193

Task

Description

For Instructions

Label the disk. Examine the disk label. Recover a corrupted disk label.

Create the disk label by using the format utility.

How to Label a Disk on page 201

Examine the disk label by How to Examine a Disk using the prtvtoc command. Label on page 203 You can attempt to recover a disk label that was damaged due to a system or power failure. Create a format.dat entry to support a third-party disk. How to Recover a Corrupted Disk Label on page 205

Create a format.dat entry.

How to Create a format.dat Entry on page 208

Automatically congure a SCSI disk.

How to Automatically You can automatically congure a SCSI disk with the Congure a SCSI Drive on page 209 SCSI-2 specication for disk device mode sense pages even if the specic drive type is not listed in the /etc/format.dat le. Identify a defective disk sector How to Identify a Defective by using the format utility. Sector by Using Surface Analysis on page 211 Fix a defective disk sector by using the format utility. How to Repair a Defective Sector on page 213

Identify a defective disk sector. If necessary, x a defective disk sector.

Identifying Disks on a System

Use the format utility to discover the types of disks that are connected to a system. You can also use the format utility to verify that a disk is known to the system. For detailed information on using the format utility, see Chapter 16.

M
Steps

How to Identify the Disks on a System

1. Become superuser or assume an equivalent role. Roles contain authorizations and privileged commands. For more information about roles, see Conguring RBAC (Task Map) in System Administration Guide: Security Services.

194

System Administration Guide: Devices and File Systems June 2005

2. Identify the disks that are recognized on the system by using the format utility.
# format

The format utility displays a list of disks that it recognizes under AVAILABLE DISK SELECTIONS.
Example 121

Identifying the Disks on a System

The following example shows format command output is from a system with one disk.
# format AVAILABLE DISK SELECTIONS: 0. c0t1d0 <FUJITSU MAN3367M SUN36G /pci@1f,0/pci@1,1/scsi@2/sd@1,0

1804 43d671f>

The output associates a disks physical and logical device name to the disks marketing name, which appears in angle brackets <>. See the example below. This method is an easy way to identify which logical device names represent the disks that are connected to your system. For a description of logical and physical device names, see Chapter 10. The following example uses a wildcard to display the four disks that are connected to a second controller.
# format /dev/rdsk/c2* AVAILABLE DISK SELECTIONS: 0. /dev/rdsk/c2t10d0s0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133> /sbus@3,0/SUNW,fas@3,8800000/sd@a,0 1. /dev/rdsk/c2t11d0s0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133> /sbus@3,0/SUNW,fas@3,8800000/sd@b,0 2. /dev/rdsk/c2t14d0s0 <SUN18G cyl 7506 alt 2 hd 19 sec 248> /sbus@3,0/SUNW,fas@3,8800000/sd@e,0 3. /dev/rdsk/c2t15d0s0 <SUN18G cyl 7506 alt 2 hd 19 sec 248> /sbus@3,0/SUNW,fas@3,8800000/sd@f,0 Specify disk (enter its number):

The following example shows how to identify the disks on a SPARC based system.
# format 0. c0t1d0 <FUJITSU MAN3367M SUN36G /pci@1f,0/pci@1,1/scsi@2/sd@1,0 Specify disk (enter its number): 1804 43d671f>

The output identies that disk 0 (target 1) is connected to the second SCSI host adapter (scsi@2), which is connected to the second PCI interface (/pci@1f0/pci@1,1...). The output also associates both the physical and logical device name to the disks marketing name, SUN36G. The following example shows how to identify the disks on an x86 based system.
# format AVAILABLE DISK SELECTIONS: 0. c0d0 <DEFAULT cyl 615 alt 2 hd 64 sec 63> Chapter 12 Administering Disks (Tasks) 195

/pci@0,0/pci-ide@7,1/ata@0/cmdk@0,0 1. c0d1 <DEFAULT cyl 522 alt 2 hd 32 sec 63> /pci@0,0/pci-ide@7,1/ata@0/cmdk@1,0 2. c1d0 <DEFAULT cyl 817 alt 2 hd 256 sec 63> /pci@0,0/pci-ide@7,1/ata@1/cmdk@0,0 Specify disk (enter its number):

The output shows that disk 0 is connected to the rst PCI host adapter (pci-ide@7...), which is connected to the ATA interface (ata...). The format output on an x86 based system does not identify disks by their marketing names.
More Information

If the format Utility Does Not Recognize a Disk ...

I I I I

Go to Chapter 13 or Chapter 14. Go to Creating a format.dat Entry on page 208. Go to How to Label a Disk on page 201. Connect the disk to the system by using your disk hardware documentation.

Formatting a Disk
Disks are typically formatted by the manufacturer or reseller. They usually do not need to be reformatted when you install the drive. A disk must be formatted before you can do the following:
I I

Write data to the disk. However, most disks are already formatted. Use the Solaris installation utility to install the system.

Caution Formatting a disk is a destructive process because it overwrites data on the disk. For this reason, disks are usually formatted only by the manufacturer or reseller. If you think disk defects are the cause of recurring problems, you can use the format utility to do a surface analysis. However, be careful to use only the commands that do not destroy data.

M
Steps

How to Determine if a Disk Is Formatted

1. Become superuser or assume an equivalent role. 2. Invoke the format utility.
# format

196

System Administration Guide: Devices and File Systems June 2005

A numbered list of disks is displayed. 3. Type the number of the disk that you want to check.
Specify disk (enter its number): 0

4. Verify that the disk you chose is formatted by noting the following message:
[disk formatted]

Example 122

Determining if a Disk Is Formatted

The following example shows that disk c1t0d0 is formatted.
# format /dev/rdsk/c1* AVAILABLE DISK SELECTIONS: 0. /dev/rdsk/c1t0d0s0 <SUN18G cyl 7506 /sbus@2,0/QLGC,isp@2,10000/sd@0,0 1. /dev/rdsk/c1t1d0s0 <SUN18G cyl 7506 /sbus@2,0/QLGC,isp@2,10000/sd@1,0 2. /dev/rdsk/c1t8d0s0 <SUN18G cyl 7506 /sbus@2,0/QLGC,isp@2,10000/sd@8,0 3. /dev/rdsk/c1t9d0s0 <SUN18G cyl 7506 /sbus@2,0/QLGC,isp@2,10000/sd@9,0 Specify disk (enter its number): 0 selecting /dev/rdsk/c1t0d0s0 [disk formatted]

alt 2 hd 19 sec 248> alt 2 hd 19 sec 248> alt 2 hd 19 sec 248> alt 2 hd 19 sec 248>

M
Steps

How to Format a Disk

1. Become superuser or assume an equivalent role. 2. Invoke the format utility.
# format

A numbered list of disks is displayed. 3. Type the number of the disk that you want to format.
Specify disk (enter its number): 0

Caution Do not select the system disk. If you format your system disk, you delete the OS and any data on this disk.

4. To begin formatting the disk, type format at the format> prompt. Conrm the command by typing y.
format> format Ready to format. Formatting cannot be interrupted Chapter 12 Administering Disks (Tasks) 197

and takes 23 minutes (estimated). Continue? yes

5. Verify that the disk format was successful by noting the following messages:
Beginning format. The current time Tue ABC xx xx:xx:xx xxxx Formatting... done Verifying media... pass 0 - pattern = 0xc6dec6de 2035/12/18 pass 1 - pattern = 0x6db6db6d 2035/12/18 Total of 0 defective blocks repaired.

6. Exit the format utility.

format> quit

Example 123

Formatting a Disk
The following example shows how to format the disk c0t6d0.
# format Searching for disks...done

AVAILABLE DISK SELECTIONS: 0. c0t0d0 <SUNW18G cyl 7506 alt 2 hd 19 sec 248 /pci@1f,0/pci@1,1/scsi@2/sd@0,0 1. c0t1d0 <FUJITSU MAN3367M SUN36G 1804 43d671f> /pci@1f,0/pci@1,1/scsi@2/sd@1,0 2. c0t2d0 <FUJITSU MAN3367M SUN36G 1804 43d671f> /pci@1f,0/pci@1,1/scsi@2/sd@2,0 3. c0t3d0 <FUJITSU MAN3367M SUN36G 1804 43d671f> /pci@1f,0/pci@1,1/scsi@2/sd@3,0 4. c0t4d0 <FUJITSU MAN3367M SUN36G 1804 43d671f> /pci@1f,0/pci@1,1/scsi@2/sd@4,0 5. c0t5d0 <FUJITSU MAN3367M SUN36G 1804 43d671f> /pci@1f,0/pci@1,1/scsi@2/sd@5,0 6. c0t6d0 <FUJITSU MAN3367M SUN36G 1804 43d671f> /pci@1f,0/pci@1,1/scsi@2/sd@6,0 Specify disk (enter its number): 6 selecting c0t6d0 [disk formatted] format> format Ready to format. Formatting cannot be interrupted and takes 332 minutes (estimated). Continue? y Beginning format. The current time is Wed Jan 7 16:16:05 2004 Formatting... 99% complete (00:00:21 remaining) done 198 System Administration Guide: Devices and File Systems June 2005

Verifying media... pass 0 - pattern = 0xc6dec6de 71132922 pass 1 - pattern = 0x6db6db6d 71132922 Total of 0 defective blocks repaired. format> quit

Displaying Disk Slices

You can use the format utility to check whether a disk has the appropriate disk slices. If you determine that a disk does not contain the slices you want to use, use the format utility to re-create them and label the disk. For information on creating disk slices, see SPARC: How to Create Disk Slices and Label a Disk on page 220 or x86: How to Create Disk Slices and Label a Disk on page 238.
Note The format utility uses the term partition instead of slice.

M
Steps

How to Display Disk Slice Information

1. Become superuser or assume an equivalent role. 2. Invoke the format utility.
# format

A numbered list of disks is displayed. 3. Type the number of the disk for which you want to display slice information.
Specify disk (enter its number):1

4. Select the partition menu.

format> partition

5. Display the slice information for the selected disk.

partition> print

6. Exit the format utility.

partition> q format> q Chapter 12 Administering Disks (Tasks) 199

7. Verify the displayed slice information by identifying specic slice tags and slices. If the screen output shows that no slice sizes are assigned, the disk probably does not have slices.
Example 124

Displaying Disk Slice Information

The following example displays slice information for a disk with a VTOC label.
# format Searching for disks...done Specify disk (enter its number):1 Selecting c0t0d0 format> partition partition> print Current partition table (original): Total disk cylinders available: 8892 + 2 (reserved cylinders) Part Tag 0 root 1 swap 2 backup 3 unassigned 4 unassigned 5 unassigned 6 unassigned 7 home partition> q format> q # Flag wm wu wm wm wm wm wm wm Cylinders 1110 - 4687 0 - 1109 0 - 8891 0 0 0 0 4688 - 8891 Size 1.61GB 512.00MB 4.01GB 0 0 0 0 1.89GB Blocks (0/3578/0) 3381210 (0/1110/0) 1048950 (0/8892/0) 8402940 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0 (0/4204/0) 3972780

For a detailed description of the slice information in these examples, see Chapter 11. The following example shows the slice information for a disk with an EFI label.
# format Searching for disks...done Specify disk (enter its number): 9 selecting c4t1d0 [disk formatted] format> partition partition> print Current partition table (original): partition> q format> q Part Tag Flag First Sector 0 root wm 34 1 usr wm 629145634 2 unassigned wm 0 3 unassigned wm 0 4 unassigned wm 0 5 unassigned wm 0 200 System Administration Guide: Devices and File Systems June 2005

Size 300.00GB 300.00GB 0 0 0 0

Last Sector 629145633 1258291233 0 0 0 0

6 8

usr reserved

wm wm

1258291234 2576924638

628.77GB 8.00MB

2576924637 2576941021

Creating and Examining a Disk Label

The labeling of a disk is usually done during system installation or when you are creating new disk slices. You might need to relabel a disk if the disk label becomes corrupted. For example, from a power failure. The format utility attempts to automatically congure any unlabeled SCSI disk. If the format utility is able to automatically congure an unlabeled disk, it displays a message similar to the following:
c0t0d1: configured with capacity of 4.00GB

Tip For information on labeling multiple disks with the same disk label, see Labeling Multiple Disks by Using the prtvtoc and fmthard Commands on page 214.

How to Label a Disk

You can use the following procedure to do the following:
I I

Label a disk with a VTOC label or a disk greater than 1 terabyte with an EFI label. Label a disk that is greater than 1 terabyte with an EFI label.

If you want to put an EFI label on disk smaller than 1 terabyte, see Example 126.
Steps

1. Become superuser or assume an equivalent role. 2. Invoke the format utility.

# format

A numbered list of disks is displayed. 3. Type the number of the disk that you want to label.
Specify disk (enter its number):1

If the format utility recognizes the disk type, the next step is to search for a backup label to label the disk. Labeling the disk with the backup label labels the disk with the correct partitioning information, the disk type, and disk geometry. 4. Select one of the following to label the disk:
Chapter 12 Administering Disks (Tasks) 201

If the disk is unlabeled and was successfully congured, go to Step 5 to label the disk. The format utility will ask if you want to label the disk. If the disk is labeled but you want to change the disk type, or if the format utility was not able to automatically congure the disk, proceed to Step 6 to set the disk type and label the disk.

5. Label the disk by typing y at the Label it now? prompt.

Disk not labeled. Label it now? y

The disk is now labeled. Go to step 10 to exit the format utility. 6. Enter type at the format> prompt.
format> type

The Available Drive Types menu is displayed. 7. Select a disk type from the list of possible disk types.
Specify disk type (enter its number)[12]: 12

Or, select 0 to automatically congure a SCSI-2 disk. For more information, see How to Automatically Congure a SCSI Drive on page 209. 8. Label the disk. If the disk is not labeled, the following message is displayed.
Disk not labeled. Label it now? y

Otherwise, you are prompted with this message:

Ready to label disk, continue? y

9. Verify the disk label.

format> verify

10. Exit the format utility.

format> q #

Example 125

Labeling a Disk
The following example shows how to automatically congure and label a 1.05-Gbyte disk.
# format c1t0d0: configured with capacity of 1002.09MB AVAILABLE DISK SELECTIONS: 0. c0t3d0 <SUN1.05 cyl 2036 alt 2 hd 14 sec 72> /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@1,0 1. c1t0d0 <SUN1.05 cyl 2036 alt 2 hd 14 sec 72> /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@1,0

202

System Administration Guide: Devices and File Systems June 2005

Specify disk (enter its number): 1 Disk not labeled. Label it now? yes format> verify format> q #

Example 126

Labeling a Disk Less Than 1 Terabyte with an EFI Label

The following example shows how to use the format -e command to label a disk that is less than 1 terabyte with an EFI label. Remember to verify that your layered software products will continue to work on systems with EFI-labeled disks. For general information on EFI label restrictions, see Restrictions of the EFI Disk Label on page 177.
# format -e Searching for disks...done AVAILABLE DISK SELECTIONS: 1. c1t0d0 <SUNW18g cyl 7506 alt 2 hd /sbus@2,0/QLGC,isp@2,10000/sd@0,0 2. c1t1d0 <SUNW18g cyl 7506 alt 2 hd /sbus@2,0/QLGC,isp@2,10000/sd@1,0 3. c1t8d0 <SUNW18g cyl 7506 alt 2 hd /sbus@2,0/QLGC,isp@2,10000/sd@8,0 4. c1t9d0 <SUNW18g cyl 7506 alt 2 hd /sbus@2,0/QLGC,isp@2,10000/sd@9,0 Specify disk (enter its number): 4 selecting c1t9d0 [disk formatted] format> label [0] SMI Label [1] EFI Label Specify Label type[0]: 1 Ready to label disk, continue? yes format> quit

19 sec 248> 19 sec 248> 19 sec 248> 19 sec 248>

How to Examine a Disk Label

Examine disk label information by using the prtvtoc command. For a detailed description of the disk label and the information that is displayed by the prtvtoc command, see Chapter 11.

Steps

1. Become superuser or assume an equivalent role. 2. Display the disk label information.
# prtvtoc /dev/rdsk/device-name

where device-name is the raw disk device you want to examine.

Example 127

Examining a Disk Label

The following example shows disk label information for a disk with a VTOC label.
Chapter 12 Administering Disks (Tasks) 203

# * * * * * * * * * * * * * * * *

prtvtoc /dev/rdsk/c0t0d0s0 /dev/rdsk/c0t0d0s0 partition map Dimensions: 512 bytes/sector 63 sectors/track 15 tracks/cylinder 945 sectors/cylinder 8894 cylinders 8892 accessible cylinders Flags: 1: unmountable 10: read-only First Sector 1048950 0 0 4430160 Sector Count 3381210 1048950 8402940 3972780 Last Sector 4430159 1048949 8402939 8402939

Partition 0 1 2 7

Tag 2 3 5 8

Flags 00 01 00 00

Mount Directory /

/export/home

The following example shows disk label information for a disk with an EFI label.
# * * * * * * * * * * * * * prtvtoc /dev/rdsk/c3t1d0s0 /dev/rdsk/c3t1d0s0 partition map Dimensions: 512 bytes/sector 2479267840 sectors 2479267773 accessible sectors Flags: 1: unmountable 10: read-only First Flags Sector 00 34 01 262178 00 524322 00 2479251422 Sector Count 262144 262144 2478727100 16384 Last Sector 262177 524321 2479251421 2479267805

Partition 0 1 6 8

Tag 2 3 4 11

Mount Directory

Recovering a Corrupted Disk Label

Sometimes, a power or system failure causes a disks label to become unrecognizable. A corrupted disk label doesnt always mean that the slice information or the disks data must be re-created or restored.
204 System Administration Guide: Devices and File Systems June 2005

The rst step to recovering a corrupted disk label is to label the disk with the correct geometry and disk type information. You can complete this step through the normal disk labeling method, by using either automatic conguration or manual disk type specication. If the format utility recognizes the disk type, the next step is to search for a backup label to label the disk. Labeling the disk with the backup label labels the disk with the correct partitioning information, the disk type, and disk geometry.

M
Steps

How to Recover a Corrupted Disk Label

1. Boot the system to single-user mode. If necessary, boot the system from a local CD-ROM or the network in single-user mode to access the disk. See Chapter 11, Booting a System (Tasks), in System Administration Guide: Basic Administration or Chapter 12, Booting a System (Tasks), in System Administration Guide: Basic Administration for information on booting the system. 2. Relabel the disk.
# format

The format utility attempts to automatically congure any unlabeled SCSI disk. If the format utility is able to congure the unlabeled and corrupted disk, it will display this message:
cwtxdy: configured with capacity of abcMB

The format utility then displays a numbered list of disks on the system. 3. Type the number of the disk that you need to recover.
Specify disk (enter its number): 1

4. Select one of the following to determine how to label the disk.

I I

If the disk was congured successfully, follow Steps 5 and 6. Then go to step 12. If the disk was not congured successfully, follow Steps 711. Then go to step 12.

5. Search for the backup label.

format> verify Warning: Could not read primary label. Warning: Check the current partitioning and label the disk or use the backup command. Backup label contents: Volume name = < > ascii name = <SUN1.05 cyl 2036 alt 2 hd 14 sec 72> pcyl = 2038 Chapter 12 Administering Disks (Tasks) 205

ncyl = 2036 acyl = 2 nhead = 14 nsect = 72 Part Tag Flag 0 root wm 1 swap wu 2 backup wm 3 unassigned wm 4 unassigned wm 5 unassigned wm 6 usr wm 7 unassigned wm

Cylinders 0 - 300 301 - 524 0 - 2035 0 0 0 525 - 2035 0

Size 148.15MB 110.25MB 1002.09MB 0 0 0 743.70MB 0

Blocks (301/0/0) 303408 (224/0/0) 225792 (2036/0/0) 2052288 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0 (1511/0/0) 1523088 (0/0/0) 0

6. If the format utility was able to nd a backup label and the backup label contents appear satisfactory, use the backup command to label the disk with the backup label.
format> backup Disk has a primary label, still continue? y Searching for backup labels...found. Restoring primary label

The disk label has been recovered. Go to Step 12. 7. If the format utility was not able to automatically congure the disk, specify the disk type by using the type command.
format> type

The Available Drives Type menu is displayed. 8. Select 0 to automatically congure the disk. Or, select a disk type from the list of possible disk types.
Specify disk type (enter its number)[12]: 12

9. If the disk was successfully congured, reply with no when the format utility asks if you want to label the disk.
Disk not labeled. Label it now? no

10. Use the verify command to search for backup labels.

format> verify Warning: Could not read primary label. Warning: Check the current partitioning and label the disk or use the backup command. . . .

206

System Administration Guide: Devices and File Systems June 2005

11. If the format utility was able to nd a backup label and the backup label contents appear satisfactory, use the backup command to label the disk with the backup label.
format> backup Disk has a primary label, still continue? y Searching for backup labels...found. Restoring primary label

The disk label has been recovered. 12. Exit the format utility.
format> q

13. Verify the le systems on the recovered disk by using the fsck command. For information on using the fsck command, see Chapter 22.

Adding a Third-Party Disk

The Solaris OS supports many third-party disks. However, for the disk to be recognized, you might need to supply either a device driver, a format.dat entry, or both. Other options for adding disks are as follows:
I

If you are adding a SCSI disk, you might to try the format utilitys automatic conguration feature. For more information, see Automatically Conguring SCSI Disk Drives on page 209. You might try hot-plugging a PCI, SCSI, or USB disk. For more information, see Chapter 5.

If the third-party disk is designed to work with standard SunOS compatible device drivers, then the creation of an appropriate format.dat entry should suffice to allow the disk to be recognized by the format utility. In other cases, you need to load a third-party device driver to support the disk.
Note Sun cannot guarantee that its format utility will work properly with all third-party disk drivers. If the disk driver is not compatible with the Solaris format utility, the disk drive vendor should supply you with a custom disk formatting program.

This section discusses what to do if some of this software support is missing. Typically, you discover that software support is missing when you invoke the format utility and nd that the disk type is not recognized.
Chapter 12 Administering Disks (Tasks) 207

Supply the missing software as described in this section. Then, refer to the appropriate conguration procedure for adding system disks or secondary disks in Chapter 13 or Chapter 14.

Creating a format.dat Entry

Unrecognized disks cannot be formatted without precise information about the disks geometry and operating parameters. This information is supplied in the /etc/format.dat le.
Note SCSI-2 disks do not require a format.dat entry. The format utility

automatically congures the SCSI-2 drivers if the disks are powered on during a reconguration boot. For step-by-step instructions on conguring a SCSI disk drive automatically, see How to Automatically Congure a SCSI Drive on page 209.

If your disk is unrecognized, use a text editor to create an entry in format.dat for the disk. You need to gather all the pertinent technical specications about the disk and its controller before you start. This information should have been provided with the disk. If not, contact the disk manufacturer or your supplier.

M
Steps

How to Create a format.dat Entry

1. Become superuser or assume an equivalent role. 2. Make a copy of the /etc/format.dat le.
# cp /etc/format.dat /etc/format.dat.gen

3. Modify the /etc/format.dat le to include an entry for the third-party disk. Use the format.dat information that is described in Chapter 16. Also, use the disks hardware product documentation to gather the required information.

208

System Administration Guide: Devices and File Systems June 2005

Automatically Conguring SCSI Disk Drives

The format utility automatically congures SCSI disk drives even if that specic type of drive is not listed in the /etc/format.dat le. This feature enables you to format, create slices for, and label any disk driver that is compliant with the SCSI-2 specication for disk device mode sense pages. Here are other options for adding disks:
I

If you are adding a SCSI disk, you might to try the format utilitys automatic conguration feature. You might try hot-plugging a PCI, SCSI, or USB disk. For more information, see Chapter 5.

The following steps are involved in conguring a SCSI drive by using automatic conguration:
I I I I I

Shutting down the system Attaching the SCSI disk drive to the system Turning on the disk drive Performing a reconguration boot Using the format utility to automatically congure the SCSI disk drive

After the reconguration boot, invoke the format utility. The format utility will attempt to congure the disk and, if successful, alert the user that the disk was congured. For step-by-step instructions on automatically conguring a SCSI disk drive, see How to Automatically Congure a SCSI Drive on page 209. Heres an example of a partition table for a 1.3-Gbyte SCSI disk drive that was displayed by the format utility.
Part 0 1 2 6 Tag root swap backup usr Flag wm wu wu wm Cylinders 0 96 97 - 289 0 - 1964 290 - 1964 Size 64.41MB 128.16MB 1.27GB 1.09GB Blocks (97/0/0) (193/0/0) (1965/0/0) (1675/0/0)

M
Steps

How to Automatically Congure a SCSI Drive

1. Become superuser or equivalent role. 2. Create the /reconfigure le that will be read when the system is booted.
# touch /reconfigure Chapter 12 Administering Disks (Tasks) 209

3. Shut down the system.

# shutdown -i0 -gn -y

-i0 -gn -y

Brings the system down to init level 0, the power-down state. Noties logged-in users that they have n seconds before the system begins to shut down. Species that the command should run without user intervention.

The ok prompt is displayed after the system is shut down. 4. Turn off the power to the system and all external peripheral devices. 5. Ensure that the disk you are adding has a different target number than the other devices on the system. Typically, a small switch is located at the back of the disk for this purpose. 6. Connect the disk to the system, and check the physical connections. Refer to the disks hardware installation guide for details. 7. Turn on the power to all external peripherals. 8. Turn on the power to the system. The system boots and displays the login prompt. 9. Log back in as superuser or assume an equivalent role. 10. Invoke the format utility, and select the disk that you want to congure automatically.
# format Searching for disks...done c1t0d0: configured with capacity of 1002.09MB AVAILABLE DISK SELECTIONS: 0. c0t1d0 <SUN1.05 cyl 2036 alt 2 hd 14 sec 72> /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@1,0 1. c0t3d0 <SUN1.05 cyl 2036 alt 2 hd 14 sec 72> /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@3,0 Specify disk (enter its number): 1

11. Type yes in response to the prompt to label the disk. Typing y causes the disk label to be generated and written to the disk by using SCSI automatic conguration.
Disk not labeled. Label it now? y

12. Verify the disk label.

format> verify

210

System Administration Guide: Devices and File Systems June 2005

13. Exit the format utility.

format> q

Repairing a Defective Sector

If a disk on your system has a defective sector, you can repair the disk by following procedures in this section. You might become aware of defective sectors when you do the following:
I

Run surface analysis on a disk For more information on the analysis feature of the format utility, see analyze Menu on page 268. The defective area reported while your system is running might not be accurate. Because the system does disk operations many sectors at a time, it is often hard to pinpoint exactly which sector caused a given error. To nd the exact sector or sectors, use How to Identify a Defective Sector by Using Surface Analysis on page 211.

Get multiple error messages from the disk driver concerning a particular portion of the disk while your system is running. Console messages that are related to disk errors appear similar to the following:
WARNING: /io-unit@f,e0200000/sbi@0,0/QLGC,isp@1,10000/sd@3,0 (sd33): Error for command read Error Level: Retryable Requested Block 126, Error Block: 179 Sense Key: Media Error Vendor name: ASC = 0x11 (unrecovered read error), ASCQ = 0x0, FRU = 0x0

This message indicates that block 179 might be defective. You would relocate the bad block by using the format utilitys repair command. Or, you would use the analyze command with the repair option enabled.

How to Identify a Defective Sector by Using Surface Analysis

1. Become superuser or assume an equivalent role. 2. Unmount the le system in the slice that contains the defective sector.
# umount /dev/dsk/device-name Chapter 12 Administering Disks (Tasks) 211

Steps

For more information, see mount(1M). 3. Invoke the format utility.

# format

4. Select the affected disk.

Specify disk (enter its number):1 selecting c0t2d0: [disk formatted] Warning: Current Disk has mounted partitions.

5. Select the analyze menu.

format> analyze

6. Set up the analysis parameters by typing setup at the analyze> prompt. Use the parameters shown here:
analyze> setup Analyze entire disk [yes]? n Enter starting block number [0, 0/0/0]: 12330 Enter ending block number [2052287, 2035/13/71]: 12360 Loop continuously [no]? y Repair defective blocks [yes]? n Stop after first error [no]? n Use random bit patterns [no]? n Enter number of blocks per transfer [126, 0/1/54]: 1 Verify media after formatting [yes]? y Enable extended messages [no]? n Restore defect list [yes]? y Create defect label [yes]? y

7. Find the defect by using the read command.

analyze> read Ready to analyze (wont harm SunOS). This takes a long time, but is interruptible with Control-C. Continue? y pass 0 2035/12/1825/7/24 pass 1 Block 12354 (18/4/18), Corrected media error (hard data ecc) 25/7/24 ^C Total of 1 defective blocks repaired.

212

System Administration Guide: Devices and File Systems June 2005

M
Steps

How to Repair a Defective Sector

1. Become superuser or assume an equivalent role. 2. Invoke the format utility.
# format

3. Select the disk that contains the defective sector.

Specify disk (enter its number): 1 selecting c0t3d0 [disk formatted] format>

4. Select the repair command.

format> repair

5. Type the defective block number.

Enter absolute block number of defect: 12354 Ready to repair defect, continue? y Repairing block 12354 (18/4/18)...ok. format>

If you are unsure of the format that is used to identify the defective sector, see How to Identify a Defective Sector by Using Surface Analysis on page 211 for more information.

Tips and Tricks for Managing Disks

Use the following tips to help you manage disks more efficiently.

Debugging format Sessions

Invoke the format -M command to enable extended and diagnostic messages for ATA and SCSI devices.
EXAMPLE 128

Debugging format Sessions

In this example, the series of numbers under Inquiry represent the hexadecimal value of the inquiry data that is displayed to the right of the numbers.

Chapter 12 Administering Disks (Tasks)

213

EXAMPLE 128

Debugging format Sessions

(Continued)

# format -M Searching for disks...done AVAILABLE DISK SELECTIONS: 0. c0t1d0 <SUN1.05 cyl 2036 alt 2 hd 14 sec 72> /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@1,0 1. c0t3d0 <SUN1.05 cyl 2036 alt 2 hd 14 sec 72> /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@3,0 Specify disk (enter its number): selecting c0t3d0 [disk formatted] format> inquiry Inquiry: 00 00 02 02 8f 00 00 12 53 45 41 53 54 31 31 32 30 30 4e 20 53 55 38 33 35 38 30 30 30 33 30 32 30 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 43 6f 70 79 72 69 67 68 74 20 39 39 32 20 53 65 61 67 61 74 65 72 69 67 68 74 73 20 72 65 73 65 30 30 30 Vendor: name Product: ST11200N SUN1.05 Revision: 8358 format> 0

47 4e 39 00 00 00 28 20 72

41 31 00 00 00 00 63 41 76

54 2e 00 00 00 00 29 6c 65

45 30 00 00 00 00 20 6c 64

20 35 00 00 00 00 31 20 20

........NAME.... ST11200N SUN1.05 835800030209.... ................ ................ ................ .Copyright (c) 1 992 NAME All rights reserved 000

Labeling Multiple Disks by Using the prtvtoc and fmthard Commands

Use the prtvtoc and fmthard commands to label multiple disks with the same disk geometry. Use the following for loop in a script to copy a disk label from one disk and replicate it on multiple disks.
# > > > for i in x y z do prtvtoc /dev/rdsk/cwtxdysz | fmthard -s - /dev/rdsk/cwt${i}d0s2 done

EXAMPLE 129

Labeling Multiple Disks

In this example, the disk label from c2t0d0s0 is copied to four other disks.

214

System Administration Guide: Devices and File Systems June 2005

EXAMPLE 129

Labeling Multiple Disks

(Continued)

# for i in 1 2 3 5 > do > prtvtoc /dev/rdsk/c2t0d0s0 | fmthard > done fmthard: New volume table of contents fmthard: New volume table of contents fmthard: New volume table of contents fmthard: New volume table of contents #

-s - /dev/rdsk/c2t${i}d0s2 now now now now in in in in place. place. place. place.

Chapter 12 Administering Disks (Tasks)

215

216

System Administration Guide: Devices and File Systems June 2005

CHAPTER

13

SPARC: Adding a Disk (Tasks)

This chapter describes how to add a disk to a SPARC system. For information on the procedures associated with adding a disk to a SPARC system, see SPARC: Adding a System Disk or a Secondary Disk (Task Map) on page 217. For overview information about disk management, see Chapter 11. For step-by-step instructions on adding a disk to an x86 based system, see Chapter 14.

SPARC: Adding a System Disk or a Secondary Disk (Task Map)

The following task map identies the procedures for adding a disk to a SPARC based system.

Task

Description

For Instructions

1. Connect the disk and boot.

System Disk Connect the new disk and boot from a local or remote Solaris CD or DVD. Secondary Disk Connect the new disk and perform a reconguration boot so that the system will recognize the disk.

SPARC: How to Connect a System Disk and Boot on page 218

SPARC: How to Connect a Secondary Disk and Boot on page 219

217

Task

Description

For Instructions

2. Create slices and label the disk.

Create disk slices and label the disk if the disk manufacturer has not already done so.

SPARC: How to Create Disk Slices and Label a Disk on page 220

3. Create le systems.

Create UFS le systems on the SPARC: How to Create a UFS disk slices by using the newfs File System on page 225 command. You must create the root (/) or /usr le system, or both, for a system disk. Restore the root (/) or /usr le system, or both, on the system disk. If necessary, restore le systems on the secondary disk. System Disk Only. Install the boot block on the root (/) le system so that the system can boot. Chapter 27

4. Restore le systems.

5. Install boot block.

SPARC: How to Install a Boot Block on a System Disk on page 226

SPARC: Adding a System Disk or a Secondary Disk

A system disk contains the root (/) or /usr le systems, or both. If the disk that contains either of these le systems becomes damaged, you have two ways to recover:
I I

You can reinstall the entire Solaris OS. Or, you can replace the system disk and restore your le systems from a backup medium.

A secondary disk does not contain the root (/) and /usr le systems. A secondary disk usually contains space for user les. You can add a secondary disk to a system for more disk space. Or, you can replace a damaged secondary disk. If you replace a secondary disk on a system, you can restore the old disks data on the new disk.

SPARC: How to Connect a System Disk and Boot

This procedure assumes that the system is shut down. 1. Disconnect the damaged system disk from the system. 2. Ensure that the disk you are adding has a different target number than the other devices on the system.

Steps

218

System Administration Guide: Devices and File Systems June 2005

Typically, a small switch is located at the back of the disk for this purpose. 3. Connect the replacement system disk to the system and check the physical connections. Refer to the disks hardware installation guide for details. 4. Follow the instructions in the following table, depending on whether you are booting from a local Solaris CD or DVD or a remote Solaris CD or DVD from the network.

Boot Type

Action

From a Solaris CD or DVD in a local drive

1. Make sure the Solaris Software 1 CD or the Solaris DVD is in the drive. 2. Boot from the media to single-user mode: ok boot cdrom -s

From the network

Boot from the network to single-user mode: ok boot net -s

After a few minutes, the root prompt (#) is displayed.

More Information

After You Connect a System Disk and Boot ...

After you boot the system, you can create slices and a disk label on the disk. Go to SPARC: How to Create Disk Slices and Label a Disk on page 220.

SPARC: How to Connect a Secondary Disk and Boot

If you are adding a disk with an EFI disk label, see Multiterabyte Disk Support With EFI Disk Label on page 175 for more information.

Steps

1. Become superuser or assume an equivalent role. 2. If the disk type is unsupported by the Solaris software, add the device driver for the disk by following the instructions included with the hardware. For information on creating a format.dat entry for the disk, see How to Create a format.dat Entry on page 208, if necessary. 3. Create the /reconfigure le that will be read when the system is booted.
# touch /reconfigure

The /reconfigure le causes the SunOS software to check for the presence of any newly installed peripheral devices when you power on or boot your system
Chapter 13 SPARC: Adding a Disk (Tasks) 219

later. 4. Shut down the system.

# shutdown -i0 -gn -y

-i0 -gn -y

Changes to run level 0, the power-down state. Noties logged-in users that they have n seconds before the system begins to shut down. Species that the command should run without user intervention.

The ok prompt is displayed after the Solaris OS is shut down. 5. Turn off the power to the system and all external peripheral devices. 6. Ensure that the disk you are adding has a different target number than the other devices on the system. Typically, a small switch is located at the back of the disk for this purpose. 7. Connect the disk to the system and check the physical connections. Refer to the disks hardware installation guide for details. 8. Turn on the power to all external peripheral devices. 9. Turn on the power to the system. The system boots and displays the login prompt.
More Information

After You Connect a Secondary Disk and Boot ...

After you boot the system, you can create slices and a disk label on the disk. Go to SPARC: How to Create Disk Slices and Label a Disk on page 220.

SPARC: How to Create Disk Slices and Label a Disk

1. Become superuser or assume an equivalent role. 2. Invoke the format utility.
# format

Steps

A numbered list of available disks is displayed. For more information, see format(1M). 3. Type the number of the disk that you want to repartition.
Specify disk (enter its number): disk-number 220 System Administration Guide: Devices and File Systems June 2005

disk-number is the number of the disk that you want to repartition. 4. Select the partition menu.
format> partition

5. Display the current partition (slice) table.

partition> print

6. Start the modication process.

partition> modify

7. Set the disk to all free hog.

Choose base (enter number) [0]?1

For more information about the free hog slice, see Using the Free Hog Slice on page 192. 8. Create a new partition table by answering y when prompted to continue.
Do you wish to continue creating a new partition table based on above table[yes]? y

9. Identify the free hog partition (slice) and the sizes of the slices when prompted. When adding a system disk, you must set up slices for:
I I

root (slice 0) and swap (slice 1) /usr (slice 6)

After you identify the slices, the new partition table is displayed. For an example of creating disk slices, see Example 131. 10. Make the displayed partition table the current partition table by answering y when prompted.
Okay to make this the current partition table[yes]? y

If you do not want the current partition table and you want to change it, answer no and go to Step 6. 11. Name the partition table.
Enter table name (remember quotes): "partition-name"

where partition-name is the name for the new partition table. 12. Label the disk with the new partition table after you have nished allocating slices on the new disk.
Ready to label disk, continue? yes

Chapter 13 SPARC: Adding a Disk (Tasks)

221

13. Quit the partition menu.

partition> q

14. Verify the disk label.

format> verify

15. Exit the format utility.

format> q

Example 131

SPARC: Creating Disk Slices and Labeling a System Disk

The following example shows the format utility being used to divide a 18-Gbyte disk into three slices: one slice for the root (/) le system, one slice for the swap area, and one slice for the /usr le system.
# format AVAILABLE DISK SELECTIONS: 0. /dev/rdsk/c1t0d0s0 <SUN18G cyl 7506 alt /sbus@2,0/QLGC,isp@2,10000/sd@0,0 1. /dev/rdsk/c1t1d0s0 <SUN18G cyl 7506 alt /sbus@2,0/QLGC,isp@2,10000/sd@1,0 2. /dev/rdsk/c1t8d0s0 <SUN18G cyl 7506 alt /sbus@2,0/QLGC,isp@2,10000/sd@8,0 3. /dev/rdsk/c1t9d0s0 <SUN18G cyl 7506 alt /sbus@2,0/QLGC,isp@2,10000/sd@9,0 Specify disk (enter its number): 0 selecting c1t0d0 [disk formatted] format> partition partition> print partition> modify Select partitioning base: 0. Current partition table (original) 1. All Free Hog Part Tag Flag Cylinders Size 0 root wm 0 0 1 swap wu 0 0 2 backup wu 0 - 7505 16.86GB 3 unassigned wm 0 0 4 unassigned wm 0 0 5 unassigned wm 0 0 6 usr wm 0 0 7 unassigned wm 0 0 Choose base (enter number) [0]? 1 table based on above table[yes]? yes Free Hog partition[6]? 6 Enter size of partition 0 [0b, 0c, Enter size of partition 1 [0b, 0c, Enter size of partition 3 [0b, 0c, Enter size of partition 4 [0b, 0c,

2 hd 19 sec 248> 2 hd 19 sec 248> 2 hd 19 sec 248> 2 hd 19 sec 248>

Blocks (0/0/0) 0 (0/0/0) 0 (7506/0/0) 35368272 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0

0.00mb, 0.00mb, 0.00mb, 0.00mb,

0.00gb]: 4gb 0.00gb]: 4gb 0.00gb]: 0.00gb]:

222

System Administration Guide: Devices and File Systems June 2005

Enter size of partition Enter size of partition Part Tag Flag 0 root wm 1 swap wu 2 backup wu 3 unassigned wm 4 unassigned wm 5 unassigned wm 6 usr wm 7 unassigned wm

5 [0b, 0c, 0.00mb, 0.00gb]: 7 [0b, 0c, 0.00mb, 0.00gb]: Cylinders Size Blocks 0 - 1780 4.00GB (1781/0/0) 8392072 1781 - 3561 4.00GB (1781/0/0) 8392072 0 - 7505 16.86GB (7506/0/0) 35368272 0 0 (0/0/0) 0 0 0 (0/0/0) 0 0 0 (0/0/0) 0 3562 - 7505 8.86GB (3944/0/0) 18584128 0 0 (0/0/0) 0

Okay to make this the current partition table[yes]? yes Enter table name (remember quotes): "disk0" Ready to label disk, continue? yes partition> quit format> verify format> quit

Example 132

SPARC: Creating Disk Slices and Labeling a Secondary Disk

The following example shows the format utility being used to divide a 18-Gbyte disk into one slice for the /export/home le system.
# format /dev/rdsk/c1* AVAILABLE DISK SELECTIONS: 0. /dev/rdsk/c1t0d0s0 <SUN18G cyl 7506 alt /sbus@2,0/QLGC,isp@2,10000/sd@0,0 1. /dev/rdsk/c1t1d0s0 <SUN18G cyl 7506 alt /sbus@2,0/QLGC,isp@2,10000/sd@1,0 2. /dev/rdsk/c1t8d0s0 <SUN18G cyl 7506 alt /sbus@2,0/QLGC,isp@2,10000/sd@8,0 3. /dev/rdsk/c1t9d0s0 <SUN18G cyl 7506 alt /sbus@2,0/QLGC,isp@2,10000/sd@9,0 Specify disk (enter its number): 1 selecting c1t1d0 [disk formatted] format> partition partition> print partition> modify Select partitioning base: 0. Current partition table (original) 1. All Free Hog Choose base (enter number) [0]? 1 Part Tag Flag Cylinders Size 0 root wm 0 0 1 swap wu 0 0 2 backup wu 0 - 7505 16.86GB 3 unassigned wm 0 0 4 unassigned wm 0 0 5 unassigned wm 0 0 6 usr wm 0 0 7 unassigned wm 0 0

2 hd 19 sec 248> 2 hd 19 sec 248> 2 hd 19 sec 248> 2 hd 19 sec 248>

Blocks (0/0/0) 0 (0/0/0) 0 (7506/0/0) 35368272 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0 (0/0/0) 0

Chapter 13 SPARC: Adding a Disk (Tasks)

223

Do you wish to continue creating a new partition table based on above table[yes]? y Free Hog partition[6]? 7 Enter size of partition 0 [0b, 0c, 0.00mb, 0.00gb]: Enter size of partition 1 [0b, 0c, 0.00mb, 0.00gb]: Enter size of partition 3 [0b, 0c, 0.00mb, 0.00gb]: Enter size of partition 4 [0b, 0c, 0.00mb, 0.00gb]: Enter size of partition 5 [0b, 0c, 0.00mb, 0.00gb]: Enter size of partition 6 [0b, 0c, 0.00mb, 0.00gb]: Part Tag Flag Cylinders Size Blocks 0 root wm 0 0 (0/0/0) 0 1 swap wu 0 0 (0/0/0) 0 2 backup wu 0 - 7505 16.86GB (7506/0/0) 35368272 3 unassigned wm 0 0 (0/0/0) 0 4 unassigned wm 0 0 (0/0/0) 0 5 unassigned wm 0 0 (0/0/0) 0 6 usr wm 0 0 (0/0/0) 0 7 unassigned wm 0 - 7505 16.86GB (7506/0/0) 35368272 Okay to make this the current partition table[yes]? yes Enter table name (remember quotes): "home" Ready to label disk, continue? y partition> q format> verify format> q #

The following example shows how to use the format utility to divide a 1.15 terabyte disk with an EFI label into three slices.
# format . . . partition> modify Select partitioning base: 0. Current partition table (original) 1. All Free Hog Choose base (enter number) [0]? 1 Part Tag Flag First Sector Size 0 root wm 0 0 1 usr wm 0 0 2 unassigned wm 0 0 3 unassigned wm 0 0 4 unassigned wm 0 0 5 unassigned wm 0 0 6 usr wm 0 0 8 reserved wm 2576924638 8.00MB Do you wish to continue creating a new partition table based on above table[yes]? y Free Hog partition[6]? 4 Enter size of partition 0 [0b, 34e, 0mb, 0gb, 0tb]: Enter size of partition 1 [0b, 34e, 0mb, 0gb, 0tb]: Enter size of partition 2 [0b, 34e, 0mb, 0gb, 0tb]: 400gb

Last Sector 0 0 0 0 0 0 0 2576941021

224

System Administration Guide: Devices and File Systems June 2005

Enter size of partition 3 [0b, 838860834e, 0mb, 0gb, 0tb]: 400gb Enter size of partition 5 [0b, 1677721634e, 0mb, 0gb, 0tb]: Enter size of partition 6 [0b, 1677721634e, 0mb, 0gb, 0tb]: Part Tag Flag First Sector Size Last Sector 0 unassigned wm 0 0 0 1 unassigned wm 0 0 0 2 usr wm 34 400.00GB 838860833 3 usr wm 838860834 400.00GB 1677721633 4 usr wm 1677721634 428.77GB 2576924637 5 unassigned wm 0 0 0 6 unassigned wm 0 0 0 8 reserved wm 2576924638 8.00MB 2576941021 Ready to label disk, continue? yes partition> q

More Information

After You Create Disk Slices and Label a Disk ...

After you create disk slices and label the disk, you can create le systems on the disk. Go to SPARC: How to Create a UFS File System on page 225.

M
Steps

SPARC: How to Create a UFS File System

1. Become superuser or assume an equivalent role. 2. Create a le system for each slice.
# newfs /dev/rdsk/cwtxdysz

where /dev/rdsk/cwtxdysx is the raw device for the le system to be created. For more information about the newfs command, see Chapter 18 or newfs(1M). 3. Verify the new le system by mounting it.
# mount /dev/dsk/cwtxdysz /mnt # ls lost+found

More Information

After Creating a UFS File System ...

I

System Disk You need to restore the root (/) and /usr le systems on the disk.
I I

Go to Chapter 27. After the root (/) and /usr le systems are restored, install the boot block. Go to SPARC: How to Install a Boot Block on a System Disk on page 226.

Chapter 13 SPARC: Adding a Disk (Tasks)

225

Secondary Disk You might need to restore le systems on the new disk. Go to Chapter 27. If you are not restoring le systems on the new disk, you are nished adding a secondary disk. For information on making the le systems available to users, see Chapter 19.

SPARC: How to Install a Boot Block on a System Disk

1. Become superuser or assume an equivalent role. 2. Install a boot block on the system disk.
# installboot /usr/platform/uname -i/lib/fs/ufs/bootblk /dev/rdsk/cwtxdys0

Steps

/usr/platform/uname -i/lib/fs /ufs/bootblk Is the boot block code. /dev/rdsk/cwtxdys0 Is the raw device of the root (/) le system. For more information, see installboot(1M). 3. Verify that the boot blocks are installed by rebooting the system to run level 3.
# init 6

Example 133

SPARC: Installing a Boot Block on a System Disk

The following example shows how to install the boot block on an Ultra 10 system.
# installboot /usr/platform/sun4u/lib/fs/ufs/bootblk /dev/rdsk/c0t0d0s0

226

System Administration Guide: Devices and File Systems June 2005

CHAPTER

14

x86: Adding a Disk (Tasks)

This chapter describes how to add a disk to an x86 based system. For information on the procedures associated with adding a disk to an x86 based system, see x86: Adding a System Disk or a Secondary Disk (Task Map) on page 227. For overview information about disk management, see Chapter 11. For step-by-step instructions on adding a disk to a SPARC based system, see Chapter 13.

x86: Adding a System Disk or a Secondary Disk (Task Map)

The following task map identies the procedures for adding a disk to an x86 based system.

Task

Description

For Instructions

1. Connect the disk and boot.

System Disk Connect the new disk and boot from a local or remote Solaris CD or DVD. Secondary Disk Connect the new disk and perform a reconguration boot so that the system will recognize the disk.

x86: How to Connect a System Disk and Boot on page 229

x86: How to Connect a Secondary Disk and Boot on page 231

227

Task

Description

For Instructions

2. (Optional) Change the fdisk partition identier.

The Solaris 10 fdisk partition How to Change the Solaris identier on x86 systems has fdisk Identier on page 230 been changed from 130 (0x82) to 191 (0xbf). You can use a new fdisk menu option to switch back and forth between the new and old identier.

3. Create slices and label the disk.

Create disk slices and label the disk if the disk manufacturer has not already done so.

x86: How to Create a Solaris fdisk Partition on page 233 and x86: How to Create Disk Slices and Label a Disk on page 238

4. Create le systems.

Create UFS le systems on the x86: How to Create File Systems on page 239 disk slices with the newfs command. You must create the root (/) or /usr le system (or both) for a system disk. Restore the root (/) or /usr le system (or both) on the system disk. If necessary, restore le systems on the secondary disk. System Disk Only. Install the boot block on the root (/) le system so that the system can boot. Chapter 27

5. Restore le systems.

6. Install boot block.

x86: How to Install a Boot Block on a System Disk on page 240

x86: Adding a System Disk or a Secondary Disk

A system disk contains the root (/) or /usr le systems, or both. If the disk that contains either of these le systems becomes damaged, you have two ways to recover:
I I

You can reinstall the entire Solaris OS. Or, you can replace the system disk and restore your le systems from a backup medium.

228

System Administration Guide: Devices and File Systems June 2005

A secondary disk doesnt contain the root (/) and /usr le systems. A secondary disk usually contains space for user les. You can add a secondary disk to a system for more disk space. Or, you can replace a damaged secondary disk. If you replace a secondary disk on a system, you can restore the old disks data on the new disk.

x86: How to Connect a System Disk and Boot

This procedure assumes that the system is down. 1. Disconnect the damaged system disk from the system. 2. Ensure that the disk you are adding has a different target number than the other devices on the system. Typically, a small switch is located at the back of the disk for this purpose. 3. Connect the replacement system disk to the system, and check the physical connections. Refer to the disks hardware installation guide for details. 4. Boot the system. If you are booting from the network, skip steps ab. a. If you are booting from a local Solaris CD or DVD, insert the Solaris Software 1 CD or the Solaris DVD into the drive. b. Insert the Solaris boot diskette into the primary diskette drive (DOS drive A). c. Press any key to reboot the system if the system displays the Type any key to continue prompt. Or, use the reset button to restart the system if the system is shut down. The Boot Solaris screen is displayed after a few minutes. d. Select the CD-ROM drive or net(work) as the boot device from the Boot Solaris screen. The Current Boot Parameters screen is displayed. e. Boot the system to single-user mode.
Select the type of installation: b -s

Steps

After a few minutes, the root prompt (#) is displayed.

More Information

After You Connect a System Disk and Boot ...

After you boot the system and the disk is less than 1 terabyte, you can create an fdisk partition. Go to x86: How to Create a Solaris fdisk Partition on page 233.
Chapter 14 x86: Adding a Disk (Tasks) 229

How to Change the Solaris fdisk Identier

A new fdisk menu option enables you to switch back and forth between the new and old identier. The fdisk identier can be changed even when the le system that is contained in the partition is mounted. Two type values in the fdisk menu reect the old and new identiers as follows:
I I

Solaris identies 0x82 Solaris2 identies 0xbf

Steps

1. Become superuser. 2. Display the current fdisk identier. For example:

Total disk size is 39890 cylinders Cylinder size is 4032 (512 byte) blocks Cylinders Start End Length ===== === ====== 1 6 6 7 39889 39883

Partition ========= 1 2

Status ====== Active

Type ============ x86 Boot Solaris2

% === 0 100

3. Select option 4 from the fdisk menu to change the fdisk partition identier back to 0x82
SELECT ONE OF THE FOLLOWING: 1. Create a partition 2. Specify the active partition 3. Delete a partition 4. Change between Solaris and Solaris2 Partition IDs 5. Exit (update disk configuration and exit) 6. Cancel (exit without updating disk configuration) Enter Selection: 4

4. Select option 5 to update your disk conguration and exit. 5. If necessary, select option 4 from the fdisk menu to change the fdisk partition identier back to 0xbf. For example:
Total disk size is 39890 cylinders Cylinder size is 4032 (512 byte) blocks Cylinders Start End Length ===== === ====== 1 6 6 7 39889 39883

Partition ========= 1 2

Status ====== Active

Type ============ x86 Boot Solaris

% === 0 100

SELECT ONE OF THE FOLLOWING: 230 System Administration Guide: Devices and File Systems June 2005

1. Create a partition 2. Specify the active partition 3. Delete a partition 4. Change between Solaris and Solaris2 Partition IDs 5. Exit (update disk configuration and exit) 6. Cancel (exit without updating disk configuration) Enter Selection: 4

6. Select option 5 to update your disk conguration and exit.

x86: How to Connect a Secondary Disk and Boot

If you are adding a disk with an EFI disk label on an x64 system, see Multiterabyte Disk Support With EFI Disk Label on page 175 for more information.

Steps

1. Become superuser or assume an equivalent role. 2. If the disk is unsupported by the Solaris software, add the device driver for the disk by following the instructions included with the hardware. 3. Create the /reconfigure le that will be read when the system is booted.
# touch /reconfigure

The /reconfigure le causes the SunOS software to check for the presence of any newly installed peripheral devices when you power on or boot your system later. 4. Shut down the system.
# shutdown -i0 -gn -y

-i0 -gn -y

Brings the system down to run level 0, the power-down state. Noties logged-in users that they have n seconds before the system begins to shut down. Species that the command should run without user intervention.

The Type any key to continue prompt is displayed. 5. Turn off the power to the system and all external peripheral devices. 6. Ensure that the disk you are adding has a different target number than the other devices on the system. Typically, a small switch is located at the back of the disk for this purpose. 7. Connect the disk to the system and check the physical connections. Refer to the disks hardware installation guide for details. 8. Turn on the power to all external peripheral devices.
Chapter 14 x86: Adding a Disk (Tasks) 231

9. Turn on the power to the system. The system boots and displays the login prompt.
More Information

After You Connect a Secondary Disk and Boot ...

After you boot the system and the disk is less than 1 terabyte, you can create an fdisk partition. Go to x86: How to Create a Solaris fdisk Partition on page 233.

x86: Guidelines for Creating an fdisk Partition

Follow these guidelines when you set up one or more fdisk partitions.
I

The fdisk command is not intended for disks greater than 1 terabyte and cannot be used on disks with an EFI label. The disk can be divided into a maximum of four fdisk partitions. One of partitions must be a Solaris partition. The Solaris partition must be made the active partition on the disk. The active partition is partition whose operating system will be booted by default at system startup. Solaris fdisk partitions must begin on cylinder boundaries. Solaris fdisk partitions must begin at cylinder 1, not cylinder 0, on the rst disk because additional boot information, including the master boot record, is written in sector 0. The Solaris fdisk partition can be the entire disk. Or, you might want to make it smaller to allow room for a DOS partition. You can also make a new fdisk partition on a disk without disturbing existing partitions (if sufficient space is available) to create a new partition.

I I

x86 only Solaris slices are also called partitions. Certain interfaces might refer to a slice as a partition.

fdisk partitions are supported only on x86 based systems. To avoid confusion, Solaris documentation tries to distinguish between fdisk partitions and the entities within the Solaris fdisk partition. These entities might be called slices or partitions.

232

System Administration Guide: Devices and File Systems June 2005

M
Before You Begin

x86: How to Create a Solaris fdisk Partition

If you need information about fdisk partitions, see x86: Guidelines for Creating an fdisk Partition on page 232. 1. Become superuser or assume an equivalent role. 2. Invoke the format utility.
# format

Steps

A numbered list of disks is displayed. For more information, see format(1M). 3. Type the number of the disk on which to create a Solaris fdisk partition.
Specify disk (enter its number): disk-number

where disk-number is the number of the disk on which you want to create a Solaris fdisk partition. 4. Select the fdisk menu.
format> fdisk

The fdisk menu that is displayed depends upon whether the disk has existing fdisk partitions. Determine the next step by using the following table.

Task

Go To

For More Information

Create a Solaris fdisk partition to span the entire disk.

Step 5

Example 141 Example 142

Create a Solaris fdisk partition and Step 6 preserve one or more existing non Solaris fdisk partitions. Create a Solaris fdisk partition and Step 6 one or more additional non Solaris fdisk partition.

Example 143

5. Create and activate a Solaris fdisk partition that spans the entire disk by specifying y at the prompt. Then, go to step 13.
No fdisk table exists. The default partition for the disk is: a 100% "SOLARIS System" partition Type "y" to accept the default partition, otherwise type "n" to edit the partition table. y

Chapter 14 x86: Adding a Disk (Tasks)

233

6. Specify n at the prompt if you do not want the Solaris fdisk partition to span the entire disk.
Type "y" to accept the default partition, otherwise type "n" to edit the partition table. n Total disk size is 3498 cylinders Cylinder size is 1199 (512 byte) blocks Cylinders Partition Status Type Start End Length % ========= ====== ======== ===== === ====== === SELECT ONE OF THE FOLLOWING: 1. Create a partition 2. Specify the active partition 3. Delete a partition 4. Change between Solaris and Solaris2 Partition IDs 5. Exit (update disk configuration and exit) 6. Cancel (exit without updating disk configuration) Enter Selection:

7. Select option 1, Create a partition, to create an fdisk partition.

Enter Selection: 1

8. Create a Solaris fdisk partition by selecting 1(=Solaris2).

Indicate the type of partition you want to create 1=SOLARIS2 2=UNIX 3=PCIXOS 4=Other 5=DOS12 6=DOS16 7=DOSEXT 8=DOSBIG 9=DOS16LBA A=x86 Boot B=Diagnostic C=FAT32 D=FAT32LBA E=DOSEXTLBA F=EFI 0=Exit? 1

9. Identify the percentage of the disk to be reserved for the Solaris fdisk partition. Keep in mind the size of any existing fdisk partitions when you calculate this percentage.
Specify the percentage of disk to use for this partition (or type "c" to specify the size in cylinders). nn

10. Activate the Solaris fdisk partition by typing y at the prompt.

Should this to become the active partition? If yes, it will be activated each time the computer is reset or turned on. Please type "y" or "n". y

The Enter Selection prompt is displayed after the fdisk partition is activated. 11. Select option 1, Create a partition, to create another fdisk partition. See steps 810 for instructions on creating an fdisk partition. 12. Update the disk conguration, and exit the fdisk menu from the selection menu.
Selection: 5 234 System Administration Guide: Devices and File Systems June 2005

13. Relabel the disk by using the label command.

format> label Ready to label disk, continue? yes format>

14. Quit the format utility.

format> quit

Example 141

x86: Creating a Solaris fdisk Partition That Spans the Entire Drive
The following example uses the format utilitys fdisk option to create a Solaris fdisk partition that spans the entire drive.
# format Searching for disks...done AVAILABLE DISK SELECTIONS: 0. c0d0 <DEFAULT cyl 2466 alt 2 hd 16 sec 63> /pci@0,0/pci-ide@7,1/ide@0/cmdk@0,0 1. c0d1 <DEFAULT cyl 522 alt 2 hd 32 sec 63> /pci@0,0/pci-ide@7,1/ide@0/cmdk@1,0 2. c1d0 <DEFAULT cyl 13102 alt 2 hd 16 sec 63> /pci@0,0/pci-ide@7,1/ide@1/cmdk@0,0 Specify disk (enter its number): 0 selecting c0d0 Controller working list found [disk formatted] format> fdisk No fdisk table exists. The default partitioning for your disk is: a 100% "SOLARIS System" partition. Type "y" to accept the default partition, otherwise type "n" to edit the partition table. y format> label Ready to label disk, continue? yes format> quit

Example 142

x86: Creating a Solaris fdisk Partition While Preserving an Existing fdisk Partition
The following example shows how to create a Solaris fdisk partition on a disk that has an existing DOS-BIG fdisk partition.
format> fdisk Total disk size is 3498 cylinders Cylinder size is 1199 (512 byte) blocks Cylinders Start End Length ===== === ======

Partition =========

Status ======

Type ============

% === 235

Chapter 14 x86: Adding a Disk (Tasks)

1 Active DOS-BIG 1 699 699 SELECT ONE OF THE FOLLOWING: 1. Create a partition 2. Specify the active partition 3. Delete a partition 4. Change between Solaris and Solaris2 Partition IDs 5. Exit (update disk configuration and exit) 6. Cancel (exit without updating disk configuration) Enter Selection: 1 Indicate the type of partition you want to create 1=SOLARIS2 2=UNIX 3=PCIXOS 4=Other 5=DOS12 6=DOS16 7=DOSEXT 8=DOSBIG 9=DOS16LBA A=x86 Boot B=Diagnostic C=FAT32 D=FAT32LBA E=DOSEXTLBA F=EFI 0=Exit?1 Indicate the percentage of the disk you want this partition to use (or enter "c" to specify in cylinders). 80 Should this become the active partition? If yes, it will be activated each time the computer is or turned on. Please type "y" or "n". y Total disk size is 3498 cylinders Cylinder size is 1199 (512 byte) blocks Cylinders Partition Status Type Start End Length ========= ====== ============ ===== === ====== 1 DOS-BIG 1 699 699 2 Active Solaris2 700 3497 2798 SELECT ONE OF THE FOLLOWING: 1. Create a partition 2. Specify the active partition 3. Delete a partition 4. Change between Solaris and Solaris2 Partition IDs 5. Exit (update disk configuration and exit) 6. Cancel (exit without updating disk configuration) Enter Selection:5 Partition 2 is now the active partition format> label Ready to label disk, continue? yes format> q

20

% === 20 80

Example 143

x86: Creating a Solaris fdisk Partition and an Additional fdisk Partition

This following example shows how to create a Solaris fdisk partition and a DOSBIG fdisk partition.
format> fdisk No fdisk table exists. The default partitioning for your disk is: a 100% "SOLARIS System" partition. Type "y" to accept the default partition, otherwise type "n" to edit the partition table. n Total disk size is 3498 cylinders Cylinder size is 1199 (512 byte) blocks

236

System Administration Guide: Devices and File Systems June 2005

Cylinders Partition Status Type Start End Length ========= ====== ============ ===== === ====== SELECT ONE OF THE FOLLOWING: 1. Create a partition 2. Specify the active partition 3. Delete a partition 4. Change between Solaris and Solaris2 Partition IDs 5. Exit (update disk configuration and exit) 6. Cancel (exit without updating disk configuration) Enter Selection: 1 Indicate the type of partition you want to create 1=SOLARIS2 2=UNIX 3=PCIXOS 4=Other 5=DOS12 6=DOS16 7=DOSEXT 8=DOSBIG 9=DOS16LBA A=x86 Boot B=Diagnostic C=FAT32 D=FAT32LBA E=DOSEXTLBA F=EFI 0=Exit? 8 Specify the percentage of disk to use for this partition (or type "c" to specify the size in cylinders)20 Should this to become the Active partition? If yes, it will be activated each time the computer is reset or turned on. again. Please type "y" or "n". n Total disk size is 3498 cylinders Cylinder size is 1199 (512 byte) blocks Cylinders Partition Status Type Start End Length ========= ====== ============ ===== === ====== 1 DOS-BIG 1 699 699 SELECT ONE OF THE FOLLOWING: 1. Create a partition 2. Specify the active partition 3. Delete a partition 4. Change between Solaris and Solaris2 Partition IDs 5. Exit (update disk configuration and exit) 6. Cancel (exit without updating disk configuration) Enter Selection: 1 Indicate the type of partition you want to create 1=SOLARIS2 2=UNIX 3=PCIXOS 4=Other 5=DOS12 6=DOS16 7=DOSEXT 8=DOSBIG 9=DOS16LBA A=x86 Boot B=Diagnostic C=FAT32 D=FAT32LBA E=DOSEXTLBA F=EFI 0=Exit? 1 Indicate the percentage of the disk you want this partition to use (or enter "c" to specify in cylinders). 80 Should this become the active partition? If yes, it will be activated each time the computer is reset or turned on. Please type "y" or "n". y Total disk size is 3498 cylinders Cylinder size is 1199 (512 byte) blocks Cylinders Partition Status Type Start End Length ========= ====== ============ ===== === ====== 1 DOS-BIG 1 699 699 2 Active Solaris2 700 3497 2798 SELECT ONE OF THE FOLLOWING: 1. Create a partition 2. Specify the active partition

% ===

% === 20

% === 20 80

Chapter 14 x86: Adding a Disk (Tasks)

237

3. Delete a partition 4. Change between Solaris and Solaris2 Partition IDs 5. Exit (update disk configuration and exit) 6. Cancel (exit without updating disk configuration) Enter Selection: 5 Partition 2 is now the Active partition format> q

More Information

After You Create a Solaris fdisk Partition ...

After you create a Solaris fdisk partition on the disk, you can create slices on the disk. Go to x86: How to Create Disk Slices and Label a Disk on page 238

M
Steps

x86: How to Create Disk Slices and Label a Disk

1. Become superuser or assume an equivalent role. 2. Invoke the format utility.
# format

A numbered list of disks is displayed. 3. Type the number of the disk that you want to repartition.
Specify disk (enter its number): disk-number

where disk-number is the number of the disk that you want to repartition. 4. Select the partition menu.
format> partition

5. Display the current partition (slice) table.

partition> print

6. Start the modication process.

partition> modify

7. Set the disk to all free hog.

Choose base (enter number) [0]? 1

For more information about the free hog slice, see Using the Free Hog Slice on page 192. 8. Create a new partition table by answering yes when prompted to continue.
Do you wish to continue creating a new partition table based on above table[yes]? yes

9. Identify the free hog partition (slice) and the sizes of the slices when prompted.
238 System Administration Guide: Devices and File Systems June 2005

When adding a system disk, you must set up slices for the following:
I I

root (slice 0) and swap (slice 1) and/or /usr (slice 6)

After you identify the slices, the new partition table is displayed. 10. Make the displayed partition table the current partition table by answering yes when prompted.
Okay to make this the current partition table[yes]? yes

If you dont want the current partition table and you want to change it, answer no and go to Step 6. 11. Name the partition table.
Enter table name (remember quotes): "partition-name"

where partition-name is the name for the new partition table. 12. Label the disk with the new partition table after you have nished allocating slices on the new disk.
Ready to label disk, continue? yes

13. Quit the partition menu.

partition> quit

14. Verify the new disk label.

format> verify

15. Exit the format utility.

format> quit

More Information

After You Create Disk Slices and Label a Disk ...

After you create disk slices and label the disk, you can create le systems on the disk. Go to x86: How to Create File Systems on page 239.

M
Steps

x86: How to Create File Systems

1. Become superuser or assume an equivalent role. 2. Create a le system for each slice.
# newfs /dev/rdsk/cwtxdysz

where /dev/rdsk/cwtxdysz is the raw device for the le system to be created. For more information about the newfs command, see Chapter 18 or newfs(1M).
Chapter 14 x86: Adding a Disk (Tasks) 239

3. Verify the new le system by mounting.

# mount /dev/dsk/cwtxdysz /mnt # ls /mnt lost+found

More Information

After You Create File Systems ...

I

System Disk You need to restore the root (/) and /usr le systems on the disk.
I I

Go to Chapter 27. After the root (/) and /usr le systems are restored, install the boot block. Go to x86: How to Install a Boot Block on a System Disk on page 240.

Secondary Disk You might need to restore le systems on the new disk. Go to Chapter 27. If you are not restoring le systems on the new disk, you are nished adding a secondary disk. For information on making the le systems available to users, see Chapter 19.

M
Steps

x86: How to Install a Boot Block on a System Disk

1. Become superuser or assume an equivalent role. 2. Install the boot block on the system disk.
# installboot /usr/platform/uname -i/lib/fs/ufs/pboot /usr/platform/ uname -i /lib/fs/ufs/bootblk /dev/rdsk/cwtxdys2

/usr/platform/uname -i/lib/fs/ufs/pboot Is the partition boot le. /usr/platform/uname -i/lib/fs/ufs/bootblk Is the boot block code. /dev/rdsk/cwtxdys2 Is the raw device name that represents the entire disk. 3. Verify that the boot blocks are installed by rebooting the system to run level 3.
# init 6

Example 144

x86: Installing a Boot Block on a System Disk

The following example shows how to install the boot block on an x86 system.
# installboot /usr/platform/i86pc/lib/fs/ufs/pboot /usr/platform/i86pc/lib/fs/ufs/bootblk /dev/rdsk/c0t6d0s2

240

System Administration Guide: Devices and File Systems June 2005

CHAPTER

15

Conguring Solaris iSCSI Initiators (Tasks)

This chapter describes how to congure Solaris iSCSI initiators in the Solaris 10 7/05 time frame. For information on the procedures associated with conguring iSCSI initiators, see Setting Up Solaris iSCSI Initiators (Task Map) on page 243.

The iSCSI Technology (Overview)

iSCSI is an acronym for Internet SCSI (Small Computer System Interface), an Internet Protocol (IP)-based storage networking standard for linking data storage subsystems. This networking standard was developed by the Internet Engineering Task Force (IETF). For more information about the iSCSI technology, see RFC 3720: http://www.ietf.org/rfc/rfc3720.txt By carrying SCSI commands over IP networks, the iSCSI protocol enables you to access block devices from across the network as if they were connected to the local system. If you want to use storage devices in your existing TCP/IP network, the following solutions are available:
I

iSCSI block devices or tape Translates SCSI commands and data from the block level into IP packets. The advantage of using iSCSI in your network is when you need to have block-level access between one system and the target device, such as a tape device or a database. Access to a block-level device is not locked so that you could not have multiple users or systems accessing a block-level device such as an iSCSI target device. NFS Transfers le data over IP. The advantage of using NFS in your network is that you can share le data across many systems. Access to le data is locked appropriately when many users are accessing data that is available in an NFS environment.
241

Here are the benets of using Solaris iSCSI initiators:

I

The iSCSI protocol runs across existing Ethernet networks.

I

I I I

You can use any supported network interface card (NIC), Ethernet hub or switch. One IP port can handle multiple iSCSI target devices. You can use existing infrastructure and management tools for IP networks.

There is no upper limit on the maximum number of congured iSCSI target devices. The protocol an be used to connect to Fibre Channel or iSCSI Storage Area Network (SAN) environments with the appropriate hardware.

Here are the current limitations or restrictions of using the Solaris iSCSI initiator software:
I I I I

No support for iSCSI devices that use SLP or iSNS is currently available. No boot support for iSCSI devices is currently available. Do not congure iSCSI targets as dump devices. iSCSI supports multiple connections per session, but the current Solaris implementation only supports a single connection per session. For more information, see RFC 3720. You should consider the impact of transferring large amounts of data over your existing network.

iSCSI Software and Hardware Requirements

I I I

iSCSI target software and devices The Solaris 10 7/05 release, the Solaris Express 02/05, or a later release The following software packages:
I I

SUNWiscsiu Sun iSCSI Device Driver (root) SUNWiscsir Sun iSCSI Management Utilities (usr)

Note The Solaris iSCSI technology includes the iSCSI initiator software only.

Any supported NIC

242

System Administration Guide: Devices and File Systems June 2005

Setting Up Solaris iSCSI Initiators (Task Map)

Task Description For Instructions

1. Identify the iSCSI software and hardware requirements.

Identify the software and hardware requirements for setting up an iSCSI-based storage network. Connect and set up your iSCSI target devices. Decide whether you want to use authentication in your Solaris iSCSI conguration: Consider using unidirectional CHAP or bidirectional CHAP Consider using a third-party RADIUS server to simplify CHAP management

iSCSI Software and Hardware Requirements on page 242 See your vendors iSCSI target device documentation for setup instructions

2. Set up your iSCSI target devices. 3. (Optional) Set up authentication in your Solaris iSCSI conguration.

How to Congure CHAP Authentication for Your iSCSI Conguration on page 246 How to Congure RADIUS for Your iSCSI Conguration on page 247 How to Prepare for a Solaris iSCSI Conguration on page 245 How to Remove Discovered iSCSI Targets on page 249 Monitoring Your iSCSI Conguration on page 250

4. Congure the iSCSI target discovery. 5. (Optional) Remove discovered iSCSI targets . 6. Monitor your iSCSI conguration. 7. (Optional) Modify your iSCSI conguration.

Congure the iSCSI target discovery method. You might need to remove a discovered iSCSI target. Monitor your iSCSI conguration with the iscsiadm command.

You might want to change How to Modify iSCSI your iSCSI target settings such Initiator and Target as the header and data digest Parameters on page 252 parameters.

Conguring Solaris iSCSI Initiators

Basically, the steps for conguring your Solaris iSCSI initiators involves the following steps:
I

Identifying the hardware and software requirements

Chapter 15 Conguring Solaris iSCSI Initiators (Tasks) 243

I I I

Conguring your IP network Connecting and setting up your iSCSI target device (Optional) Congure iSCSI authentication between the iSCSI initiator and the iSCSI target, if necessary Conguring the iSCSI target discovery method Creating le systems on your iSCSI disks Monitoring your iSCSI conguration

I I I

The iSCSI conguration information is stored in the /etc/iscsi directory. This information requires no administration.

iSCSI Terminology
Review the following terminology before conguring iSCSI initiators.

Term

Description

Initiator Target device Discovery Discovery method

The driver that initiates SCSI requests from the iSCSI target. Represents the iSCSI storage component. Discovery is the process that presents the initiator with a list of available targets Describes the way in which the iSCSI targets can be found. Two discovery methods are currently available: I SendTargets Potential targets are discovered by using a discovery-address. I Static Static target address is congured.

Conguring Dynamic or Static Target Discovery

Determine whether you want to congure the dynamic iSCSI SendTargets feature or use static iSCSI initiator targets to perform device discovery.
I

Dynamic device discovery If an iSCSI node exposes a large number of targets, such as an iSCSI to Fibre-Channel bridge, you can supply the iSCSI node IP address/port combination and allow the iSCSI initiator to use the SendTargets features to perform the device discovery. Static device discovery If an iSCSI node has a small number of targets or if you want to restrict the targets that the initiator attempts to access, you can statically congure the target-name by using the following static target address naming convention: target-name,target-address[:port-number] You can also determine the static target address from the arrays management tool.

244

System Administration Guide: Devices and File Systems June 2005

The preferred method for target discovery is SendTargets discovery.

Note Do not congure an iSCSI target to be discovered by both static and dynamic

device discovery methods. The consequence of using redundant discovery methods might be slow performance when communicating with the iSCSI target device.

M
Steps

How to Prepare for a Solaris iSCSI Conguration

1. Become superuser. 2. Verify that the iSCSI software packages are installed.
# pkginfo SUNWisciu SUNWiscsir system SUNWiscsiu Sun iSCSI Device Driver (root) system SUNWiscsir Sun iSCSI Management Utilities (usr)

3. Verify that you are running a Solaris release that supports the iSCSI protocol.
I

Solaris Express 2/05 release

% cat /etc/release Nevada nv_07 X86 Copyright 2005 Sun Microsystems, Inc. All Rights Reserved. Use is subject to license terms. Assembled 25 January 2005

Solaris 10 7/05 release

% cat /etc/release Solaris 10 7/05 X86 Copyright 2005 Sun Microsystems, Inc. All Rights Reserved. Use is subject to license terms. Assembled 03 March 2005

Solaris 10 release with the iSCSI patch On a SPARC system:

# showrev -p | grep 119090

On an x86 system:
# showrev -p | grep 119091

4. Conrm that your TCP/IP network is setup. 5. Connect your third-party iSCSI target devices and conrm that they are congured. For example, determine if the iSCSI target device is reachable by using the telnet command to connect to the iSCSI target device using port 3260. If the connection is refused, see Troubleshooting iSCSI Conguration Problems on page 254.
Chapter 15 Conguring Solaris iSCSI Initiators (Tasks) 245

For information about connecting your third-party iSCSI target devices, see your third-party hardware documentation.

Conguring Authentication in Your iSCSI-Based Storage Network

Setting up authentication for your iSCSI devices is optional. In a secure environment, authentication is not required because only trusted initiators can access the targets. In a less secure environment, the target cannot determine if a connection request is truly from a given host. In that case, the target can authenticate an initiator by using the Challenge-Handshake Authentication Protocol (CHAP). CHAP authentication uses the notion of a challenge and response, which means that the target challenges the initiator to prove its identity. For the challenge/response method to work, the target must know the initiators secret key and the initiator must be set up to respond to a challenge. Refer to the array vendors documentation for instructions on setting up the secret key on the array. iSCSI supports unidirectional and bidirectional authentication:
I

Unidirectional authentication enables the target to authenticate the identity of the initiator. Bidirectional authentication adds a second level of security by providing a means for the initiator to authenticate the identity of the target.

M How to Congure CHAP Authentication for Your iSCSI

Conguration
This procedure assumes that you are logged in to the local system where you want to securely access the congured iSCSI target device.
Steps

1. Become superuser. 2. Determine whether you want to congure unidirectional or bidirectional CHAP.
I

Unidirectional authentication enables the target to validate the initiator. This method is the default method. Complete steps 34 only. Bidirectional authentication adds a second level of security by providing a means for the initiator to authenticate the target. Complete steps 56 only.

3. Unidirectional CHAP Set the secret key on the initiator.

246 System Administration Guide: Devices and File Systems June 2005

For example, the following command initiates a dialogue to dene the CHAP secret key.
# iscsiadm modify initiator-node --CHAP-Secret

Note The CHAP secret length must be 16 or more characters.

4. Unidirectional CHAP Enable CHAP authentication on the initiator after the secret has been set.
# iscsiadm modify initiator-node --authentication CHAP

5. Bidirectional CHAP Set the target device secret key on the initiator. For example, the following command initiates a dialogue to dene the CHAP secret key.
# iscsiadm modify target-param --CHAP-Secret eui.5000ABCD78945E2B

6. Bidirectional CHAP Enable bidirectional authentication parameters on the target. For example:
# iscsiadm modify target-param -B enable eui.5000ABCD78945E2B

Using a Third-Party Radius Server to Simplify CHAP Management in Your iSCSI Conguration
You can use a third-party RADIUS server to simplify CHAP secret management. A RADIUS server is a centralized authentication service. While you must still specify the initiators CHAP secret, you are no longer required to specify each targets CHAP secret on each initiator when using bidirectional authentication with a RADIUS server. For more information, see RFC 1994 (CHAP) and RFC 2865 (RADIUS).

M How to Congure RADIUS for Your iSCSI Conguration

Steps

1. Become superuser. 2. Congure the initiator node with the IP address and port (the default port is 1812) of the RADIUS server. For example:
# iscsiadm modify initiator-node --radius-server 10.0.0.72:1812 Chapter 15 Conguring Solaris iSCSI Initiators (Tasks) 247

3. Congure the initiator node with the shared secret of the RADIUS server.
# iscsiadm modify initiator-node --radius-shared-secret

Note The Solaris iSCSI implementation requires that the RADIUS server is congured with a shared secret before the Solaris iSCSI software can interact with the RADIUS server.

4. Enable the RADIUS server.

# iscsiadm modify initiator-node --radius-access enable

Solaris iSCSI and RADIUS Server Error Messages

This section describes the messages that are related to a Solaris iSCSI and RADIUS server conguration with potential solutions for recovery. empty RADIUS shared secret Cause: The RADIUS server is enabled on the initiator but the RADIUS shared secret is not set.
Solution: Congure the initiator with RADIUS shared secret. For more information,

see How to Congure RADIUS for Your iSCSI Conguration on page 247. WARNING: RADIUS packet authentication failed Cause: The initiator failed to authenticate the RADIUS data packet. This error can occur if the shared secret congured on the initiator-node is different from the shared secret on the RADIUS server.
Solution: Recongure the initiator with the correct RADIUS shared secret. For more

information, see How to Congure RADIUS for Your iSCSI Conguration on page 247.

How to Congure iSCSI Target Discovery

This procedure assumes that you are logged in to the local system where you want to congure access to an iSCSI target device.

Steps

1. Become superuser. 2. Congure the SendTargets device discovery method or the static discovery method:
I

Congure the iSCSI device discovery method.

248

System Administration Guide: Devices and File Systems June 2005

For example:
# iscsiadm add discovery-address 10.0.0.1:3260

The iSCSI connection is not initiated until the discovery method is enabled. See the next step.
I

Congure the static discovery method. For example:

# iscsiadm add static-config eui.5000ABCD78945E2B,10.0.0.1

The iSCSI connection is not initiated until the discovery method is enabled. See the next step. 3. Enable the iSCSI target discovery method using one of the following:
I

If you have congured the SendTargets method of discovery, enable SendTargets discovery.
# iscsiadm modify discovery --send-targets enable

If you have congured static targets, enable the static target discovery method.
# iscsiadm modify discovery --static enable

4. Create the iSCSI device links for the local system.

# devfsadm -i iscsi

How to Remove Discovered iSCSI Targets

This optional procedure assumes that you are logged in to the local system where access to an iSCSI target device has already been congured.

Steps

1. Become superuser. 2. (Optional) Disable an iSCSI target discovery method using one of the following:
I

If you need to disable the SendTargets method of discovery, use the following command:
# iscsiadm modify discovery --send-targets disable

If you need to disable the static targets, use the following command:
# iscsiadm modify discovery --static disable

3. Remove an iSCSI device discovery entry:

I

Remove an iSCSI SendTargets discovery entry. For example:

# iscsiadm remove discovery-address 10.0.0.1:3260 Chapter 15 Conguring Solaris iSCSI Initiators (Tasks) 249

Remove a static iSCSI initiator entry. For example:

# iscsiadm remove static-config eui.5000ABCD78945E2B,10.0.0.1

4. Reboot the system if you want to remove the iSCSI target device. The iSCSI target device is still available until the system is rebooted.

Accessing iSCSI Disks

If you want to make the iSCSI drive available on reboot, create the le system, and add an entry to the /etc/vfstab le as you would with a UFS le system on a SCSI device. After the devices have been discovered by the Solaris iSCSI initiator, the login negotiation occurs automatically. The Solaris iSCSI driver determines that the number of LUNs available and creates the device nodes. Then, the iSCSI devices can be treated as any other SCSI device. You can view the iSCSI disks on the local system with the format utility. In the following format output, disks 1,2, and 3 are iSCSI LUNs that are not under MPxIO control. Disks 21 and 22 are iSCSI LUNs under MPxIO control.
# format AVAILABLE DISK SELECTIONS: 0. c0t1d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424> /pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w500000e010685cf1,0 1. c0t2d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424> /pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w500000e0106e3ba1,0 2. c3t0d0 <ABCSTORAGE-100E-00-2.2 cyl 20813 alt 2 hd 16 sec 63> /iscsi/[email protected]%3A6-8a0900-477d70401b0fff044352423a2-hostname-020000,0 3. c3t1d0 <ABCSTORAGE-100E-00-2.2 cyl 20813 alt 2 hd 16 sec 63> /iscsi/[email protected]%3A6-8a0900-3fcd70401 -085ff04434f423a2-hostname-010000,0 . . . 21. c4t60A98000686F694B2F59775733426B77d0 <ABCSTORAGE-LUN-0.2 cyl 4606 alt 2 hd 16 sec 256> /scsi_vhci/ssd@g60a98000686f694b2f59775733426b77 22. c4t60A98000686F694B2F59775733434C41d0 <ABCSTORAGE-LUN-0.2 cyl 4606 alt 2 hd 16 sec 256> /scsi_vhci/ssd@g60a98000686f694b2f59775733434c41

Monitoring Your iSCSI Conguration

You can display information about the iSCSI initiator and target devices by using the iscsiadm list command.

250

System Administration Guide: Devices and File Systems June 2005

Steps

1. Become superuser. 2. Display information about the iSCSI initiator. For example:
# iscsiadm list initiator-node Initiator node name: iqn.1986-03.com.sun:01:0003ba4d233b.425c293c Initiator node alias: zzr1200 Login Parameters (Default/Configured): Header Digest: NONE/Data Digest: NONE/Authentication Type: NONE RADIUS Server: NONE RADIUS access: unknown

3. Display information about which discovery methods are in use. For example:
# iscsiadm list discovery Discovery: Static: enabled Send Targets: disabled

Example 151

Listing Information About a Specic iSCSI Target

The following example shows how to list information about a specic iSCSI target.
# iscsiadm list target-param iqn.1992-08.com.abcstorage:sn.33592219 Target: iqn.1992-08.com.abcstorage:sn.33592219 Alias: Bi-directional Authentication: disabled Authentication Type: NONE Login Parameters (Default/Configured): Data Sequence In Order: yes/Data PDU In Order: yes/Default Time To Retain: 20/Default Time To Wait: 2/Error Recovery Level: 0/First Burst Length: 65536/Immediate Data: yes/Initial Ready To Transfer (R2T): yes/Max Burst Length: 262144/Max Outstanding R2T: 1/Max Receive Data Segment Length: 65536/Max Connections: 1/Header Digest: NONE/Data Digest: NONE/-

Modifying iSCSI Initiator and Target Parameters

You can modify parameters on both the iSCSI initiator and the iSCSI target device. However, the only parameters that can be modied on the iSCSI initiator are the following:
Chapter 15 Conguring Solaris iSCSI Initiators (Tasks) 251

I I I

Header digest The value can be none, the default value, or CRC32. Data digest The value can be none, the default value, or CRC32. Authentication and CHAP secret For more information about setting up authentication, see How to Congure CHAP Authentication for Your iSCSI Conguration on page 246.

The iSCSI driver provides default values for the iSCSI initiator and iSCSI target device parameters. If you modify the parameters of the iSCSI initiator, the modied parameters are inherited by the iSCSI target device, unless the iSCSI target device is already set to a different value.
Caution Ensure that the target software supports the parameter to be modied. Otherwise, you might be unable to log in to the iSCSI target device. See your array documentation for a list of supported parameters.

Modifying iSCSI parameters should be done when I/O between the initiator and the target is complete. The iSCSI driver reconnects the session after the changes are made with the iscsiadm modify command.

M How to Modify iSCSI Initiator and Target Parameters

The rst part of this procedure illustrates how modifying parameters of the iSCSI initiator are inherited by the iSCSI target device. The second part of this procedure shows how to actually modify parameters on the iSCSI target device.
Steps

1. Become superuser. 2. List the current parameters of the iSCSI initiator and target device. a. List the current parameters of the iSCSI initiator. For example:
# iscsiadm list initiator-node Initiator node name: iqn.1986-03.com.sun:01:0003ba4d233b.425c293c Initiator node alias: zzr1200 Login Parameters (Default/Configured): Header Digest: NONE/Data Digest: NONE/Authentication Type: NONE RADIUS Server: NONE RADIUS access: unknown

b. List the current parameters of the iSCSI target device. For example:
# iscsiadm list target-param -v iqn.1992-08.com.abcstorage:sn.84186266 Target: iqn.1992-08.com.abcstorage:sn.84186266 Alias: 252 System Administration Guide: Devices and File Systems June 2005

Bi-directional Authentication: disabled Authentication Type: NONE Login Parameters (Default/Configured): Data Sequence In Order: yes/Data PDU In Order: yes/Default Time To Retain: 20/Default Time To Wait: 2/Error Recovery Level: 0/First Burst Length: 65536/Immediate Data: yes/Initial Ready To Transfer (R2T): yes/Max Burst Length: 262144/Max Outstanding R2T: 1/Max Receive Data Segment Length: 65536/Max Connections: 1/Header Digest: NONE/Data Digest: NONE/-

Note that both header digest and data digest parameters are currently set to NONE for both the iSCSI initiator and the iSCSI target device. To review the default parameters of the iSCSI target device, see the iscsiadm list target-param output in Example 151. 3. Modify the parameter of the iSCSI initiator. For example, set header digest to CRC32.
# iscsiadm modify initiator-node -h CRC32

4. Verify that the parameter was modied. a. Display the updated parameter information for the iSCSI initiator. For example:
# iscsiadm list initiator-node Initiator node name: iqn.1986-03.com.sun:01:0003ba4d233b.425c293c Initiator node alias: zzr1200 Login Parameters (Default/Configured): Header Digest: NONE/CRC32 Data Digest: NONE/Authentication Type: NONE RADIUS Server: NONE RADIUS access: unknown

Note that the header digest is now set to CRC32. b. Display the updated parameter information for the iSCSI target device. For example:
# iscsiadm list target-param -v iqn.1992-08.com.abcstorage:sn.84186266 Target: iqn.1992-08.com.abcstorage:sn.84186266 Alias: Bi-directional Authentication: disabled Authentication Type: NONE Login Parameters (Default/Configured): Data Sequence In Order: yes/Chapter 15 Conguring Solaris iSCSI Initiators (Tasks) 253

Data PDU In Order: yes/Default Time To Retain: 20/Default Time To Wait: 2/Error Recovery Level: 0/First Burst Length: 65536/Immediate Data: yes/Initial Ready To Transfer (R2T): yes/Max Burst Length: 262144/Max Outstanding R2T: 1/Max Receive Data Segment Length: 65536/Max Connections: 1/Header Digest: CRC32/Data Digest: NONE/-

Note that the header digest is now set to CRC32. 5. Verify that the iSCSI initiator has reconnected to the iSCSI target.
# iscsiadm list target -v iqn.1992-08.com.abcstorage:sn.84186266 Target: iqn.1992-08.com.abcstorage:sn.84186266 Target Portal Group Tag: 2 Connections: 1 CID: 0 IP address (Local): nnn.nn.nn.nnn:64369 IP address (Peer): nnn.nn.nn.nnn:3260 Discovery Method: SendTargets Login Parameters (Negotiated): . . . Header Digest: CRC32 Data Digest: NONE

6. Unset an iSCSI initiator parameter or an iSCSI target device parameter. You can unset a parameter by either setting it to none with the iscsiadm modify command. Or, you can use the iscsiadm remove command to reset all target properties to the default settings. The following example shows how to reset the header digest to none:
# iscsiadm modify target-param -h none iqn.1992-08.com.abcstorage:sn...

For information about using the iscsiadm remove target-param command, see iscsiadm.1m.

Troubleshooting iSCSI Conguration Problems

The following tools are available to troubleshoot general iSCSI conguration problems:
254 System Administration Guide: Devices and File Systems June 2005

I I

snoop This tool has been updated to support iSCSI packets. ethereal This freeware product is available from http://www.ethereal.com.

Both tools can lter iSCSI packets on port 3260. The following sections describe various iSCSI troubleshooting and error message resolution scenarios.

No Connections to the iSCSI Target From the Local System

M How to Troubleshoot iSCSI Connection Problems
Steps

1. Become superuser. 2. List your iSCSI target information. For example:

# iscsiadm list target Target: iqn.2001-05.com.abcstorage:6-8a0900-37ad70401-bcfff02df8a421df -zzr1200-01 Target Portal Group Tag: default Connections: 0

3. If no connections are listed in the iscsiadm list target output, check the /var/adm/messages le for possible reasons why the connection failed. You can also verify whether the connection is accessible by using the ping command or by connecting to the storage devices iSCSI port with the telnet command to ensure the iSCSI service is available. The default port is 3260. 4. If your target is not listed in the iscsiadm list target output, check the /var/adm/messages le for possible causes. If you are using SendTargets as the discovery method, try listing the discovery-address using the -v option to ensure that the expected targets are visible to the host. For example:
# iscsiadm list discovery-address -v 10.0.0.1 Discovery Address: 10.0.0.1:3260 Target name: eui.210000203787dfc0 Target address: 10.0.0.1:11824 Target name: eui.210000203787e07b Target address: 10.0.0.1:11824

Chapter 15 Conguring Solaris iSCSI Initiators (Tasks)

255

iSCSI Device or Disk Is Not Available on the Local System

M How to Troubleshoot iSCSI Device or Disk Unavailability
Steps

1. Become superuser. 2. Identify the LUNs that were discovered on this target during enumeration. For example:
# iscsiadm list target -S Target: iqn.2001-05.com.abcstorage:6-8a0900-37ad70401-bcfff02df8a421df-zzr1200-01 Target Portal Group Tag: default Connections: 1 LUN: 0 Vendor: ABCSTOR Product: 0010 OS Device Name: /dev/rdsk/c3t34d0s2

The -S option shows which LUNs where discovered on this target during enumeration. If you think a LUN should be listed but it is not, review the /var/adm/messages le to see if an error was reported. Check the storage devices log les for errors. Also, ensure that any storage device LUN masking is properly congured.

General iSCSI Error Messages

This section describes the iSCSI messages that might be found in the /var/adm/messages le and potential solutions for recovery. The message format is as follows:
iscsi TYPE (OID) STRING (STATUS-CLASS#/STATUS-DETAIL#)

TYPE OID

Is either connection or session. Is the object ID of the connection or session. This ID is unique for an OS instance. Is a description of the condition. These values are returned in an iSCSI login response as dened by RFC 3270.

STRING <STATUS-CLASS#>/<STATUS-DETAIL#>

iscsi connection(OID) login failed - Miscellaneous iSCSI initiator errors. Cause: The device login failed due to some form of initiator error.
256 System Administration Guide: Devices and File Systems June 2005

iscsi connection(OID) login failed - Initiator could not be successfully authenticated. Cause: The device could not successfully authenticate the initiator.
Solution: If applicable, verify that the passwords or RADIUS information are

accurate. iscsi connection(OID) login failed - Initiator is not allowed access to the given target. Cause: The device will not allow the initiator access to the iSCSI target device.
Solution: Verify your initiator name and conrm that it is properly masked or provisioned by the storage device.

iscsi connection(OID) login failed - Requested ITN does not exist at this address. Cause: The device does not provide access to the iSCSI target name (ITN) that you are requesting.
Solution: Verify the initiator discovery information is entered properly and that the

storage device is congured properly. iscsi connection(OID) login failed - Requested ITN has been removed and no forwarding address is provided. Cause: The device can no longer provide access to the iSCSI target name (ITN) that you are requesting.
Solution: Verify that the initiator discovery information has been specied properly and the storage device has been congured properly.

iscsi connection(OID) login failed - Requested iSCSI version range is not supported by the target. Cause: The initiators iSCSI version is not supported by the storage device. iscsi connection(OID) login failed - No more connections can be accepted on this Session ID (SSID). Cause: The storage device cannot accept another connection for this initiator node to the iSCSI target device. iscsi connection(OID) login failed - Missing parameters (e.g., iSCSI initiator and/or target name). Cause: The storage device is reporting that the initiator or target name has not been properly specied.
Solution: Properly specify the iSCSI initiator or target name.

Chapter 15 Conguring Solaris iSCSI Initiators (Tasks)

257

iscsi connection(OID) login failed - Target hardware or software error. Cause: The storage device encountered a hardware or software error.
Solution: Consult the storage documentation or contact the storage vendor for

further assistance. iscsi connection(OID) login failed - iSCSI service or target is not currently operational. Cause: The storage device is currently not operational.
Solution: Consult the storage documentation or contact the storage vendor for

further assistance. iscsi connection(OID) login failed - Target has insufficient session, connection or other resources. Cause: The storage device has insufficient resources.
Solution: Consult the storage documentation or contact the storage vendor for

further assistance. iscsi connection(OID) login failed - unable to initialize authentication iscsi connection(OID) login failed - unable to set authentication iscsi connection(OID) login failed - unable to set username iscsi connection(OID) login failed - unable to set password iscsi connection(OID) login failed - unable to set ipsec iscsi connection(OID) login failed - unable to set remote authentication Cause: The initiator was unable to initialize or set authentication properly.
Solution: Verify that your initiator settings for authentication are properly

congured. iscsi connection(OID) login failed - unable to make login pdu Cause: The initiator was unable to make a login payload data unit (PDU) based on the initiator or storage device settings.
Solution: Try resetting any target login parameters or other nondefault settings.

258

System Administration Guide: Devices and File Systems June 2005

iscsi connection(OID) login failed - failed to transfer login iscsi connection(OID) login failed - failed to receive login response Cause: The initiator failed to transfer or receive a login payload data unit (PDU) across the network connection.
Solution: Verify that the network connection is reachable.

iscsi connection(OID) login failed - received invalid login response (OP CODE) Cause: The storage device has responded to a login with an unexpected response. iscsi connection(OID) login failed - login failed to authenticate with target Cause: The initiator was unable to authenticate the storage device.
Solution: Verify that your initiator settings for authentication are properly

congured. iscsi connection(OID) login failed - initiator name is required Cause: An initiator name must be congured to perform all actions.
Solution: Verify that the initiator name is congured.

iscsi connection(OID) login failed - authentication receive failed iscsi connection(OID) login failed - authentication transmit failed Cause: The initiator was unable to transmit or receive authentication information.
Solution: Verify the network connectivity with storage device or the RADIUS server

as applicable. iscsi connection(OID) login failed - login redirection invalid Cause: The storage device attempted to redirect the initiator to an invalid destination.
Solution: Consult the storage documentation or contact the storage vendor for

further assistance. iscsi connection(OID) login failed - target protocol group tag mismatch, expected <TPGT>, received <TPGT> Cause: The initiator and target had a TPGT (target portal group tag) mismatch.
Solution: Verify your TPGT discovery settings on the initiator or the storage device.

Chapter 15 Conguring Solaris iSCSI Initiators (Tasks)

259

iscsi connection(OID) login failed - cant accept PARAMETER in security stage Cause: The device responded with an unsupported login parameter during the security phase of login.
Solution: The parameter name is noted for reference. Consult the storage

documentation or contact the storage vendor for further assistance. iscsi connection(OID) login failed - HeaderDigest=CRC32 is required, cant accept V ALUE iscsi connection(OID) login failed - DataDigest=CRC32 is required, cant accept V ALUE Cause: The initiator is only congured to accept HeaderDigest or DataDigest that is set to CRC32 for this target. The device returned the value of V ALUE.
Solution: Verify that the initiator and device digest settings are compatible.

iscsi connection(OID) login failed - HeaderDigest=None is required, cant accept V ALUE iscsi connection(OID) login failed - DataDigest=None is required, cant accept V ALUE Cause: The initiator is only congured to accept HeaderDigest or DataDigest that is set to none for this target. The device returned the value of V ALUE.
Solution: Verify that the initiator and device digest settings are compatible.

iscsi connection(OID) login failed - cant accept PARAMETER Cause: The initiator does not support this parameter. iscsi connection(OID) login failed - cant accept MaxOutstandingR2T V ALUE Cause: The initiator does not accept MaxOutstandingR2T of the noted V ALUE. iscsi connection(OID) login failed - cant accept MaxConnections V ALUE Cause: The initiator does not accept the maximum connections of the noted V ALUE. iscsi connection(OID) login failed - cant accept ErrorRecoveryLevel V ALUE Cause: The initiator does not accept an error recovery level of the noted V ALUE. iscsi session(OID) NAME offline Cause: All connections for this target NAME have been removed or failed. iscsi connection(OID) failure - unable to schedule enumeration Cause: The initiator was unable to enumerate the LUNs on this target.
260 System Administration Guide: Devices and File Systems June 2005

Solution: You can force LUN enumeration by running the devfsadm -i iscsi command. For more information, see devfsadm(1M).

iscsi connection(OID) unable to connect to target NAME (errno:ERRNO) Cause: The initiator failed to establish a network connection.
Solution: For information about the specic ERRNO on the connection failure, see

the /usr/include/sys/errno.h le.

Chapter 15 Conguring Solaris iSCSI Initiators (Tasks)

261

262

System Administration Guide: Devices and File Systems June 2005

CHAPTER

16

The format Utility (Reference)

This chapter describes the format utilitys menus and commands. This is a list of the reference information in this chapter.
I I I I I

Recommendations and Requirements for Using the format Utility on page 263 format Menu and Command Descriptions on page 264 format.dat File on page 270 Rules for Input to format Commands on page 275 Getting Help on the format Utility on page 277

For a overview of when to use the format utility, see format Utility on page 185.

Recommendations and Requirements for Using the format Utility

You must be superuser or have assumed an equivalent role to use the format utility. Otherwise, the following error message is displayed when you try to use the format utility:
$ format Searching for disks...done No permission (or no disks found)!

Keep the following guidelines in mind when you use the format utility and want to preserve the existing data:
I I

Back up all les on the disk drive. Save all your defect lists in les by using the format utilitys dump command. The le name should include the drive type, model number, and serial number. Save the paper copies of the manufacturers defect list that was shipped with your drive.
263

format Menu and Command Descriptions

The format main menu appears similar to the following:
FORMAT MENU: disk type partition current format fdisk repair label analyze defect backup verify save inquiry volname !<cmd> quit format> select a disk select (define) a disk type select (define) a partition table describe the current disk format and analyze the disk run the fdisk program (x86 only) repair a defective sector write label to the disk surface analysis defect list management search for backup labels read and display labels save new disk/partition definitions show vendor, product and revision set 8-character volume name execute <cmd>, then return

The following table describes the main menu items for the format utility.
TABLE 161

The Main Menu Item Descriptions for the format Utility

Command or Menu? Description

Menu Item

disk

Command

Lists all of the systems drives. Also lets you choose the disk you want to use in subsequent operations. This disk is referred to as the current disk. Identies the manufacturer and model of the current disk. Also displays a list of known drive types. Choose the Auto configure option for all SCSI-2 disk drives. Creates and modies slices. For more information, see partition Menu on page 266.

type

Command

partition

Menu

264

System Administration Guide: Devices and File Systems June 2005

TABLE 161 Menu Item

The Main Menu Item Descriptions for the format Utility

Command or Menu? Description

(Continued)

current

Command

Displays the following information about the current disk: I Device name and device type I Number of cylinders, alternate cylinders, heads and sectors I Physical device name Formats the current disk by using one of these sources of information in this order: 1. Information that is found in the format.dat le 2. Information from the automatic conguration process 3. Information that you type at the prompt if no format.dat entry exists This command does not apply to IDE disks. IDE disks are preformatted by the manufacturer.

format

Command

fdisk

Menu

x86 platform only: Runs the fdisk program to create a Solaris fdisk partition. The fdisk command is not intended for disks greater than 1 terabyte and cannot be used on disks with an EFI label.

repair label analyze

Command Command Menu Menu

Repairs a specic block on the current disk. Writes a new label to the current disk. Runs read, write, and compare tests. For more information, see analyze Menu on page 268. Retrieves and displays defect lists. For more information, see defect Menu on page 269. This feature does not apply to IDE disks. IDE disks manage defects automatically. VTOC Searches for backup labels. EFI Not supported.

defect

backup

Command

verify

Command

Displays the following information about the current disk: I Device name and device type I Number of cylinders, alternate cylinders, heads and sectors I Partition table VTOC Saves new disk and partition information. EFI Not applicable.

save

Command

Chapter 16 The format Utility (Reference)

265

TABLE 161 Menu Item

The Main Menu Item Descriptions for the format Utility

Command or Menu? Description

(Continued)

inquiry

Command Command Command

SCSI disks only Displays the vendor, product name, and revision level of the current drive. Labels the disk with a new eight-character volume name that you specify. Exits the format menu.

volname

quit

partition Menu
The partition menu appears similar to the following:
format> partition PARTITION MENU: 0 - change 0 partition 1 - change 1 partition 2 - change 2 partition 3 - change 3 partition 4 - change 4 partition 5 - change 5 partition 6 - change 6 partition 7 - change 7 partition select - select a predefined table modify - modify a predefined partition table name - name the current table print - display the current table label - write partition map and label to the disk quit partition>

The following table describes the partition menu items.

TABLE 162

Descriptions for partition Menu Items

Description

Subcommand

change n partition

Enables you to specify the following information for the new partition: I Identication tag I Permission ags I Starting cylinder I Size Enables you to choose a predened partition table.

select

266

System Administration Guide: Devices and File Systems June 2005

TABLE 162

Descriptions for partition Menu Items

Description

(Continued)

Subcommand

modify

Enables you to change all the slices in the partition table. This command is preferred over the individual change x partition commands. Enables you to specify a name for the current partition table. Displays the current partition table. Writes the partition map and the label to the current disk. Exits the partition menu.

name print label quit

x86: fdisk Menu

The fdisk menu appears on x86 based systems only and appears similar to the following.
format> fdisk Total disk size is 14169 cylinders Cylinder size is 2510 (512 byte) blocks Cylinders Start End Length ===== === ====== 1 9 9 10 14168 14159

Partition ========= 1 2

Status ====== Active

Type ============ x86 Boot Solaris2

% === 0 100

SELECT ONE OF THE FOLLOWING: 1. Create a partition 2. Specify the active partition 3. Delete a partition 4. Change between Solaris and Solaris2 Partition IDs 5. Exit (update disk configuration and exit) 6. Cancel (exit without updating disk configuration) Enter Selection:

The following table describes the fdisk menu items.

TABLE 163 Menu Item

x86: Descriptions for fdisk Menu Items

Description

Create a partition

Creates an fdisk partition. You must create a separate partition for each OS such as Solaris or DOS. There is a maximum of four partitions per disk. You are prompted for the size of the fdisk partition as a percentage of the disk.

Chapter 16 The format Utility (Reference)

267

TABLE 163 Menu Item

x86: Descriptions for fdisk Menu Items

Description

(Continued)

Specify the active partition

Enables you to specify the partition to be used for booting. This menu item identies where the rst stage boot program looks for the second stage boot program. Deletes a previously created partition. This command destroys all the data in the partition. Changes partition IDs from 130 (0x82) to 191 (0xbf) and vice versa. Writes a new version of the partition table and exits the fdisk menu. Exits the fdisk menu without modifying the partition table.

Delete a partition

Change between Solaris and Solaris2 Partition IDs Exit (update disk configuration and exit) Cancel (exit without updating disk configuration)

analyze Menu
The analyze menu appears similar to the following.
format> analyze ANALYZE MENU: read refresh test write compare purge verify print setup config quit analyze>

read only test (doesnt harm read then write (doesnt harm pattern testing (doesnt harm write then read (corrupts write, read, compare (corrupts write, read, write (corrupts write entire disk, then verify display data buffer set analysis parameters show analysis parameters

SunOS) data) data) data) data) data) (corrupts data)

The following table describes the analyze menu items.

TABLE 164

Descriptions for analyze Menu Items

Description

Subcommand

read

Reads each sector on the current disk. Repairs defective blocks as a default.

268

System Administration Guide: Devices and File Systems June 2005

TABLE 164

Descriptions for analyze Menu Items

Description

(Continued)

Subcommand

refresh

Reads then writes data on the current disk without harming the data. Repairs defective blocks as a default. Writes a set of patterns to the disk without harming the data. Repairs defective blocks as a default. Writes a set of patterns to the disk then reads back the data on the disk. Destroys existing data on the disk. Repairs defective blocks as a default. Writes a set of patterns to the disk, reads back the data, and then compares it to the data in the write buffer. Destroys existing data on the disk. Repairs defective blocks as a default. Removes all data from the disk so that the data cannot be retrieved by any means. Data is removed by writing three distinct patterns over the entire disk (or a section of the disk). If the verication passes, a hex-bit pattern is written over the entire disk (or a section of the disk). Repairs defective blocks as a default.

test

write

compare

purge

verify

In the rst pass, writes unique data to each block on the entire disk. In the next pass, reads and veries the data. Destroys existing data on the disk. Repairs defective blocks as a default. Displays the data in the read/write buffer. Enables you to specify the following analysis parameters:
Analyze entire disk? yes Starting block number: depends on drive Ending block number: depends on drive Loop continuously? no Number of passes: 2 Repair defective blocks? yes Stop after first error? no Use random bit patterns? no Number of blocks per transfer: 126 (0/n/nn) Verify media after formatting? yes Enable extended messages? no Restore defect list? yes Restore disk label? yes

print setup

config quit

Displays the current analysis parameters. Exits the analyze menu.

defect Menu
The defect menu appears similar to the following:
format> defect

Chapter 16 The format Utility (Reference)

269

DEFECT MENU: primary grown both print dump quit defect>

extract manufacturers defect list extract manufacturers and repaired defects lists extract both primary and grown defects lists display working list dump working list to file

The following table describes the defect menu items.

TABLE 165

The defect Menu Item Descriptions

Description

Subcommand

primary

Reads the manufacturers defect list from the disk drive and updates the in-memory defect list. Reads the grown defect list and then updates the in-memory defect list. Grown defects are defects that have been detected during analysis. Reads both the manufacturers defect list and the grown defect list. Then, updates the in-memory defect list. Displays the in-memory defect list. Saves the in-memory defect list to a le. Exits the defect menu.

grown

both

print dump quit

format.dat File
The format.dat le that is shipped with the Solaris OS supports many standard disks. If your disk drive is not listed in the format.dat le, you can do the following:
I I

Add an entry to the format.dat le for the disk. Add entries with the format utility by selecting the type command and choosing the other option.

Adding an entry to the format.dat le can save time if the disk drive will be used throughout your site. To use the format.dat le on other systems, copy the le to each system that will use the specic disk drive that you added to the format.dat le. You might need to modify the /etc/format.dat le for your system if you have one of the following:
I

A disk that is not supported by the Solaris OS

270

System Administration Guide: Devices and File Systems June 2005

A disk with a partition table that is different from the Solaris OSs default conguration

Note Do not alter default entries in the /etc/format.dat le. If you want to alter

the default entries, copy the entry, give the entry a different name, and make the appropriate changes to avoid confusion.

The /etc/format.dat is not applicable for disks with EFI labels.

Contents of the format.dat File

The format.dat contains disk drive information that is used by the format utility. Three items are dened in the format.dat le:
I I I

Search paths Disk types Slice tables

Syntax of the format.dat File

The following syntax rules apply to the /etc/format.dat le:
I

The pound sign (#) is the comment character. Any text on a line after a pound sign is not interpreted by the format utility. Each denition in the format.dat le appears on a single logical line. If the denition is longer than one line long, all lines but the last line of the denition must end with a backslash (\). A denition consists of a series of assignments that have an identier on the left side and one or more values on the right side. The assignment operator is the equal sign (=). The assignments within a denition must be separated by a colon (:). White space is ignored by the format utility. If you want an assigned value to contain white space, enclose the entire value in double quotation marks ("). This syntax causes the white space within the quotes to be preserved as part of the assignment value. Some assignments can have multiple values on the right side. Separate values by a comma.

Keywords in the format.dat File

The format.dat le contains disk denitions that are read by the format utility when it is started. Each denition starts with one of the following keywords: disk_type or partition. These keywords are described in the following table.
Chapter 16 The format Utility (Reference) 271

TABLE 166 Keyword

Keyword Descriptions for the format.dat File

Description

disk_type

Denes the controller and disk model. Each disk_type denition contains information that concerns the physical geometry of the disk. The default data le contains denitions for the controllers and disks that the Solaris OS supports. You need to add a new disk_type denition only if you have an unsupported disk. You can add as many disk_type denitions to the data le as you want.

partition

Denes a partition table for a specic disk type. The partition table contains the partition information, plus a name that lets you refer to it in the format utility. The default format.dat le contains default partition denitions for several kinds of disk drives. Add a partition denition if you recreated partitions on any of the disks on your system. Add as many partition denitions to the data le as you need.

Disk Type (format.dat)

The disk_type keyword in the format.dat le denes the controller and disk model. Each disk_type denition contains information about the physical geometry of the disk. The default format.dat le contains denitions for the controllers and disks that the Solaris OS supports. You need to add a new disk_type only if you have an unsupported disk. You can add as many disk_type denitions to the data le as you want. The keyword itself is assigned the name of the disk type. This name appears in the disks label, and is used to identify the disk type whenever the format utility is run. Enclose the name in double quotation marks to preserve any white space in the name. The following table describes the identiers that must also be assigned values in all disk_type denitions.
TABLE 167 Identier

Required disk_type Identiers (format.dat)

Description

ctlr

Identies the controller type for the disk type. Currently, the supported values are SCSI and ATA. Species the number of data cylinders in the disk type. This determines how many logical disk cylinders the system will be allowed to access. Species the number of alternate cylinders in the disk type. These cylinders are used by the format utility to store information such as the defect list for the drive. You should always reserve at least two cylinders for alternates.

ncyl

acyl

272

System Administration Guide: Devices and File Systems June 2005

TABLE 167 Identier

Required disk_type Identiers (format.dat)

Description

(Continued)

pcyl

Species the number of physical cylinders in the disk type. This number is used to calculate the boundaries of the disk media. This number is usually equal to ncyl plus acyl. Species the number of heads in the disk type. This number is used to calculate the boundaries of the disk media. Species the number of data sectors per track in the disk type. This number is used to calculate the boundaries of the disk media. Note that this number includes only the data sectors. Any spares are not reected in the number of data sections per track. Species the rotations per minute of the disk type. This information is put in the label and later used by the le system to calculate the optimal placement of le data.

nhead

nsect

rpm

Other identiers might be necessary, depending on the controller. The following table describes the identiers that are required for SCSI controllers.
TABLE 168 Identier

Required disk_type Identiers for SCSI Controllers format.dat

Description

fmt_time

Species a number that indicates how long it takes to format a given drive. See the controller manual for more information. Species a number that controls the operation of the on-board cache while the format utility is operating. See the controller manual for more information. Species a number that identies how many tracks that exist per defect zone, to be used in alternate sector mapping. See the controller manual for more information. Species a number that identies how many sectors are available for alternate mapping within a given defect zone. See the controller manual for more information.

cache

trks_zone

asect

EXAMPLE 161

Required disk_type Identiers for SCSI Controllers (format.dat)

The following are examples of disk_type denitions:

disk_type : : : : disk_type : : : = "SUN1.3G" \ ctlr = SCSI : fmt_time = 4 \ trks_zone = 17 : asect = 6 : atrks = 17 \ ncyl = 1965 : acyl = 2 : pcyl = 3500 : nhead = 17 : nsect = 80 \ rpm = 5400 : bpt = 44823 = "SUN2.1G" \ ctlr = SCSI : fmt_time = 4 \ ncyl = 2733 : acyl = 2 : pcyl = 3500 : nhead = 19 : nsect = 80 \ rpm = 5400 : bpt = 44823 Chapter 16 The format Utility (Reference) 273

EXAMPLE 161

(Continued)

Required disk_type Identiers for SCSI Controllers (format.dat)

disk_type : : :

= "SUN2.9G" \ ctlr = SCSI : fmt_time = 4 \ ncyl = 2734 : acyl = 2 : pcyl = 3500 : nhead = 21 : nsect = 99 \ rpm = 5400

Partition Tables (format.dat)

A partition table in the format.dat le denes a slice table for a specic disk type. The partition keyword in the format.dat le is assigned the name of the partition table. Enclose the name in double quotation marks to preserve any white space in the name. The following table describes the identiers that must be assigned values in all partition tables.
TABLE 169 Identier

Required Identiers for Partition Tables (format.dat)

Description

disk

The name of the disk_type that this partition table is dened for. This name must appear exactly as it does in the disk_type denition. The disk controller type that this partition table can be attached to. Currently, the supported values are ATA for ATA controllers and SCSI for SCSI controllers. The controller type that is specied here must also be dened for the disk_type that you specied in the disk_type denition.

ctlr

The other identiers in a slice denition describe the actual partition information. The identiers are the numbers 0 through 7. These identiers are optional. Any partition that is not explicitly assigned is set to 0 length. The value of each of these identiers is a pair of numbers separated by a comma. The rst number is the starting cylinder for the partition. The second is the number of sectors in the slice.
EXAMPLE 162

Required Identiers for Partition Tables (format.dat)

The following are some examples of slice denitions:

partition = "SUN1.3G" \ : disk = "SUN1.3G" : ctlr = SCSI \ : 0 = 0, 34000 : 1 = 25, 133280 : 2 = 0, 2672400 : 6 = 123, 2505120 partition = "SUN2.1G" \ : disk = "SUN2.1G" : ctlr = SCSI \ : 0 = 0, 62320 : 1 = 41, 197600 : 2 = 0, 4154160 : 6 = 171, 3894240 partition = "SUN2.9G" \ : disk = "SUN2.9G" : ctlr = SCSI \ 274 System Administration Guide: Devices and File Systems June 2005

EXAMPLE 162

Required Identiers for Partition Tables (format.dat)

(Continued)

: 0 = 0, 195426 : 1 = 94, 390852 : 2 = 0, 5683986 : 6 = 282, 5097708

Specifying an Alternate Data File for the format Utility

The format utility determines the location of an alternate le by the following methods in this order: 1. If a le name is given with the format -x option, that le is always used as the data le. 2. If the -x option is not specied, then the format utility searches the current directory for a le named format.dat. If the le exists, it is used as the data le. 3. If neither of these methods yields a data le, the format utility uses the /etc/format.dat le as the data le. This le is shipped with the Solaris OS and should always be present.

Rules for Input to format Commands

When you use the format utility, you need to provide various kinds of information. This section describes the rules for this information. For information on using formats help facility when you specify data, see Getting Help on the format Utility on page 277.

Specifying Numbers to format Commands

Several places in the format utility require number as input. You must either specify the appropriate data or select a number from a list of choices. In either case, the help facility causes format to display the upper and lower limits of the number expected. Simply enter the appropriate number. The number is assumed to be in decimal format unless a base is explicitly specied as part of the number (for example, 0x for hexadecimal). The following are examples of integer input:
Enter number of passes [2]: 34 Enter number of passes [34] Oxf

Chapter 16 The format Utility (Reference)

275

Specifying Block Numbers to format Commands

Whenever you are required to specify a disk block number, there are two ways to do so:
I I

Specify the block number as an integer Specify the block number in the cylinder/head/sector format

You can specify the information as an integer that represents the logical block number. You can specify the number in any base, but the default is decimal. The maximum operator (a dollar sign, $) can also be used here so that format utility can select the appropriate value. Logical block format is used by the SunOS disk drivers in error messages. The other way to specify a block number is by using cylinder/head/sector format. In this method, you must specify explicitly the three logical components of the block number: the cylinder, head, and sector values. These values are still logical. However, they allow you to dene regions of the disk that are related to the layout of the media. If any of the cylinder/head/sector numbers are not specied, the value is assumed to be zero. You can also use the maximum operator in place of any of the numbers. Then, the format utility will select the appropriate value. The following are some examples of cylinder, head, and sector values:
Enter Enter Enter Enter Enter Enter Enter defective defective defective defective defective defective defective block block block block block block block number: number: number: number: number: number: number: 34/2/3 23/1/ 457// 12345 Oxabcd 334/$/2 892//$

The format utility always displays block numbers in both formats. Also, the help facility shows you the upper and lower limits of the block number expected, in both formats.

Specifying format Command Names

Command names are needed as input whenever the format utility displays a menu prompt. You can abbreviate the command names, as long as what you type is sufficient to uniquely identify the command desired. For example, use p to access the partition menu from the format menu. Then, type p to display the current slice table.
format> p PARTITION MENU: 0 - change 0 partition 1 - change 1 partition 2 - change 2 partition 276 System Administration Guide: Devices and File Systems June 2005

3 4 5 6 7 select modify name print label quit partition> p

change 3 partition change 4 partition change 5 partition change 6 partition change 7 partition select a predefined table modify a predefined partition table name the current table display the current table write partition map and label to the disk

Specifying Disk Names to format Commands

At certain points in the format utility, you must name something. In these cases, you are free to specify any string you want for the name. If the name has white space in it, the entire name must be enclosed in double quotation marks ("). Otherwise, only the rst word of the name is used. For example, if you want to identify a specic partition table for a disk, you can use the name subcommand that is available from the partition menu:
partition> name Enter table name (remember quotes): "new disk3"

Getting Help on the format Utility

The format utility provides a help facility that you can use whenever the format utility is expecting input. You can request help about what input is expected by typing a question mark (?). The format utility displays a brief description of what type of input is needed. If you type a ? at a menu prompt, a list of available commands is displayed. The man pages associated with the format utility include the following:
I

format(1M) Describes the basic format utility capabilities and provides descriptions of all command-line variables. format.dat(4) Describes disk drive conguration information for the format utility.

Chapter 16 The format Utility (Reference)

277

278

System Administration Guide: Devices and File Systems June 2005

CHAPTER

17

Managing File Systems (Overview)

Managing le systems is one of your most important system administration tasks. This is a list of the overview information in this chapter.
I I I I I I I I I I

Whats New in File Systems? on page 279 Where to Find File System Management Tasks on page 283 Overview of File Systems on page 284 Types of File Systems on page 284 Commands for File System Administration on page 290 Default Solaris File Systems on page 291 Swap Space on page 289 UFS File System on page 293 Mounting and Unmounting File Systems on page 296 Determining a File Systems Type on page 300

Whats New in File Systems?

This section describes new le system features in the Solaris 10 release.

UFS Logging Is Enabled by Default

Solaris 10 Logging is enabled by default for all UFS le systems, except under the following conditions:
I I

When logging is explicitly disabled. If there is insufficient le system space for the log.

In previous Solaris releases, you had to manually enable UFS logging. For more information about UFS logging, see UFS Logging on page 294.
279

Keep the following issues in mind when using UFS logging in this release:
I

Ensure that you have enough disk space for your general system needs, such as for users and applications, and for UFS logging. If you dont have enough disk space for logging data, a message similar to the following is displayed:
# mount /dev/dsk/c0t4d0s0 /mnt /mnt: No space left on device Could not enable logging for /mnt on /dev/dsk/c0t4d0s0. #

However, the le system is still mounted. For example:

# df -h /mnt Filesystem /dev/dsk/c0t4d0s0 #
I

size 142M

used 142M

avail capacity 0K 100%

Mounted on /mnt

A UFS le system with logging enabled that is generally empty will have some disk space consumed for the log. If you upgrade to this Solaris release from a previous Solaris release, your UFS le systems will have logging enabled, even if the logging option was not specied in the /etc/vfstab le. To disable logging, add the nologging option to the UFS le system entries in the /etc/vfstab le.

NFS Version 4
This Solaris release includes the Sun implementation of the NFS version 4 distributed le access protocol. NFS version 4 integrates le access, le locking, and mount protocols into a single, unied protocol to ease traversal through a rewall and improve security. The Solaris implementation of NFS version 4 is fully integrated with Kerberos V5, also known as SEAM, thus providing authentication, integrity, and privacy. NFS version 4 also enables the negotiation of security avors to be used between the client and the server. With NFS version 4, a server can offer different security avors for different le systems. For more information about NFS Version 4 features, see Whats New With the NFS Service in System Administration Guide: Network Services.

NFS Version 4 and CacheFS Compatibility Issues

If both the CacheFS client and the CacheFS server are running NFS version 4, les are no longer cached in a front le system. All le access is provided by the back le system. Also, since no les are being cached in the front le system, CacheFS-specic mount options, which are meant to affect the front le system, are ignored. CacheFS-specic mount options do not apply to the back le system.
280 System Administration Guide: Devices and File Systems June 2005

Note The rst time you congure your system for NFS version 4, a warning appears on the console to indicate that caching is no longer performed.

If you want to implement your CacheFS mounts as in previous Solaris releases, then specify NFS version 3 in your CacheFS mount commands. For example:
mount -F cachefs -o backfstype=nfs,cachedir=/local/mycache,vers=3 starbug:/docs /docs

64-bit: Support of Multiterabyte UFS File Systems

This Solaris release provides support for multiterabyte UFS le systems on systems that run a 64-bit Solaris kernel. Previously, UFS le systems were limited to approximately 1 terabyte on both 64-bit and 32-bit systems. All UFS le system commands and utilities have been updated to support multiterabyte UFS le systems. For example, the ufsdump command has been updated with a larger block size for dumping large UFS le systems:
# ufsdump 0f /dev/md/rdsk/d97 /dev/md/rdsk/d98 DUMP: Date of this level 0 dump: Tue Jan 07 14:23:36 2003 DUMP: Date of last level 0 dump: the epoch DUMP: Dumping /dev/md/rdsk/d98 to /dev/md/rdsk/d97. DUMP: Mapping (Pass I) [regular files] DUMP: Mapping (Pass II) [directories] DUMP: Forcing larger tape block size (2048). DUMP: Writing 32 Kilobyte records DUMP: Estimated 4390629500 blocks (2143862.06MB). DUMP: Dumping (Pass III) [directories] DUMP: Dumping (Pass IV) [regular files]

Administering UFS le systems that are less than 1 terabyte remains the same. No administration differences exist between UFS le systems that are less than one terabyte and le systems that are greater than 1 terabyte. You can initially create a UFS le system that is less than 1 terabyte and specify that it can eventually be expanded into a multiterabyte le system by using the newfs -T option. This option sets the inode and fragment density to scale appropriately for a multiterabyte le system. Using the newfs -T option when you create a UFS le system less than 1 terabyte on a system running a 32-bit kernel enables you to eventually expand this le system by using the growfs command when you boot this system under a 64-bit kernel. For more information, see newfs(1M). You can use the growfs command to expand a UFS le system to the size of the slice or the volume without loss of service or data. For more information, see growfs(1M).
Chapter 17 Managing File Systems (Overview) 281

Two new related features are multiterabyte volume support with the EFI disk label and multiterabyte volume support with Solaris Volume Manager. For more information, see Multiterabyte Disk Support With EFI Disk Label on page 175 and the Solaris Volume Manager Administration Guide.

Features of Multiterabyte UFS File Systems

Multiterabyte UFS le systems include the following features:
I I

Provides the ability to create a UFS le system up to 16 terabytes in size. Provides the ability to create a le system less than 16 terabytes that can later be increased in size up to 16 terabytes. Multiterabyte le systems can be created on physical disks, Solaris Volume Managers logical volumes, and Veritas VxVM logical volumes. Multiterabyte le systems benet from the performance improvements of having UFS logging enabled. Multiterabyte le systems also benet from the availability of logging because the fsck command might not have to be run when logging is enabled. When you create a partition for your multiterabyte UFS le system, the disk will be labeled automatically with an EFI disk label. For more information on EFI disk labels, see Multiterabyte Disk Support With EFI Disk Label on page 175. Provides the ability to snapshot a multiterabyte le system by creating multiple backing store les when a le system is over 512 Gbytes.

Limitations of Multiterabyte UFS File Systems

Limitations of multiterabyte UFS le systems are as follows:
I I

This feature is not supported on 32-bit systems. You cannot mount a le system greater than 1 terabyte on a system that is running a 32-bit Solaris kernel. You cannot boot from a le system greater than 1 terabyte on a system that is running a 64-bit Solaris kernel. This limitation means that you cannot put a root (/) le system on a multiterabyte le system. There is no support for individual les greater than 1 terabyte. The maximum number of les is 1 million les per terabyte of a UFS le system. For example, a 4-terabyte le system can contain 4 million les. This limit is intended to reduce the time it takes to check the le system with the fsck command.

I I

The maximum quota that you can set on a multiterabyte UFS le system is 2 terabytes of 1024-byte blocks.

282

System Administration Guide: Devices and File Systems June 2005

Where to Find Multiterabyte UFS Tasks

Use these references to nd step-by-step instructions for working with multiterabyte UFS le systems.

Multiterabyte UFS Task

For More Information

Create multiterabyte UFS le systems

How to Create a Multiterabyte UFS File System on page 306 How to Expand a Multiterabyte UFS File System on page 307 How to Expand a UFS File System to a Multiterabyte UFS File System on page 308

Create a multiterabyte UFS snapshot Troubleshoot multiterabyte UFS problems

Example 262 Troubleshooting Multiterabyte UFS File System Problems on page 309

libc_hwcap
The mount output on an x86 system might include a loopback mount of a libc_hwcap library, a hardware-optimized implementation of libc. This libc implementation is intended to optimize the performance of 32-bit applications. This loopback mount requires no administration and consumes no disk space.

Where to Find File System Management Tasks

Use these references to nd step-by-step instructions for managing le systems.
File System Management Task For More Information

Create new le systems. Make local and remote les available to users. Connect and congure new disk devices.

Chapter 18 and Chapter 20 Chapter 19 Chapter 11

Chapter 17 Managing File Systems (Overview)

283

File System Management Task

For More Information

Design and implement a backup schedule and Chapter 24 restore les and le systems, as needed. Check for and correct le system inconsistencies. Chapter 22

Overview of File Systems

A le system is a structure of directories that is used to organize and store les. The term le system is used to describe the following:
I I I I

A particular type of le system: disk-based, network-based, or virtual The entire le tree, beginning with the root (/) directory The data structure of a disk slice or other media storage device A portion of a le tree structure that is attached to a mount point on the main le tree so that the les are accessible

Usually, you know from the context which meaning is intended. The Solaris OS uses the virtual le system (VFS) architecture, which provides a standard interface for different le system types. The VFS architecture enables the kernel to handle basic operations, such as reading, writing, and listing les. The VFS architecture also makes it easier to add new le systems.

Types of File Systems

The Solaris OS supports three types of le systems:
I I I

Disk-based Network-based Virtual

To identify the le system type, see Determining a File Systems Type on page 300.

Disk-Based File Systems

Disk-based le systems are stored on physical media such as hard disks, CD-ROMs, and diskettes. Disk-based le systems can be written in different formats. The available formats are described in the following table.
284 System Administration Guide: Devices and File Systems June 2005

Disk-Based File System

Format Description

UFS

UNIX le system (based on the BSD Fast File system that was provided in the 4.3 Tahoe release). UFS is the default disk-based le system for the Solaris OS. Before you can create a UFS le system on a disk, you must format the disk and divide it into slices. For information on formatting disks and dividing disks into slices, see Chapter 11.

HSFS

High Sierra, Rock Ridge, and ISO 9660 le system. High Sierra is the rst CD-ROM le system. ISO 9660 is the official standard version of the High Sierra le system. The HSFS le system is used on CD-ROMs, and is a read-only le system. Solaris HSFS supports Rock Ridge extensions to ISO 9660. When present on a CD-ROM, these extensions provide all UFS le system features and le types, except for writability and hard links. PC le system, which allows read- and write- access to data and programs on DOS-formatted disks that are written for DOS-based personal computers. The Universal Disk Format (UDF) le system, the industry-standard format for storing information on the optical media technology called DVD (Digital Versatile Disc or Digital Video Disc).

PCFS

UDF

Each type of disk-based le system is customarily associated with a particular media device, as follows:
I I I I

UFS with hard disk HSFS with CD-ROM PCFS with diskette UDF with DVD

However, these associations are not restrictive. For example, CD-ROMs and diskettes can have UFS le systems created on them.

The Universal Disk Format (UDF) File System

The UDF le system is the industry-standard format for storing information on DVD (Digital Versatile Disc or Digital Video Disc) optical media. The UDF le system is provided as dynamically loadable 32-bit and 64-bit modules, with system administration utilities for creating, mounting, and checking the le system on both SPARC and x86 platforms. The Solaris UDF le system works with supported ATAPI and SCSI DVD drives, CD-ROM devices, and disk and diskette drives. In addition, the Solaris UDF le system is fully compliant with the UDF 1.50 specication. The UDF le system provides the following features:
Chapter 17 Managing File Systems (Overview) 285

Ability to access the industry-standard CD-ROM and DVD-ROM media when they contain a UDF le system Flexibility in exchanging information across platforms and operating systems A mechanism for implementing new applications rich in broadcast-quality video, high-quality sound, and interactivity using the DVD video specication based on UDF format

I I

The following features are not included in the UDF le system:

I

Support for write-once media, (CD-RW, and DVD-RAM), with either the sequential disk-at-once recording and incremental recording UFS components such as quotas, ACLs, transaction logging, le system locking, and le system threads, which are not part of the UDF 1.50 specication

The UDF le system requires the following:

I I I

At least the Solaris 7 11/99 release Supported SPARC or x86 platform Supported CD-ROM or DVD-ROM device

The Solaris UDF le system implementation provides the following:

I I

Support for industry-standard read/write UDF version 1.50 Fully internationalized le system utilities

Network-Based File Systems

Network-based le systems can be accessed from the network. Typically, network-based le systems reside on one system, typically a server, and are accessed by other systems across the network. With NFS, you can administer distributed resources (les or directories) by exporting them from a server and mounting them on individual clients. For more information, see The NFS Environment on page 299.

Virtual File Systems

Virtual le systems are memory-based le systems that provide access to special kernel information and facilities. Most virtual le systems do not use le system disk space. However, the CacheFS le system uses a le system on the disk to contain the cache. Also, some virtual le systems, such as the temporary le system (TMPFS), use the swap space on a disk.

CacheFS File System

The CacheFS le system can be used to improve the performance of remote le systems or slow devices such as CD-ROM drives. When a le system is cached, the data that is read from the remote le system or CD-ROM is stored in a cache on the local system.
286 System Administration Guide: Devices and File Systems June 2005

If you want to improve the performance and scalability of an NFS or CD-ROM le system, you should use the CacheFS le system. The CacheFS software is a general purpose caching mechanism for le systems that improves NFS server performance and scalability by reducing server and network load. Designed as a layered le system, the CacheFS software provides the ability to cache one le system on another. In an NFS environment, CacheFS software increases the client per server ratio, reduces server and network loads, and improves performance for clients on slow links, such as Point-to-Point Protocol (PPP). You can also combine a CacheFS le system with the AutoFS service to help boost performance and scalability. For detailed information about the CacheFS le system, see Chapter 20.

Temporary File System

The temporary le system (TMPFS) uses local memory for le system reads and writes. Typically, using memory for le system reads and writes is much faster than using a UFS le system. Using TMPFS can improve system performance by saving the cost of reading and writing temporary les to a local disk or across the network. For example, temporary les are created when you compile a program. The OS generates a much disk activity or network activity while manipulating these les. Using TMPFS to hold these temporary les can signicantly speed up their creation, manipulation, and deletion. Files in TMPFS le systems are not permanent. These les are deleted when the le system is unmounted and when the system is shut down or rebooted. TMPFS is the default le system type for the /tmp directory in the Solaris OS. You can copy or move les into or out of the /tmp directory, just as you would in a UFS le system. The TMPFS le system uses swap space as a temporary backing store. If a system with a TMPFS le system does not have adequate swap space, two problems can occur:
I I

The TMPFS le system can run out of space, just as regular le systems do. Because TMPFS allocates swap space to save le data (if necessary), some programs might not execute because of insufficient swap space.

For information about creating TMPFS le systems, see Chapter 18. For information about increasing swap space, see Chapter 21.

The Loopback File System

The loopback le system (LOFS) lets you create a new virtual le system so that you can access les by using an alternative path name. For example, you can create a loopback mount of the root (/) directory on /tmp/newroot. This loopback mounts make the entire le system hierarchy appear as if it is duplicated under /tmp/newroot, including any le systems mounted from NFS servers. All les will be accessible either with a path name starting from root (/), or with a path name that starts from /tmp/newroot.
Chapter 17 Managing File Systems (Overview) 287

For information on how to create LOFS le systems, see Chapter 18.

Process File System

The process le system (PROCFS) resides in memory and contains a list of active processes, by process number, in the /proc directory. Information in the /proc directory is used by commands such as ps. Debuggers and other development tools can also access the address space of the processes by using le system calls.
Caution Do not delete les in the /proc directory. The deletion of processes from the

/proc directory does not kill them. /proc les do not use disk space, so there is no reason to delete les from this directory.

The /proc directory does not require administration.

Additional Virtual File Systems

These additional types of virtual le systems are listed for your information. They do not require administration.

Virtual File System

Description

CTFS

CTFS (the contract le system) is the interface for creating, controlling, and observing contracts. A contract enhances the relationship between a process and the system resources it depends on by providing richer error reporting and (optionally) a means of delaying the removal of a resource. The service management facility (SMF) uses process contracts (a type of contract) to track the processes which compose a service, so that a failure in a part of a multi-process service can be identied as a failure of that service.

FIFOFS (rst-in rst-out) FDFS (le descriptors) MNTFS NAMEFS

Named pipe les that give processes common access to data Provides explicit names for opening les by using le descriptors Provides read-only access to the table of mounted le systems for the local system Used mostly by STREAMS for dynamic mounts of le descriptors on top of les

288

System Administration Guide: Devices and File Systems June 2005

Virtual File System

Description

OBJFS

The OBJFS (object) le system describes the state of all modules currently loaded by the kernel. This le system is used by debuggers to access information about kernel symbols without having to access the kernel directly. Provides access to character special devices and block devices Used by the kernel for swapping

SPECFS (special) SWAPFS

Extended File Attributes

The UFS, NFS, and TMPFS le systems have been enhanced to include extended le attributes. Extended le attributes enable application developers to associate specic attributes to a le. For example, a developer of an application used to manage a windowing system might choose to associate a display icon with a le. Extended le attributes are logically represented as les within a hidden directory that is associated with the target le. You can use the runat command to add attributes and execute shell commands in the extended attribute namespace. This namespace is a hidden attribute directory that is associated with the specied le. To use the runat command to add attributes to a le, you rst have to create the attributes le.
$ runat filea cp /tmp/attrdata attr.1

Then, use the runat command to list the attributes of the le.
$ runat filea ls -l

For more information, see the runat(1) man page. Many Solaris le system commands have been modied to support le system attributes by providing an attribute-aware option. Use this option to query, copy, or nd le attributes. For more information, see the specic man page for each le system command.

Swap Space
The Solaris OS uses some disk slices for temporary storage rather than for le systems. These slices are called swap slices, or swap space. Swap space is used for virtual memory storage areas when the system does not have enough physical memory to handle current processes. Since many applications rely on swap space, you should know how to plan for, monitor, and add more swap space, when needed. For an overview about swap space and instructions for adding swap space, see Chapter 21.
Chapter 17 Managing File Systems (Overview) 289

Commands for File System Administration

Most commands for le system administration have both a generic component and a le systemspecic component. Whenever possible, you should use the generic commands, which call the le systemspecic component. The following table lists the generic commands for le system administration. These commands are located in the /usr/sbin directory.
TABLE 171 Command

Generic Commands for File System Administration

Description Man Page

clri df ff fsck

Clears inodes Reports the number of free disk blocks and les Lists le names and statistics for a le system Checks the integrity of a le system and repairs any damage found Debugs the le system Determines the le system type Lists or provides labels for le systems when they are copied to tape (for use only by the volcopy command) Creates a new le system Mounts local and remote le systems Mounts all le systems that are specied in the virtual le system table (/etc/vfstab) Generates a list of path names with their inode numbers Unmounts local and remote le systems Unmounts all le systems that are specied in the virtual le system table (/etc/vfstab) Creates an image copy of a le system

clri(1M) df(1M) ff(1M) fsck(1M)

fsdb fstyp labelit

fsdb(1M) fstyp(1M) labelit(1M)

mkfs mount mountall

mkfs(1M) mount(1M) mountall(1M)

ncheck

ncheck(1M)

umount umountall

mount(1M) mountall(1M)

volcopy

volcopy(1M)

290

System Administration Guide: Devices and File Systems June 2005

How File System Commands Determine the File System Type

The generic le system commands determine the le system type by following this sequence: 1. From the -F option, if supplied. 2. By matching a special device with an entry in the /etc/vfstab le (if the special device is supplied). For example, fsck rst looks for a match against the fsck device eld. If no match is found, the command then checks the special device eld. 3. By using the default specied in the /etc/default/fs le for local le systems and in the /etc/dfs/fstypes le for remote le systems.

Manual Pages for Generic and Specic File System Commands

Both the generic commands and specic commands have manual pages in the man pages section 1M: System Administration Commands. The manual pages for the generic le system commands provide information about generic command options only. The manual page for a specic le system command has information about options for that le system. To look at a manual page for a specic le system, append an underscore and the abbreviation for the le system type to the generic command name. For example, to see the specic manual page for mounting a UFS le system, type the following:
$ man mount_ufs

Default Solaris File Systems

The Solaris UFS le system is hierarchical, starting with the root directory (/) and continuing downwards through a number of directories. The Solaris installation process enables you to install a default set of directories and uses a set of conventions to group similar types of les together. For a description of the contents of Solaris le systems and directories, see filesystem(5). The following table provides a summary of the default Solaris le systems.

Chapter 17 Managing File Systems (Overview)

291

TABLE 172

The Default Solaris File Systems

File System Type Description

File System or Directory

root (/)

UFS

The top of the hierarchical le tree. The root (/) directory contains the directories and les that are critical for system operation, such as the kernel, the device drivers, and the programs used to boot the system. The root (/) directory also contains the mount point directories where local and remote le systems can be attached to the le tree. System les and directories that can be shared with other users. Files that run only on certain types of systems are in the /usr le system (for example, SPARC executables). Files that can be used on all types of systems, such as the man pages, are in the /usr/share directory. The mount point for user home directories, which store user work les. By default, the /home directory is an automounted le system. On stand-alone systems, the /home directory might be a UFS le system on a local disk slice. System les and directories that are likely to change or grow over the life of the local system. These include system logs, vi and ex backup les, and uucp les. Optional mount point for third-party software. On some systems, the /opt directory might be a UFS le system on a local disk slice. Temporary les, which are removed each time the system is booted or the /tmp le system is unmounted. A list of active processes, by process number. A virtual le system that provides read-only access to the table of mounted le systems for the local system. A memory-based le system for storing temporary les that are not needed after the system is booted. A virtual le system that maintains contract information. A virtual le system that is used by debuggers to access information about kernel symbols without having to access the kernel directly.

/usr

UFS

/export/home or /home

NFS, UFS

/var

UFS

/opt

NFS, UFS

/tmp

TMPFS PROCFS MNTFS TMPFS

/proc /etc/mnttab

/var/run

/system/contract CTFS /system/object OBJFS

292

System Administration Guide: Devices and File Systems June 2005

The root (/) and /usr le systems are required to run a system. Some of the most basic commands in the /usr le system (like mount) are also included in the root (/) le system. As such, they are available when the system boots or is in single-user mode, and /usr is not mounted. For more detailed information on the default directories for the root (/) and /usr le systems, see Chapter 23.

UFS File System

UFS is the default disk-based le system in Solaris OS. Most often, when you administer a disk-based le system, you are administering UFS le systems. UFS provides the following features.

UFS Feature

Description

Extended fundamental types (EFT) Large le systems

Provides 32-bit user ID (UID), group ID (GID), and device numbers. Allows les of about 1 terabyte in size in a le system that can be up to 16 terabytes in size. You can create a multiterabyte UFS le system on a disk with an EFI disk label. UFS logging bundles the multiple metadata changes that comprise a complete UFS operation into a transaction. Sets of transactions are recorded in an on-disk log and are applied to the actual UFS le systems metadata. A multiterabyte le system enables creation of a UFS le system up to approximately 16 terabytes of usable space, minus approximately one percent overhead Shows the state of the le system: clean, stable, active, logging, or unknown. These ags eliminate unnecessary le system checks. If the le system is clean, stable, or logging, le system checks are not run.

Logging

Multiterabyte le systems

State ags

For detailed information about the UFS le system structure, see Chapter 23.

Planning UFS File Systems

When laying out le systems, you need to consider possible conicting demands. Here are some suggestions:
Chapter 17 Managing File Systems (Overview) 293

Distribute the workload as evenly as possible among different I/O systems and disk drives. Distribute the /export/home le system and swap space evenly across disks. Keep pieces of projects or members of groups within the same le system. Use as few le systems per disk as possible. On the system (or boot) disk, you should have three le systems: root (/), /usr, and swap space. On other disks, create one or at most two le systems, with one le system preferrably being additional swap space. Fewer, roomier le systems cause less le fragmentation than many small, over crowded le systems. Higher-capacity tape drives and the ability of the ufsdump command to handle multiple volumes make it easier to back up larger le systems. If you have some users who consistently create very small les, consider creating a separate le system with more inodes. However, most sites do not need to keep similar types of user les in the same le system.

I I

For information on default le system parameters as well as procedures for creating new UFS le systems, see Chapter 18.

UFS Logging
UFS logging bundles the multiple metadata changes that comprise a complete UFS operation into a transaction. Sets of transactions are recorded in an on-disk log. Then, they are applied to the actual UFS le systems metadata. At reboot, the system discards incomplete transactions, but applies the transactions for completed operations. The le system remains consistent because only completed transactions are ever applied. This consistency remains even when a system crashes. A system crash might interrupt system calls and introduces inconsistencies into a UFS le system. UFS logging provides two advantages:
I

If the le system is already consistent due to the transaction log, you might not have to run the fsck command after a system crash or an unclean shutdown. For more information on unclean shutdowns, see What the fsck Command Checks and Tries to Repair on page 373. Starting in the Solaris 9 12/02 release, the performance of UFS logging improves or exceeds the level of performance of non logging le systems. This improvement can occur because a le system with logging enabled converts multiple updates to the same data into single updates. Thus, reduces the number of overhead disk operations required.

The UFS transaction log has the following characteristics:

I I

Is allocated from free blocks on the le system Sized at approximately 1 Mbyte per 1 Gbyte of le system, up to a maximum of 64 Mbytes

294

System Administration Guide: Devices and File Systems June 2005

I I

Continually ushed as it lls up Also ushed when the le system is unmounted or as a result of any lockfs command.

UFS logging is enabled by default for all UFS le systems. If you need to disable UFS logging, add the nologging option to the le systems entry in the /etc/vfstab le or when you manually mount the le system. If you need to enable UFS logging, specify the -o logging option with the mount command in the /etc/vfstab le or when you manually mount the le system. Logging can be enabled on any UFS le system, including the root (/) le system. Also, the fsdb command has new debugging commands to support UFS logging. In some operating systems, a le system with logging enabled is known as a journaling le system.

UFS Snapshots
You can use the fssnap command to create a read-only snapshot of a le system. A snapshot is a le systems temporary image that is intended for backup operations. See Chapter 26 for more information.

UFS Direct Input/Output (I/O)

Direct I/O is intended to boost bulk I/O operations. Bulk I/O operations use large buffer sizes to transfer large les (larger than 256 Kbytes). Using UFS direct I/O might benet applications, such as database engines, that do their own internal buffering. Starting with the Solaris 8 1/01 release, UFS direct I/O has been enhanced to allow the same kind of I/O concurrency that occurs when raw devices are accessed. Now you can get the benet of le system naming and exibility with very little performance penalty. Check with your database vendor to see if it can enable UFS direct I/O in its product conguration options. Direct I/O can also be enabled on a le system by using the forcedirectio option to the mount command. Enabling direct I/O is a performance benet only when a le system is transferring large amounts of sequential data. When a le system is mounted with this option, data is transferred directly between a users address space and the disk. When forced direct I/O is not enabled for a le system, data transferred between a users address space and the disk is rst buffered in the kernel address space. The default behavior is no forced direct I/O on a UFS le system. For more information, see mount_ufs(1M).
Chapter 17 Managing File Systems (Overview) 295

Mounting and Unmounting File Systems

Before you can access the les on a le system, you need to mount the le system. When you mount a le system, you attach that le system to a directory (mount point) and make it available to the system. The root (/) le system is always mounted. Any other le system can be connected or disconnected from the root (/) le system. When you mount a le system, any les or directories in the underlying mount point directory are unavailable as long as the le system is mounted. These les are not permanently affected by the mounting process. They become available again when the le system is unmounted. However, mount directories are typically empty because you usually do not want to obscure existing les. For example, the following gure shows a local le system, starting with a root (/) le system and the sbin, etc, and opt subdirectories.
/

sbin

etc

opt

mount

fs

opt

SUNWrtvc

rc0

ufs

group

shutdown

mount

system

ttydefs

FIGURE 171

Sample root (/) File System

To access a local le system from the /opt le system that contains a set of unbundled products, you must do the following:
I

First, you must create a directory to use as a mount point for the le system you want to mount, for example, /opt/unbundled. Once the mount point is created, you can mount the le system by using the mount command. This command makes all of the les and directories in /opt/unbundled available, as shown in the following gure.

296

System Administration Guide: Devices and File Systems June 2005

sbin

etc

opt

mount

fs

opt

SUNWrtvc

unbundled

rc0

ufs

group

app1

app2

file1 shutdown mount system file2 ttydefs file3

file1

file2

file3

Mount point File system

FIGURE 172

Mounting a File System

For step-by-step instructions on how to mount le systems, see Chapter 19.

The Mounted File System Table

Whenever you mount or unmount a le system, the /etc/mnttab (mount table) le is modied with the list of currently mounted le systems. You can display the contents of this le by using the cat or more commands. However, you cannot edit this le. Here is an example of an /etc/mnttab le:
$ more /etc/mnttab /dev/dsk/c0t0d0s0 / ufs rw,intr,largefiles,logging,xattr,onerror =panic,dev=2200008 1093882623 /devices /devices devfs dev=4340000 1093882603 ctfs /system/contract ctfs dev=4380001 1093882603 proc /proc proc dev=43c0000 1093882603 mnttab /etc/mnttab mntfs dev=4400001 1093882603 swap /etc/svc/volatile tmpfs xattr,dev=4440001 1093882603 /dev/dsk/c0t0d0s6 /usr ufs rw,intr,largefiles,logging,xattr,onerror Chapter 17 Managing File Systems (Overview) 297

=panic,dev=220000e 1093882623 objfs /system/object objfs dev=44c0001 1094150403 fd /dev/fd fd rw,dev=45c0001 1093882624 swap /var/run tmpfs xattr,dev=4440002 1093882625 swap /tmp tmpfs xattr,dev=4440003 1093882625 /dev/dsk/c0t0d0s7 /export/home ufs rw,intr,largefiles,logging,xattr ,onerror=panic,dev=220000f 1093882637 $

The Virtual File System Table

Manually mount le systems every time you wanted to access them would be a very time-consuming and error-prone. To avoid these problems, the virtual le system table (the /etc/vfstab le) provides a list of le systems and information on how to mount them. The /etc/vfstab le provides two important features:
I I

You can specify le systems to automatically mount when the system boots. You can mount le systems by using only the mount point name. The /etc/vfstab le contains the mapping between the mount point and the actual device slice name.

A default /etc/vfstab le is created when you install a system, depending on the selections during installation. However, you can edit the /etc/vfstab le on a system whenever you want. To add an entry, the information you need to specify is as follows:
I I I I

The device where the le system resides The le system mount point File system type Whether you want the le system to mount automatically when the system boots (by using the mountall command) Any mount options

The following is an example of an /etc/vfstab le. Comment lines begin with #. This example shows an /etc/vfstab le for a system with two disks (c0t0d0 and c0t3d0).
$ more /etc/vfstab #device device #to mount to fsck # fd /proc /dev/dsk/c0t0d0s1 /dev/dsk/c0t0d0s0 /dev/rdsk/c0t0d0s0 /dev/dsk/c0t0d0s6 /dev/rdsk/c0t0d0s6 298 mount point /dev/fd /proc / /usr FS type fd proc swap ufs ufs fsck pass 1 1 mount mount at boot options no no no no no -

System Administration Guide: Devices and File Systems June 2005

/dev/dsk/c0t0d0s7 /dev/dsk/c0t0d0s5 /devices ctfs objfs swap $

/dev/rdsk/c0t0d0s7 /dev/rdsk/c0t0d0s5 -

/export/home /opt /devices /system/contract /system/object /tmp

ufs ufs devfs ctfs objfs tmpfs

2 2 -

yes yes no no no yes

In this example, the UFS le system entry for /export/home on the /dev/dsk/c0t0d0s7 slice will be automatically mounted on the /test mount point when the system boots. Note that, for root (/) and /usr, the mount at boot eld value is specied as no. These le systems are mounted by the kernel as part of the boot sequence before the mountall command is run. For descriptions of each /etc/vfstab eld and information on how to edit and use the le, see Chapter 19.

The NFS Environment

NFS is a distributed le system service that can be used to share resources (les or directories) from one system, typically a server, with other systems on the network. For example, you might want to share third-party applications or source les with users on other systems. NFS makes the actual physical location of the resource irrelevant to the user. Instead of placing copies of commonly used les on every system, NFS allows you to place one copy on one systems disk and let all other systems access it from the network. Under NFS, remote les are virtually indistinguishable from local les. For more information, see Chapter 4, Managing Network File Systems (Overview), in System Administration Guide: Network Services. A system becomes an NFS server if it has resources to share on the network. A server keeps a list of currently shared resources and their access restrictions (such as read/write or read-only access). When you share a resource, you make it available for mounting by remote systems. You can share a resource in these ways:
I I

By using the share or shareall command By adding an entry to the /etc/dfs/dfstab (distributed le system table) le and rebooting the system

For information on how to share resources, see Chapter 19. For a complete description of NFS, see Chapter 4, Managing Network File Systems (Overview), in System Administration Guide: Network Services.

Chapter 17 Managing File Systems (Overview)

299

Automounting or AutoFS
You can mount NFS le system resources by using a client-side service called automounting (or AutoFS). AutoFS enables a system to automatically mount and unmount NFS resources whenever you access them. The resource remains mounted as long as you remain in the directory and are using a le within that directory. If the resource is not accessed for a certain period of time, it is automatically unmounted. AutoFS provides the following features:
I

NFS resources dont need to be mounted when the system boots, which saves booting time. Users dont need to know the root password to mount and unmount NFS resources. Network traffic might be reduced because NFS resources are mounted only when they are in use.

The AutoFS service is initialized by the automount utility, which runs automatically when a system is booted. The automountd daemon runs continuously and is responsible for the mounting and unmounting of NFS le systems on an as-needed basis. By default, the /home le system is mounted by the automount daemon. With AutoFS, you can specify multiple servers to provide the same le system. This way, if one of these servers is down, AutoFS can try to mount the le system from another machine. For complete information on how to set up and administer AutoFS, see System Administration Guide: IP Services.

Determining a File Systems Type

You can determine a le systems type by using one of the following:
I I I

The FS type eld in the virtual le system table (the /etc/vfstab le) The /etc/default/fs le for local le systems The /etc/dfs/fstypes le for NFS le systems

How to Determine a File Systems Type

This procedure works whether or not the le system is mounted. Determine a le systems type by using the grep command.
$ grep mount-point fs-table 300 System Administration Guide: Devices and File Systems June 2005

mount-point fs-table

Species the mount point name of the le system for which you want to know the le system type. For example, the /var directory. Species the absolute path to the le system table in which to search for the le systems type. If the le system is mounted, fs-table should be /etc/mnttab. If the le system isnt mounted, fs-table should be /etc/vfstab.

Information for the mount point is displayed.

Note If you have the raw device name of a disk slice, you can use the fstyp command to determine a le systems type (if the disk slice contains a le system). For more information, see fstyp(1M).

EXAMPLE 171

Determining a File Systems Type

The following example uses the /etc/vfstab le to determine the le system type for the /export le system.
$ grep /export /etc/vfstab /dev/dsk/c0t3d0s6 /dev/rdsk/c0t3d0s6 $ /export ufs 2 yes -

The following example uses the /etc/mnttab le to determine the le system type of the currently mounted diskette. The diskette was previously mounted by vold.
$ grep /floppy /etc/mnttab /vol/dev/diskette0/unnamed_floppy nohidden,nofoldcase,dev=16c0009 $ /floppy/unnamed_floppy 89103376 pcfs rw,

Chapter 17 Managing File Systems (Overview)

301

302

System Administration Guide: Devices and File Systems June 2005

CHAPTER

18

Creating UFS, TMPFS, and LOFS File Systems (Tasks)

This chapter describes how to create UFS, temporary (TMPFS), and loopback (LOFS) le systems. For UFS le systems, this chapter shows you how to create a le system by using the newfs command. Because TMPFS and LOFS are virtual le systems, you actually access them by mounting them. This is a list of the step-by-step instructions in this chapter.
I I I

How to Create a UFS File System on page 304 How to Create a TMPFS File System on page 310 How to Create an LOFS File System on page 312

Note For instructions on how to create UFS and DOS le systems on removable

media, see Chapter 1.

Creating a UFS File System

Before you can create a UFS le system on a disk, the disk must be formatted and divided into slices. A disk slice is a physical subset of a disk that is composed of a single range of contiguous blocks. A slice can be used either as a raw device that provides, for example, swap space, or to hold a disk-based le system. See Chapter 11 for complete information on formatting disks and dividing disks into slices. Disk and storage management products, such as Solaris Volume Manager, create more sophisticated volumes. Volumes expand beyond single-slice or single-disk boundaries. For more information about using volumes, see Solaris Volume Manager Administration Guide.

303

Note Solaris device names use the term slice (and the letter s in the device name) to

refer to the slice number. Slices are also called partitions.

You need to create UFS le systems only occasionally, because the Solaris OS automatically creates them as part of the installation process. You need to create (or re-create) a UFS le system when you want to do the following:
I I I

Add or replace disks Change the existing partitioning structure of a disk Fully restore of a le system

The newfs command is the standard way to create UFS le systems. The newfs command is a convenient front end to the mkfs command, which actually creates the new le system. The newfs command reads parameter defaults, such as tracks per cylinder and sectors per track, from the label for the disk that will contain the new le system. The options you choose are passed to the mkfs command to build the le system. For information about the default parameters that are used by the newfs command, see newfs(1M).

M
Before You Begin

How to Create a UFS File System

Ensure that you have met the following prerequisites:
I I I

The disk must be formatted and divided into slices. If you are re-creating an existing UFS le system, unmount it. You need to know the device name of the slice that will contain the le system.

For information on nding disks and disk slice numbers, see Chapter 12. For information on formatting disks and dividing disks into slices, see Chapter 11.
Steps

1. You must be superuser or assume an equivalent role. 2. Create the UFS le system.
# newfs [-N] [-b size] [-i bytes] /dev/rdsk/device-name

-N

Displays what parameters the newfs command would pass to the mkfs command without actually creating the le system. This option is a good way to test the newfs command. Species the block size for the le system, either 4096 or 8192 bytes per block. The default is 8192.

-b size

304

System Administration Guide: Devices and File Systems June 2005

-i bytes device-name

Species the number of bytes per inode. The default varies depending on the disk size. For more information, see newfs(1M). Species the disk device name on which to create the new le system.

The system asks for conrmation.

Caution Be sure you have specied the correct device name for the slice before performing this step. If you specify the wrong slice, you will erase its contents when the new le system is created. This error might cause the system to panic.

3. To verify the creation of the UFS le system, check the new le system.
# fsck /dev/rdsk/device-name

where device-name argument species the name of the disk device that contains the new le system. The fsck command checks the consistency of the new le system, reports any problems, and prompts you before it repairs the problems. For more information on the fsck command, see Chapter 22 or fsck(1M).
Example 181

Creating a UFS File System

The following example shows how to create a UFS le system on /dev/rdsk/c0t1d0s7.

# newfs /dev/rdsk/c0t1d0s7 /dev/rdsk/c0t1d0s7: 725760 sectors in 720 cylinders of 14 tracks, 72 sectors 354.4MB in 45 cyl groups (16 c/g, 7.88MB/g, 3776 i/g) super-block backups (for fsck -F ufs -o b=#) at: 32, 16240, 32448, 48656, 64864, 81072, 97280, 113488, 129696, 145904, 162112, 178320, 194528, 210736, 226944, 243152, 258080, 274288, 290496, 306704, 322912, 339120, 355328, 371536, 387744, 403952, 420160, 436368, 452576, 468784, 484992, 501200, 516128, 532336, 548544, 564752, 580960, 597168, 613376, 629584, 645792, 662000, 678208, 694416, 710624, fsck /dev/rdsk/c0t1d0s7 #

More Information

After You Create a UFS File System ...

To mount the UFS le system and make it available, go to Chapter 19.

Chapter 18 Creating UFS, TMPFS, and LOFS File Systems (Tasks)

305

How to Create a Multiterabyte UFS File System

Support for a multiterabyte UFS le system assumes the availability of multiterabyte LUNs, provided as Solaris Volume Manager or VxVM volumes, or as physical disks greater than 1 terabyte. Before you can create a multiterabyte UFS le system, verify that you have done either of the following:
I

Created a multiterabyte disk partition by using the format utility or the Solaris installation utilities Set up a multiterabyte volume with Solaris Volume Manager

For more information about multiterabyte UFS le systems, see 64-bit: Support of Multiterabyte UFS File Systems on page 281.
Steps

1. Become superuser. 2. Create a multiterabyte UFS le system on a logical volume. For example, this command creates a UFS le system for a 1.8 terabyte volume:
# newfs /dev/md/rdsk/d99 newfs: construct a new file system /dev/md/rdsk/d99: (y/n)? y /dev/md/rdsk/d99: 3859402752 sectors in 628158 cylinders of 48 tracks, 128 sectors 1884474.0MB in 4393 cyl groups (143 c/g, 429.00MB/g, 448 i/g) super-block backups (for fsck -F ufs -o b=#) at: 32, 878752, 1757472, 2636192, 3514912, 4393632, 5272352, 6151072, 702... Initializing cylinder groups: ........................................................................ super-block backups for last 10 cylinder groups at: 3850872736, 3851751456, 3852630176, 3853508896, 3854387616, 3855266336, 3856145056, 3857023776, 3857902496, 3858781216, #

3. Verify the integrity of the newly created le system. For example:

# fsck /dev/md/rdsk/d99 ** /dev/md/rdsk/d99 ** Last Mounted on ** Phase 1 - Check Blocks and Sizes ** Phase 2 - Check Pathnames ** Phase 3 - Check Connectivity ** Phase 4 - Check Reference Counts ** Phase 5 - Check Cyl groups 2 files, 2 used, 241173122 free (0 frags, 241173122 blocks, 0.0% fragmentation) #

4. Mount and verify the newly created le system.

306 System Administration Guide: Devices and File Systems June 2005

For example:
# mount /dev/md/dsk/d99 /bigdir # df -h /bigdir Filesystem size used /dev/md/dsk/d99 1.8T 64M

avail capacity 1.8T 1%

Mounted on /bigdir

How to Expand a Multiterabyte UFS File System

After a multiterabyte UFS le system is created, you can use the growfs command to expand the le system. For example, using the le system that was created for the volume in the preceding procedure, you can add another disk to this volume. Then, expand the le system.

Steps

1. Become superuser. 2. Add another disk to the volume. For example:

# metattach d99 c4t5d0s4 d99: component is attached # metastat d99: Concat/Stripe Size: 5145882624 blocks (2.4 TB) Stripe 0: Device Start Block Dbase c0t1d0s4 36864 Yes Stripe 1: Device Start Block Dbase c3t7d0s4 0 No Stripe 2: Device Start Block Dbase c1t1d0s4 0 No Stripe 3: Device Start Block Dbase c4t5d0s4 0 No

Reloc Yes Reloc Yes Reloc Yes Reloc Yes

3. Expand the le system. For example:

# growfs -v /dev/md/rdsk/d99 /usr/lib/fs/ufs/mkfs -G /dev/md/rdsk/d99 5145882624 /dev/md/rdsk/d99: 5145882624 sectors in 837546 cylinders of 48 tracks, 128 sectors 2512638.0MB in 5857 cyl groups (143 c/g, 429.00MB/g, 448 i/g) super-block backups (for fsck -F ufs -o b=#) at: 32, 878752, 1757472, 2636192, 3514912, 4393632, 5272352, 6151072, 702... Initializing cylinder groups: ......................................................................... super-block backups for last 10 cylinder groups at: Chapter 18 Creating UFS, TMPFS, and LOFS File Systems (Tasks) 307

5137130400, 5138009120, 5138887840, 5139766560, 5140645280, 5141524000, 5142402720, 5143281440, 5144160160, 5145038880, #

4. Mount and verify the expanded le system. For example:

# mount /dev/md/dsk/d99 /bigdir # df -h /bigdir Filesystem size used /dev/md/dsk/d99 2.4T 64M

avail capacity 2.4T 1%

Mounted on /bigdir

How to Expand a UFS File System to a Multiterabyte UFS File System

Use the following procedure to expand a UFS le system to greater than 1 terabyte in size. This procedure assumes that the newfs -T option was used initially to create the UFS le system.

Steps

1. Become superuser. 2. Identify the size of the current disk or volume. For example, the following volume is 800 gigabytes:
# metastat d98 d98: Concat/Stripe Size: 1677754368 blocks (800 GB) Stripe 0: Device Start Block Dbase c0t1d0s4 0 No Stripe 1: Device Start Block Dbase c3t7d0s4 0 No

Reloc Yes Reloc Yes

3. Increase the volume to greater than 1 terabyte. For example:

# metattach d98 c1t1d0s4 d98: component is attached # metastat d98 d98: Concat/Stripe Size: 2516631552 blocks (1.2 TB) Stripe 0: Device Start Block Dbase c0t1d0s4 0 No Stripe 1: Device Start Block Dbase c3t7d0s4 0 No

Reloc Yes Reloc Yes

308

System Administration Guide: Devices and File Systems June 2005

Stripe 2: Device c1t1d0s4

Start Block 0

Dbase No

Reloc Yes

4. Expand the UFS le system for the disk or volume to greater than 1 terabyte. For example:
growfs -v /dev/md/rdsk/d98 /usr/lib/fs/ufs/mkfs -G /dev/md/rdsk/d98 2516631552 /dev/md/rdsk/d98: 2516631552 sectors in 68268 cylinders of 144 tracks, 256 sectors 1228824.0MB in 2731 cyl groups (25 c/g, 450.00MB/g, 448 i/g) super-block backups (for fsck -F ufs -o b=#) at: 32, 921888, 1843744, 2765600, 3687456, 4609312, 5531168, 6453024, 737... 8296736, Initializing cylinder groups: ...................................................... super-block backups for last 10 cylinder groups at: 2507714848, 2508636704, 2509558560, 2510480416, 2511402272, 2512324128, 2513245984, 2514167840, 2515089696, 2516011552,

5. Mount and verify the expanded le system. For example:

# mount /dev/md/dsk/d98 /datadir # df -h /datadir Filesystem size used /dev/md/dsk/d98 1.2T 64M

avail capacity 1.2T 1%

Mounted on /datadir

Troubleshooting Multiterabyte UFS File System Problems

Use the following error messages and solutions to troubleshoot problems with multiterabyte UFS le systems. Error Message (similar to the following):
mount: /dev/rdsk/c0t0d0s0 is not this fstype.

Cause You attempted to mount a UFS le system that is greater than 1 terabyte on a system running a Solaris release prior to the Solaris 10 release. Solution Mount a UFS le system that is greater than 1 terabyte on a system running the Solaris 10 or later release. Error Message
"File system was not set up with the multi-terabyte format." cannot be increased to a terabyte or more." "Its size

Chapter 18 Creating UFS, TMPFS, and LOFS File Systems (Tasks)

309

Cause You attempted to expand a le system that was not created by using the newfs -T command. Solution 1. Back up the data for the le system that you want to expand to greater than 1 terabyte. 2. Re-create the le system by using the newfs command to create a multiterabyte le system. 3. Restore the backup data into the newly created le system.

Creating a Temporary File System (TMPFS)

A temporary le system (TMPFS) uses local memory for le system reads and writes, which is typically much faster than reads and writes in a UFS le system. TMPFS le systems can improve system performance by saving the cost of reading and writing temporary les to a local disk or across the network. Files in TMPFS le systems do not survive across reboots or unmounts. If you create multiple TMPFS le systems, be aware that they all use the same system resources. Files created under one TMPFS le system use up space available for any other TMPFS le system, unless you limit TMPFS sizes by using the -o size option of the mount command. For more information, see the tmpfs(7FS).

M
Steps

How to Create a TMPFS File System

1. Become superuser or assume an equivalent role. 2. Create the directory that you want to mount as the TMPFS le system, if necessary.
# mkdir /mount-point

where mount-point is the directory on which the TMPFS le system is mounted. 3. Mount the TMPFS le system.
# mount -F tmpfs [-o size=number] swap mount-point

-o size=number
310

Species the size limit of the TMPFS le system in Mbytes.

System Administration Guide: Devices and File Systems June 2005

mount-point

Species the directory on which the TMPFS le system is mounted.

To set up the system to automatically mount a TMPFS le system at boot time, see Example 183. 4. Verify that the TMPFS le system has been created.
# mount -v

Example 182

Creating a TMPFS File System

The following example shows how to create, mount, and limit the size of the TMPFS le system, /export/reports, to 50 Mbytes.
# # # # mkdir chmod mount mount /export/reports 777 /export/reports -F tmpfs -o size=50m swap /export/reports -v

Example 183

Mounting a TMPFS File System at Boot Time

You can set up the system to automatically mount a TMPFS le system at boot time by adding an /etc/vfstab entry. The following example shows an entry in the /etc/vfstab le that mounts /export/test as a TMPFS le system at boot time. Because the size=number option is not specied, the size of the TMPFS le system on /export/test is limited only by the available system resources.
swap - /export/test tmpfs yes -

For more information on the /etc/vfstab le, see Field Descriptions for the /etc/vfstab File on page 319.

Creating a Loopback File System (LOFS)

An LOFS le system is a virtual le system that provides an alternate path to an existing le system. When other le systems are mounted onto an LOFS le system, the original le system does not change. For more information, see the lofs(7FS).

Chapter 18 Creating UFS, TMPFS, and LOFS File Systems (Tasks)

311

Note Be careful when creating LOFS le systems. Because LOFS le systems are virtual le systems, the potential for confusing both users and applications is enormous.

M
Steps

How to Create an LOFS File System

1. Become superuser or assume an equivalent role. 2. Create the directory you want to mount as an LOFS le system, if necessary.
# mkdir loopback-directory

3. Grant the appropriate permissions and ownership on the newly created directory. 4. Create the mount point where you want to mount the LOFS le system, if necessary.
# mkdir /mount-point

5. Mount the LOFS le system.

# mount -F lofs loopback-directory /mount-point

loopback-directory /mount-point

Species the le system to be mounted on the loopback mount point. Species the directory on which to mount the LOFS le system.

6. Verify that the LOFS le system has been mounted.

# mount -v

Example 184

Creating and Mounting an LOFS File System

The following example shows how to create, mount, and test new software in the /new/dist directory as a loopback le system without actually having to install it.
# mkdir /tmp/newroot # mount -F lofs /new/dist /tmp/newroot # chroot /tmp/newroot newcommand

312

System Administration Guide: Devices and File Systems June 2005

Example 185

Mounting an LOFS File System at Boot Time

You can set up the system to automatically mount an LOFS le system at boot time by adding an entry to the end of the /etc/vfstab le. The following example shows an entry in the /etc/vfstab le that mounts an LOFS le system for the root (/) le system on /tmp/newroot.
/ - /tmp/newroot lofs yes -

Ensure that the loopback entries are the last entries in the /etc/vfstab le. Otherwise, if the /etc/vfstab entry for a loopback le system precedes the le systems to be included in it, the loopback le system cannot be mounted.
See Also

For more information on the /etc/vfstab le, see Field Descriptions for the /etc/vfstab File on page 319.

Chapter 18 Creating UFS, TMPFS, and LOFS File Systems (Tasks)

313

314

System Administration Guide: Devices and File Systems June 2005

CHAPTER

19

Mounting and Unmounting File Systems (Tasks)

This chapter describes how to mount and unmount le systems in the Solaris OS. This is a list of the step-by-step instructions in this chapter.
I I I I I

I I

I I I

How to Determine Which File Systems Are Mounted on page 320 How to Add an Entry to the /etc/vfstab File on page 321 How to Mount a File System (/etc/vfstab File) on page 322 How to Mount a UFS File System (mount Command) on page 323 How to Mount a UFS File System Without Large Files (mount Command) on page 324 How to Mount an NFS File System (mount Command) on page 325 x86: How to Mount a PCFS (DOS) File System From a Hard Disk (mount Command) on page 326 How to Verify a File System is Unmounted on page 328 How to Stop All Processes Accessing a File System on page 328 How to Unmount a File System on page 329

Overview of Mounting File Systems

After you create a le system, you need to make it available to the system so that you can use it. You make a le system available by mounting it, which attaches the le system to the system directory tree at the specied mount point. The root (/) le system is always mounted. The following table provides guidelines on mounting le systems based on how you use them.

315

Mount Type Needed

Suggested Mount Method

Local or remote le systems that need to be mounted infrequently Local le systems that need to be mounted frequently Remote le systems,, such as home directories, that need to be mounted frequently

The mount command that you type manually from the command line. The /etc/vfstab le, which mounts the le system automatically when the system is booted in multi user state.
I

The /etc/vfstab le, which automatically mounts the le system when the system is booted in multiuser state. AutoFS, which automatically mounts the le system when you access it or unmounts the le system when you change to another directory.

To enhance performance, you can also cache the remote le systems by using the CacheFS le system.

You can mount removable media that contains a le system by inserting the media into the drive and running the volcheck command, if necessary. For more information on mounting removable media, see Chapter 1.

Commands for Mounting and Unmounting File Systems

The following table lists the commands in the /usr/sbin directory that you use to mount and unmount le systems.
TABLE 191 Command

Commands for Mounting and Unmounting File Systems

Description Man Page

mount mountall

Mounts le systems and remote resources. Mounts all le systems that are specied in the /etc/vfstab le. The mountall command runs automatically when the system enters multiuser mode. Unmounts le systems and remote resources. Unmounts all le systems that are specied in the /etc/vfstab le.

mount(1M) mountall(1M)

umount umountall

mount(1M) mountall(1M)

Keep the following key points in mind when using the mount and mountall commands:
316 System Administration Guide: Devices and File Systems June 2005

The mount and mountall commands cannot mount a read/write le system that has known inconsistencies. If you receive an error message from the mount or mountall command, you might need to check the le system. See Chapter 22 for information on how to check the le system. The umount and umountall commands do not unmount a le system that is busy. A le system is considered busy if one of the following is true:
I I I

A user is accessing a le or directory in the le system. A program has a le open in that le system. The le system is shared.

You can use the remount option when remounting from read-only access to read-write access only. You cannot remount from read-write access to read-only access.

Commonly Used Mount Options

The following table describes the commonly used options that you can specify with the mount -o option. If you specify multiple options, separate them with commas (no spaces). For example, -o ro,nosuid. For a complete list of mount options for each le system type, refer to the specic mount man page (for example, mount_ufs(1M)).
TABLE 192

Commonly Used -o Mount Options

File System Description

mount Option

bg | fg

NFS

If the rst mount attempt fails, retries another mount in the background (bg) or in the foreground (fg). This option is safe for non critical vfstab entries. The default is fg. Species the procedure if the server does not respond. The soft option indicates that an error is returned. The hard option indicates that the retry request is continued until the server responds. The default is hard. Species whether keyboard interrupts are delivered to a hung process while waiting for a response on a hard-mounted le system. The default is intr (interrupts allowed).

hard | soft

NFS

intr | nointr

NFS

Chapter 19 Mounting and Unmounting File Systems (Tasks)

317

TABLE 192

Commonly Used -o Mount Options

File System

(Continued)

mount Option

Description

largefiles | nolargefiles

UFS

Enables you to create les larger than 2 Gbytes. The largefiles option means that a le system mounted with this option might contain les larger than 2 Gbytes. If the nolargefiles option is specied, the le system cannot be mounted on a system that is running Solaris 2.6 or compatible versions. The default is largefiles. Enables or disables logging for the le system. UFS logging is the process of storing transactions (changes that comprise a complete UFS operation) into a log before the transactions are applied to the UFS le system. Logging helps prevent UFS le systems from becoming inconsistent, which means fsck can be bypassed. Bypassing fsck reduces the time to reboot a system if it crashes, or after a system is shut down uncleanly. The log is allocated from free blocks on the le system, and is sized at about 1 Mbyte per 1 Gbyte of le system, up to a maximum of 64 Mbytes. The default is logging.

logging | nologging

UFS

atime | noatime

UFS

Suppresses access time updates on les, except when they coincide with updates to the time of the last le status change or the time of the last le modication. For more information, see stat(2). This option reduces disk activity on le systems where access times are unimportant (for example, a Usenet news spool). The default is normal access time (atime) recording. Changes the mount options associated with an already-mounted le system. This option can generally be used with any option except ro. However, what can be changed with this option depends on the le system type. Retries the mount operation when it fails. n is the number of times to retry. Species read/write (rw) or read-only (ro). If you do not specify this option, the default is rw. The default option for HSFS is ro. Allows or disallows setuid execution. The default is to allow setuid execution.

remount

All

retry=n

NFS CacheFS, NFS, PCFS, UFS, HSFS CacheFS, HSFS, NFS, UFS

ro | rw

suid | nosuid

318

System Administration Guide: Devices and File Systems June 2005

Field Descriptions for the /etc/vfstab File

An entry in the /etc/vfstab le has seven elds, which are described in the following table.
TABLE 193 Field Name

Field Descriptions for the /etc/vfstab File

Description

device to mount

This eld identies one of the following: The block device name for a local UFS le system (for example, /dev/dsk/c0t0d0s0). I The resource name for a remote le system (for example, myserver:/export/home). For more information about NFS, see System Administration Guide: IP Services. I The block device name of the slice on which to swap (for example, /dev/dsk/c0t3d0s1). I A directory for a virtual le system.
I

device to fsck

The raw (character) device name that corresponds to the UFS le system identied by the device to mount eld (for example, /dev/rdsk/c0t0d0s0). This eld determines the raw interface that is used by the fsck command. Use a dash (-) when there is no applicable device, such as for a read-only le system or a remote le system. Identies where to mount the le system (for example, /usr).

mount point

FS type fsck pass

Identies the type of le system. The pass number used by the fsck command to decide whether to check a le system. When the eld contains a dash (-), the le system is not checked. When the eld contains a zero, UFS le systems are not checked. However, non-UFS le systems are checked. When the eld contains a value greater than zero, the le system is always checked. All le systems with a value of 1 in this eld are checked one at a time in the order they appear in the vfstab le. When the fsck command is run on multiple UFS le systems that have fsck pass values greater than 1 and the preen option (-o p) is used, the fsck command automatically checks the le systems on different disks in parallel to maximize efficiency. Otherwise, the value of the pass number does not have any effect.

mount at boot

Set to yes or no for whether the le system should be automatically mounted by the mountall command when the system is booted. Note that this eld has nothing to do with AutoFS. The root (/), /usr and /var le systems are not mounted from the vfstab le initially. This eld should always be set to no for these le systems and for virtual le systems such as /proc and /dev/fd.

Chapter 19 Mounting and Unmounting File Systems (Tasks)

319

TABLE 193 Field Name

Field Descriptions for the /etc/vfstab File

Description

(Continued)

mount options

A list of comma-separated options (with no spaces) that are used for mounting the le system. Use a dash (-) to indicate no options. For a list of commonly used mount options, see Table 192.

Note You must have an entry in each eld in the /etc/vfstab le. If there is no value for a eld, be sure to specify a dash (-). Otherwise, the system might not boot successfully. Similarly, white space should not be used as a eld value.

Mounting File Systems

The following sections describe how to mount a le system by adding an entry in the /etc/vfstab le or by using the mount command from the command line.

How to Determine Which File Systems Are Mounted

You can determine which le systems are already mounted by using the mount command.
$ mount [ -v ]

The -v displays the list of mounted le systems in verbose mode.

EXAMPLE 191

Determining Which File Systems Are Mounted

This example shows how to use the mount command to display information about the le systems that are currently mounted.
$ mount / on /dev/dsk/c0t0d0s0 read/write/setuid/intr/largefiles/xattr/onerror=... /devices on /devices read/write/setuid/dev=46c0000 on Thu Sep ... /system/contract on ctfs read/write/setuid/devices/dev=43c0001 ... /usr on /dev/dsk/c0t0d0s6 read/write/setuid/intr/largefiles/xattr/... /proc on /proc read/write/setuid/dev=4700000 on Thu Sep 2 ... /etc/mnttab on mnttab read/write/setuid/dev=47c0000 on Thu Sep 2 ... /etc/svc/volatile on swap read/write/setuid/devices/xattr/dev=4480001 ... /system/object on objfs read/write/setuid/devices/dev=44c0001 ... /dev/fd on fd read/write/setuid/dev=4800000 on Thu Sep 2 ... /var/run on swap read/write/setuid/xattr/dev=1 on Thu Sep 2 ... 320 System Administration Guide: Devices and File Systems June 2005

EXAMPLE 191

Determining Which File Systems Are Mounted

(Continued)

/tmp on swap read/write/setuid/xattr/dev=2 on Thu Sep 2 ... /stuff on /dev/dsk/c0t0d0s5 read/write/setuid/intr/largefiles/xattr... /export/home on /dev/dsk/c0t0d0s7 read/write/setuid/intr/largefiles/... /home/rimmer on pluto:/export/home/rimmer remote/read/write/setuid/xattr/... $

M
Steps

How to Add an Entry to the /etc/vfstab File

1. Become superuser or assume an equivalent role. 2. Create a mount point for the le system to be mounted, if necessary.
# mkdir /mount-point

There must be a mount point on the local system to mount a le system. A mount point is a directory to which the mounted le system is attached. 3. Edit the /etc/vfstab le and add an entry. Ensure that you do the following: a. Separate each eld with white space (a space or a tab). b. Specify a dash (-) if a eld has no contents. c. Save the changes. For detailed information about the /etc/vfstab eld entries, see Table 193.
Note Because the root (/) le system is mounted read-only by the kernel during

the boot process, only the remount option (and options that can be used in conjunction with remount) affect the root (/) entry in the /etc/vfstab le.

Example 192

Adding an Entry to the /etc/vfstab File

The following example shows how to mount the disk slice /dev/dsk/c0t3d0s7 as a UFS le system to the mount point /files1. The raw character device /dev/rdsk/c0t3d0s7 is specied as the device to fsck. The fsck pass value of 2 means that the le system will be checked, but not sequentially.

#device device mount #to mount to fsck point # /dev/dsk/c0t3d0s7 /dev/rdsk/c0t3d0s7 /files1

FS type ufs

fsck pass 2

mount at boot yes

mount options -

The following example shows how to mount the /export/man directory from the system pluto as an NFS le system on mount point /usr/man. Neither a device to fsck nor a fsck pass is specied because its an NFS le system. In this example, mount options are ro (read-only) and soft.
Chapter 19 Mounting and Unmounting File Systems (Tasks) 321

#device device #to mount to fsck pluto:/export/man -

mount FS point type /usr/man nfs

fsck pass -

mount at boot yes

mount options ro,soft

The following example shows how to mount the root (/) le system on a loopback mount point, /tmp/newroot. LOFS le systems must always be mounted after the le systems that are in the LOFS le system.
#device #to mount # / device to fsck mount point FS type fsck pass mount at boot yes mount options -

/tmp/newroot lofs -

M
Steps

How to Mount a File System (/etc/vfstab File)

1. Become superuser or assume an equivalent role. 2. Mount a le system listed in the /etc/vfstab le.
# mount /mount-point

where /mount-point species an entry in the mount point or device to mount eld in the /etc/vfstab le. It is usually easier to specify the mount point.
Example 193

Mounting a File System (/etc/vfstab File)

The following example shows how to mount the /usr/dist le system that is listed in the /etc/vfstab le.
# mount /usr/dist

Example 194

Mounting All File Systems (/etc/vfstab File)

The following example shows the messages that are displayed when you use the mountall command and the le systems are already mounted.
# mountall /dev/rdsk/c0t0d0s7 already mounted mount: /tmp already mounted mount: /dev/dsk/c0t0d0s7 is already mounted, /export/home is busy, or the allowable number of mount points has been exceeded

When using the mountall command, all the le systems with a device to fsck entry are checked and xed, if necessary, before they are mounted. The following example shows how to mount all the local systems that are listed in the /etc/vfstab le.
# mountall -l # mount / on /dev/dsk/c0t0d0s0 read/write/setuid/intr/largefiles/xattr/onerror=... 322 System Administration Guide: Devices and File Systems June 2005

/devices on /devices read/write/setuid/dev=46c0000 on Thu Sep ... /system/contract on ctfs read/write/setuid/devices/dev=43c0001 ... /usr on /dev/dsk/c0t0d0s6 read/write/setuid/intr/largefiles/xattr/... /proc on /proc read/write/setuid/dev=4700000 on Thu Sep 2 ... /etc/mnttab on mnttab read/write/setuid/dev=47c0000 on Thu Sep 2 ... /etc/svc/volatile on swap read/write/setuid/devices/xattr/dev=4480001 ... /system/object on objfs read/write/setuid/devices/dev=44c0001 ... /dev/fd on fd read/write/setuid/dev=4800000 on Thu Sep 2 ... /var/run on swap read/write/setuid/xattr/dev=1 on Thu Sep 2 ... /tmp on swap read/write/setuid/xattr/dev=2 on Thu Sep 2 ... /stuff on /dev/dsk/c0t0d0s5 read/write/setuid/intr/largefiles/xattr... /export/home on /dev/dsk/c0t0d0s7 read/write/setuid/intr/largefiles/...

The following example shows how to mount all the remote le systems that are listed in the /etc/vfstab le.
# mountall -r # mount / on /dev/dsk/c0t0d0s0 read/write/setuid/intr/largefiles/xattr/onerror=... /devices on /devices read/write/setuid/dev=46c0000 on Thu Sep ... /system/contract on ctfs read/write/setuid/devices/dev=43c0001 ... /usr on /dev/dsk/c0t0d0s6 read/write/setuid/intr/largefiles/xattr/... /proc on /proc read/write/setuid/dev=4700000 on Thu Sep 2 ... /etc/mnttab on mnttab read/write/setuid/dev=47c0000 on Thu Sep 2 ... /etc/svc/volatile on swap read/write/setuid/devices/xattr/dev=4480001 ... /system/object on objfs read/write/setuid/devices/dev=44c0001 ... /dev/fd on fd read/write/setuid/dev=4800000 on Thu Sep 2 ... /var/run on swap read/write/setuid/xattr/dev=1 on Thu Sep 2 ... /tmp on swap read/write/setuid/xattr/dev=2 on Thu Sep 2 ... /stuff on /dev/dsk/c0t0d0s5 read/write/setuid/intr/largefiles/xattr... /stuff on /dev/dsk/c0t0d0s5 read/write/setuid/intr/largefiles/xattr... /export/home on /dev/dsk/c0t0d0s7 read/write/setuid/intr/largefiles/... /home/rimmer on pluto:/export/home/rimmer remote/read/write/setuid/xattr/...

How to Mount a UFS File System (mount Command)

1. Become superuser or assume an equivalent role. 2. Create a mount point for the le system to be mounted, if necessary.
# mkdir /mount-point

Steps

There must be a mount point on the local system to mount a le system. A mount point is a directory to which the mounted le system is attached. 3. Mount the UFS le system.
# mount [-o mount-options] /dev/dsk/device-name /mount-point

-o mount-options

Species mount options that you can use to mount a UFS le system. For a list of options, see Table 192 or mount_ufs(1M).
Chapter 19 Mounting and Unmounting File Systems (Tasks) 323

/dev/dsk/device-name

Species the disk device name for the slice that contains the le system (for example, /dev/dsk/c0t3d0s7). To view slice information for a disk, see How to Display Disk Slice Information on page 199. Species the directory on which to mount the le system.

/mount-point

Example 195

Mounting a UFS File System (mount Command)

The following example shows how to mount /dev/dsk/c0t3d0s7 on the /files1 directory.
# mount /dev/dsk/c0t3d0s7 /files1

How to Mount a UFS File System Without Large Files (mount Command)
When you mount a le system, the largefiles option is selected by default. This option enables you to create les larger than 2 Gbytes. Once a le system contains large les, you cannot remount the le system with the nolargefiles option or mount it on a system that is running Solaris 2.6 or compatible versions, until you remove any large les and run the fsck command to reset the state to nolargefiles. This procedure assumes that the le system is in the /etc/vfstab le.

Steps

1. Become superuser or assume an equivalent role. 2. Create a mount point for the le system to be mounted, if necessary.
# mkdir /mount-point

There must be a mount point on the local system to mount a le system. A mount point is a directory to which the mounted le system is attached. 3. Ensure that no large les exist in the le system.
# cd /mount-point # find . -xdev -size +20000000 -exec ls -l {} \;

where /mount-point identies the mount point of the le system you want to check for large les. 4. Remove or move any large les in this le system to another le system, if necessary.

324

System Administration Guide: Devices and File Systems June 2005

5. Unmount the le system.

# umount /mount-point

6. Reset the le system state.

# fsck /mount-point

7. Remount the le system with the nolargefiles option.

# mount -o nolargefiles /mount-point

Example 196

Mounting a File System Without Large Files (mount Command)

The following example shows how to check the /datab le system and remount it with the nolargefiles option.
# # # # # cd /datab find . -xdev -size +20000000 -exec ls -l {} \; umount /datab fsck /datab mount -o nolargefiles /datab

How to Mount an NFS File System (mount Command)

1. Become superuser or assume an equivalent role. 2. Create a mount point for the le system to be mounted, if necessary.
# mkdir /mount-point

Steps

There must be a mount point on the local system to mount a le system. A mount point is a directory to which the mounted le system is attached. 3. Ensure that the resource (le or directory) is available from a server. To mount an NFS le system, the resource must be made available on the server by using the share command. For information on how to share resources, see About the NFS Service in System Administration Guide: Network Services. 4. Mount the NFS le system.
# mount -F nfs [-o mount-options] server:/directory /mount-point

-o mount-options

Species mount options that you can use to mount an NFS le system. See Table 192 for the list of commonly used mount options or mount_nfs(1M) for a complete list of options.

Chapter 19 Mounting and Unmounting File Systems (Tasks)

325

server:/directory /mount-point
Example 197

Species the servers host name that contains the shared resource, and the path to the le or directory to mount. Species the directory on which to mount the le system.

Mounting an NFS File System (mount Command)

The following example shows how to mount the /export/packages directory on /mnt from the server pluto.
# mount -F nfs pluto:/export/packages /mnt

x86: How to Mount a PCFS (DOS) File System From a Hard Disk (mount Command)
Use the following procedure to mount a PCFS (DOS) le system from a hard disk. 1. Become superuser or assume an equivalent role. 2. Create a mount point for the le system to be mounted, if necessary.
# mkdir /mount-point

Steps

There must be a mount point on the local system to mount a le system. A mount point is a directory to which the mounted le system is attached. 3. Mount the PCFS le system.
# mount -F pcfs [-o rw | ro] /dev/dsk/device-name:logical-drive /mount-point

-o rw | ro

Species that you can mount a PCFS le system read/write (rw) or read-only (ro). If you do not specify this option, the default is rw. Species the device name of the whole disk (for example, /dev/dsk/c0t0d0p0). Species either the DOS logical drive letter (c through z) or a drive number (1 through 24). Drive c is equivalent to drive 1 and represents the primary DOS slice on the drive. All other letters or numbers represent DOS logical drives within the extended DOS slice. Species the directory on which to mount the le system.

/dev/dsk/device-name logical-drive

/mount-point

Note that the device-name and logical-drive must be separated by a colon.

326

System Administration Guide: Devices and File Systems June 2005

Example 198

x86: Mounting a PCFS (DOS) File System From a Hard Disk (mount Command)
The following example shows how to mount the logical drive in the primary DOS slice on the /pcfs/c directory.
# mount -F pcfs /dev/dsk/c0t0d0p0:c /pcfs/c

The following example shows how to mount read-only the rst logical drive in the extended DOS slice on the /mnt directory.
# mount -F pcfs -o ro /dev/dsk/c0t0d0p0:2 /mnt

Unmounting File Systems

The unmounting of a le system removes it from the le system mount point, and deletes the entry from the /etc/mnttab le. Some le system administration tasks cannot be performed on mounted le systems. You should unmount a le system when the following occurs:
I

The le system is no longer needed or has been replaced by a le system that contains more current software. You need to check and repair the le system by using the fsck command. For more information about the fsck command, see Chapter 22. File systems should be unmounted before doing a complete backup. For more information about doing backups, see Chapter 25.
Note File systems are automatically unmounted as part of the system shutdown procedure.

In an emergency situation, you can use the umount -f option to forcibly unmount a busy le system. This practice is not recommended under normal circumstances because the unmounting of a le system with open les could cause a loss of data. This option is only available for UFS and NFS le systems.

Prerequisites for Unmounting File Systems

The prerequisites for unmounting le systems include the following:
I

You must be superuser or assume an equivalent role.

Chapter 19 Mounting and Unmounting File Systems (Tasks) 327

A le system must be available for unmounting. You cannot unmount a le system that is busy. A le system is considered busy if a user is accessing a directory in the le system, if a program has a le open in that le system, or if the le system is being shared. You can make a le system available for unmounting by doing the following:
I I I

Changing to a directory in a different le system. Logging out of the system. Using the fuser command to list all processes that are accessing the le system and to stop them, if necessary. For more details, see How to Stop All Processes Accessing a File System on page 328. Notify users if you need to unmount a le system that they are using. Unsharing the le system. For information about unsharing a le system, see unshare(1M).

How to Verify a File System is Unmounted

To verify that you unmounted a le system or a number of le systems, examine the output from the mount command.
$ mount | grep unmounted-le-system $

M
Steps

How to Stop All Processes Accessing a File System

1. Become superuser or assume an equivalent role. 2. List all the processes that are accessing the le system so that you know which processes you are going to stop.
# fuser -c [ -u ] /mount-point

-c -u /mount-point

Reports on les that are mount points for le systems and any les within those mounted le systems. Displays the user login name for each process ID. Species the name of the le system for which you want to stop processes.

3. Stop all processes that are accessing the le system.

# fuser -c -k /mount-point

A SIGKILL is sent to each process that is using the le system.

328

System Administration Guide: Devices and File Systems June 2005

Note You should not stop a users processes without rst warning the user.

4. Verify that no processes are accessing the le system.

# fuser -c /mount-point

Example 199

Stopping All Processes Accessing a File System

The following example shows how to stop process 4006c that is using the /export/home le system.
# fuser -c /export/home /export/home: 4006c # fuser -c -k /export/home /export/home: 4006c # fuser -c /export/home /export/home:

How to Unmount a File System

Use the following procedure to unmount a le system, except for the root (/), /usr, or /var le systems.
Note The root (/), /usr, and /var le systems can be unmounted only during a shutdown. The system needs these le systems to function.

Steps

1. Ensure that you have met the prerequisites listed in Prerequisites for Unmounting File Systems on page 327. 2. Unmount the le system.
# umount /mount-point

where /mount-point is the name of the le system that you want to unmount. This can be one of the following:
I I I I

The directory name where the le system is mounted The device name path of the le system The resource for an NFS le system The loopback directory for an LOFS le system

Chapter 19 Mounting and Unmounting File Systems (Tasks)

329

Example 1910

Unmounting a File System

The following example shows how to unmount a local home le system.
# umount /export/home

The following example shows how to unmount the le system on slice 7.

# umount /dev/dsk/c0t0d0s7

The following example shows how to forcibly unmount the /export le system.
# umount -f /export #

The following example shows how to unmount all le systems in the /etc/vfstab le, except for the root (/), /proc, /var, and /usr le systems.
# umountall

All le systems are unmounted, except for those le systems that are busy.

330

System Administration Guide: Devices and File Systems June 2005

CHAPTER

20

Using The CacheFS File System (Tasks)

This chapter describes how to set up and maintain CacheFS le systems. This is a list of task maps in this chapter.
I I I I I

High-Level View of Using the CacheFS File System (Task Map) on page 331 Creating and Mounting a CacheFS File System (Task Map) on page 334 Maintaining a CacheFS File System (Task Map) on page 339 Packing a Cached File System (Task Map) on page 345 Collecting CacheFS Statistics (Task Map) on page 354

For information on troubleshooting CacheFS errors, see Troubleshooting cachefspack Errors on page 350.
Note For important information about NFS version 4 and the CacheFS software, see NFS Version 4 and CacheFS Compatibility Issues on page 280.

High-Level View of Using the CacheFS File System (Task Map)

Use this task map to identify all the tasks for using CacheFS le systems. Each task points to a series of additional tasks such as creating and mounting CacheFS le systems, and packing and maintaining the cache.

331

Task

Description

For Instructions

1. Create and mount a CacheFS le system. 2. Maintain a CacheFS le system.

Create the cache and mount the le system in the cache. Display and modify a CacheFS le system by unmounting, removing, or re-creating the cache.

Creating and Mounting a CacheFS File System (Task Map) on page 334 Maintaining a CacheFS File System (Task Map) on page 339

3. (Optional) Pack and unpack Determine whether you want Packing a Cached File a CacheFS le system. to pack the cache and use System (Task Map) on page packing lists. Packing the 345 cache ensures that certain les and directories are always updated in the cache. 4. Collect CacheFS statistics. Determine cache performance Collecting CacheFS Statistics and appropriate cache size. (Task Map) on page 354

Overview of the CacheFS File System

The CacheFS le system is a general purpose caching mechanism that improves NFS server performance and scalability by reducing server and network load. Designed as a layered le system, the CacheFS le system provides the ability to cache one le system on another le system. In an NFS environment, the CacheFS le system increases the client per server ratio, reduces server and network loads, and improves performance for clients on slow links, such as Point-to-Point Protocol (PPP).

How a CacheFS File System Works

You create a CacheFS le system on a client system so that le systems you cache can be accessed by the client locally instead of across the network. The following gure shows the relationship of the components that are involved in using CacheFS le systems.

332

System Administration Guide: Devices and File Systems June 2005

Server

Back file system Network

Client

Cached file systems

FIGURE 201

How a CacheFS File System Works

The back le system is the le system that you specify to be mounted in the cache. A back le system can be either NFS or HSFS (High Sierra File System). When the user attempts to access les that are part of the back le system, those les are placed in the cache. The front le system is the le system that is mounted in the cache and is accessed from the local mount point. The front le system type must be UFS. To the user, the initial request to access a le in a CacheFS le system might seem slow. However, subsequent uses of the same le are faster.

CacheFS File System Structure and Behavior

Each cache has a set of parameters that determines the cache structure and how it behaves. The parameters are set to the default values listed in the following table. The default values specify that the entire front le system is used for caching, which is the recommended method of caching le systems.
TABLE 201

CacheFS File System Parameters and Their Default Values

Default Value Denition

CacheFS File System Parameter

maxblocks

90 %

Sets the maximum number of blocks that a CacheFS le system is allowed to claim within the front le system.

Chapter 20 Using The CacheFS File System (Tasks)

333

TABLE 201

CacheFS File System Parameters and Their Default Values

Default Value Denition

(Continued)

CacheFS File System Parameter

minblocks

0%

Sets the minimum number of blocks that a CacheFS le system is allowed to claim within the front le system. Sets the number of blocks that must be available in the front le system before a CacheFS le system can claim more than the blocks specied by minblocks. Sets the maximum number of available inodes (number of les) that a CacheFS le system is allowed to claim within the front le system. Sets the minimum number of available inodes that a CacheFS le system is allowed to claim within the front le system. Sets the number of inodes that must be available in the front le system before a CacheFS le system can claim more les than is specied in minfiles.

threshblocks

85 %

maxfiles

90 %

minfiles

0%

threshfiles

85 %

Typically, you should not change any of these parameter values. They are set to default values to achieve optimal cache behavior. However, you might want to modify the maxblocks and maxfiles values if you have some room in the front le system that is not used by the cache, and you want to use it for some other le system. You do so by using the cfsadmin command. For example:
$ cfsadmin -o maxblocks=60

Creating and Mounting a CacheFS File System (Task Map)

Use the procedures in this task map to create and mount a CacheFS le system.
Task Description For Instructions

1. Share the le system to be cached. 2. Create the cache.

Verify that the le system you want to cache is shared. Use the cfsadmin command to create the cache.

share(1M) How to Create the Cache on page 335

334

System Administration Guide: Devices and File Systems June 2005

Task

Description

For Instructions

3. Mount a le system in the cache.

Mount a le system in a cache by using one of the following methods: Mount a CacheFS le system by using the mount command. Mount a CacheFS le system by editing the /etc/vfstab le. How to Mount a CacheFS File System (mount) on page 336 How to Mount a CacheFS File System (/etc/vfstab) on page 338

Mount a cached le system by How to Mount a CacheFS using AutoFS. File System (AutoFS) on page 339

M
Steps

How to Create the Cache

1. Become superuser on the client system. 2. Create the cache.
# cfsadmin -c /cache-directory

where cache-directory indicates the name of the directory where the cache resides. For more information, see cfsadmin(1M).
Note After you have created the cache, do not perform any operations within the cache directory itself. Doing so could cause conicts within the CacheFS software.

Example 201

Creating the Cache

The following example shows how to create a cache in the /local/mycache directory by using the default cache parameter values.
# mkdir /local # cfsadmin -c /local/mycache

Mounting a File System in the Cache

You specify a le system to be mounted in the cache so that users can locally access les in that le system. The les do not actually get placed in the cache until the user accesses the les.
Chapter 20 Using The CacheFS File System (Tasks) 335

The following table describes three ways to mount a CacheFS le system.

Mount Type for CacheFS File System

Frequency of CacheFS Mount Type

Using the mount command Editing the /etc/vfstab le Using AutoFS

Every time the system reboots in order to access the same le system. Only once. The /etc/vfstab le remains unchanged after the system reboots. Only once. AutoFS maps remain unchanged after the system reboots.

Choose the method of mounting le systems that best suits your environment. You can mount only le systems that are shared. For information on sharing le systems, see share(1M).
Note The caching of the root (/) and /usr le systems is not supported in a CacheFS le system.

M
Steps

How to Mount a CacheFS File System (mount)

1. Become superuser on the client system. 2. Create the mount point, if necessary.
# mkdir /mount-point

You can create the mount point from anywhere, but it must be a UFS le system. The CacheFS options used with the mount command, as shown in the next step, determine that the mount point you create is cached in the cache directory you specify. 3. Mount a le system in the cache.
# mount -F cachefs -o backfstype=fstype,cachedir=/cache-directory[,options] /back-lesystem /mount-point

fstype /cache-directory

Indicates the le system type of the back le system, which can be either NFS or HSFS. Indicates the name of the UFS directory where the cache resides. This name is the same name you specied when you created the cache in How to Create the Cache on page 335.

336

System Administration Guide: Devices and File Systems June 2005

options

Species other mount options that you can include when you mount a le system in a cache. For a list of CacheFS mount options, see mount_cachefs(1M). Species the mount point of the back le system to cache. If the back le system is an NFS le system, you must specify the host name of the server from which you are mounting the le system and the name of the le system to cache, separated by a colon. For example, merlin: /data/abc. Indicates the directory where the le system is mounted.

/back-lesystem

/mount-point

4. Verify that the cache you created was actually mounted.

# cachefsstat /mount-point

The /mount-point is the CacheFS le system that you created. For example:
# cachefsstat /docs /docs cache hit rate: consistency checks: modifies: garbage collection:

100% (0 hits, 0 misses) 1 (1 pass, 0 fail) 0 0

If the le system was not mounted in the cache, an error message similar to the following is displayed:
# cachefsstat /mount-point cachefsstat: mount-point: not a cachefs mountpoint

For more information about the cachefsstat command, see Collecting CacheFS Statistics on page 354.
Example 202

Mounting a CacheFS File System (mount)

The following example shows how to mount the NFS le system merlin:/docs as a CacheFS le system named /docs in the cache named /local/mycache.

# mkdir /docs # mount -F cachefs -o backfstype=nfs,cachedir=/local/mycache merlin:/docs /docs

The following example shows how to make a Solaris 9 SPARC CD (HSFS le system) available as a CacheFS le system named /cfssrc. Because you cannot write to the CD, the ro argument is specied to make the CacheFS le system read-only. This example assumes that the vold daemon is not running.
# mount -F hsfs -o ro /dev/dsk/c0t6d0s0 /sol9 # mount -F cachefs -o backfstype=hsfs,cachedir=/cfs/cache,ro,noconst, backpath=/sol9 /dev/dsk/c0t6d0s0 /cfssrc # ls /cfssrc Copyright Solaris_9

Chapter 20 Using The CacheFS File System (Tasks)

337

The following example shows how to mount a Solaris 9 SPARC CD as a CacheFS le system with vold running.
# mount -F cachefs -o backfstype=hsfs,cachedir=/cfs/cache,ro,noconst, backpath=/cdrom/sol_9_sparc/s0 /vol/dev/dsk/c0t2d0/sol_9_sparc/s0 /cfssrc

The following example shows how to mount a CD as a CacheFS le system with vold running.
# mount -F cachefs -o backfstype=hsfs,cachedir=/cfs/cache,ro,noconst, backpath=/cdrom/epson /vol/dev/dsk/c0t2d0/epson /drvrs

The following example uses the demandconst option to specify consistency checking on demand for the NFS CacheFS le system /docs, whose back le system is merlin:/docs. For more information, see Consistency Checking of a CacheFS File System on page 342.
# mount -F cachefs -o backfstype=nfs,cachedir=/local/mycache,demandconst merlin:/docs /docs

How to Mount a CacheFS File System (/etc/vfstab)

1. Become superuser on the client system. 2. Using an editor, specify the le systems to be mounted in the /etc/vfstab le. See the example that follows. For more information on the /etc/vfstab le, see Field Descriptions for the /etc/vfstab File on page 319. 3. Mount the CacheFS le system.
# mount /mount-point

Steps

Or, reboot the system.

Example 203

Mounting a CacheFS File System (/etc/vfstab)

The following example shows the /etc/vfstab entry for the /data/abc directory from the remote system starbug that is mounted in the cached directory, /opt/cache.

#device device mount FS fsck #to mount to fsck point type pass # starbug:/data/abc /local/abc /opt/cache cachefs 7 nosuid,demandconst,backfstype=nfs,cachedir=/opt/cache

mount mount at boot options yes local-access,bg,

338

System Administration Guide: Devices and File Systems June 2005

How to Mount a CacheFS File System (AutoFS)

You can mount a le system in a cache with AutoFS by specifying the -fstype=cachefs mount option in your automount map. Note that the CacheFS mount options, for example, backfstype and cachedir, are also specied in the automount map. For details on automount maps, see Task Overview for Autofs Administration in System Administration Guide: Network Services or automount(1M).

Steps

1. Become superuser on the client system. 2. Using an editor, add the following line to the auto_direct map:
/mount-point -fstype=cachefs,cachedir=/directory,backfstype=nfs server:/le-system

3. Using an editor, add the following line to the auto_master map:

/-

The /- entry is a pointer to check the auto_direct map. 4. Reboot the system. 5. Verify that the entry was made correctly by changing to the le system you mounted in the cache, and then list the contents.
# cd /lesystem # ls

Example 204

Mounting a CacheFS File System (AutoFS)

The following auto_direct entry automatically mounts the CacheFS le system in the /docs directory.
/docs -fstype=cachefs,cachedir=/local/mycache,backfstype=nfs merlin:/docs

Maintaining a CacheFS File System (Task Map)

After a CacheFS le system is set up, it requires little maintenance. Use the optional procedures in this task map if you need to perform maintenance tasks on your CacheFS le systems.

Chapter 20 Using The CacheFS File System (Tasks)

339

Task

Description

For Instructions

Modify a CacheFS le system. Modify CacheFS le system behavior by unmounting, deleting, or re-creating the cache. Display CacheFS le system information.

Modifying a CacheFS File System on page 340

Display information about How to Display Information CacheFS le systems by using About a CacheFS File System the cfsadmin command. on page 341 How to Specify Cache Consistency Checking on Demand on page 342 How to Delete a CacheFS File System on page 342

Perform consistency checking. Perform consistency checking on demand by using the cfsadmin command. Delete a CacheFS le system. Delete a CacheFS le system by using the umount command and the cfsadmin command. Check the integrity of a CacheFS le system by using the fsck_cachefs command.

Check the integrity of a CacheFS le system.

How to Check the Integrity of a CacheFS File System on page 344

Maintaining a CacheFS File System

This section describes how to maintain a CacheFS le system. If you are using the /etc/vfstab le to mount le systems, you modify the cache by editing the le system options in the /etc/vfstab le. If you are using AutoFS, you modify the cache by editing the le system options in the AutoFS maps.

Modifying a CacheFS File System

When you modify a le system in the cache, you need to delete the cache and then re-create it. You might also need to reboot your machine in single-user mode, depending on how your le systems are shared and accessed. In the following example, the cache is deleted, re-created, and then mounted again by using demandconst option specied for the /docs le system.
# shutdown -g30 -y . . . 340 System Administration Guide: Devices and File Systems June 2005

Root password for system maintenance (control-d to bypass): single-user privilege assigned to /dev/console. . . . Here is where you might be prompted to run fsck on the le system where the cache is located. # fsck /local # mount /local # cfsadmin -d all /local/mycache # cfsadmin -c /local/mycache # init 6 . . . console login: password: # mount -F cachefs -o backfstype=nfs,cachedir=/local/cache1,demandconst merlin:/docs /docs #

How to Display Information About a CacheFS File System

1. Become superuser on the client system. 2. Display information about all le systems cached under a specied cache.
# cfsadmin -l /cache-directory

Steps

where /cache-directory is the name of the directory where the cache resides.
Example 205

Displaying Information About CacheFS File Systems

The following example shows information about the /local/mycache cache directory. In this example, the /docs le system is cached in /local/mycache. The last line displays the name of the CacheFS le system.
# cfsadmin -l /local/mycache cfsadmin: list cache FS information maxblocks 90% minblocks 0% threshblocks 85% maxfiles 90% minfiles 0% threshfiles 85% maxfilesize 3MB merlin:_docs:_docs # Chapter 20 Using The CacheFS File System (Tasks) 341

Consistency Checking of a CacheFS File System

To ensure that the cached directories and les remain current, the CacheFS software periodically checks the consistency of les stored in the cache. To check consistency, the CacheFS software compares the current modication time to the previous modication time. If the modication times are different, all data and attributes for the directory or le are purged from the cache. Then, new data and attributes are retrieved from the back le system.

Consistency Checking on Demand

Consistency checks can be performed only when you explicitly request checks for le systems that are mounted by using the -o demandconst option. If you mount a le system in a cache with this option, then use the cfsadmin command with the -s option to request a consistency check. By default, consistency checking is performed le by le as the les are accessed. If no les are accessed, no checks are performed. Using the -o demandconst option avoids the situation where the network is ooded with consistency checks. For more information, see mount_cachefs(1M).

How to Specify Cache Consistency Checking on Demand

1. Become superuser on the client system. 2. Mount the le system in the cache and specify cache consistency checking.
# mount -F cachefs -o backfstype=nfs,cachedir=/directory,demandconst server:/le-system /mount-point

Steps

3. Initiate consistency checking on a specic CacheFS le system.

# cfsadmin -s /mount-point

M
Steps

How to Delete a CacheFS File System

1. Become superuser on the client system. 2. Unmount the CacheFS le system.
# umount /mount-point

where /mount-point species the CacheFS le system that you want to delete.
342 System Administration Guide: Devices and File Systems June 2005

3. Determine the name of the CacheFS le system (cache ID).

# cfsadmin -l /cache-directory cfsadmin: list cache FS information maxblocks 90% minblocks 0% threshblocks 85% maxfiles 90% minfiles 0% threshfiles 85% maxfilesize 3MB cache-ID #

4. Delete the CacheFS le system from the specied cache.

# cfsadmin -d cache-ID /cache-directory

cache-ID

Indicates the name of the CacheFS le system, which is the last line of the cfsadmin -l output. For more information, see How to Display Information About a CacheFS File System on page 341. You can delete all the CacheFS le systems in a particular cache by specifying all for cache-ID. Species the directory where the cache resides.

/cache-directory

5. Verify that the CacheFS le system has been deleted. The cache ID of the le system you just deleted should be missing from the cfsadmin -l output.
# cfsadmin -l /cache-directory cfsadmin: list cache FS information maxblocks 90% minblocks 0% threshblocks 85% maxfiles 90% minfiles 0% threshfiles 85% maxfilesize 3MB #

For more information about the elds that are specied in the command output, refer to cfsadmin(1M). 6. Update the resource counts for the cache.
# fsck -F cachefs /cache-directory

For more information, see How to Check the Integrity of a CacheFS File System on page 344.
Example 206

Deleting a CacheFS File System

The following example shows how to delete the le systems from the cache.
Chapter 20 Using The CacheFS File System (Tasks) 343

# # # # #

umount /cfssrc cfsadmin -l /cfssrc cfsadmin -d _dev_dsk_c0t6d0s0:_cfssrc cfsadmin -l fsck -F cachefs /cache-directory

How to Check the Integrity of a CacheFS File System

Use the fsck command to check the integrity of CacheFS le systems. The CacheFS version of the fsck command automatically corrects problems without requiring user interaction. You should not need to run the fsck command manually for CacheFS le systems because the fsck command is run automatically at boot time or when the le system is mounted. If you want to manually check the integrity, you can use the following procedure. For more information, see fsck_cachefs(1M).

Steps

1. Become superuser on the client system. 2. Check the le systems in the specied cache.
# fsck -F cachefs [-m -o noclean] /cache-directory

-m -o noclean /cache-directory
Example 207

Causes the fsck command to check a CacheFS le system without making any repairs. Forces a check on the CacheFS le systems only. Does not make any repairs. Indicates the name of the directory where the cache resides.

Checking the Integrity of CacheFS File Systems

The following example shows how to check the le systems cached in the /local/mycache cache.
# fsck -F cachefs /local/mycache #

344

System Administration Guide: Devices and File Systems June 2005

Packing a Cached File System (Task Map)

The following task map describes the procedures that are associated with packing a CacheFS le system. All of these procedures are optional.

Task

Description

For Instructions

Pack les in the cache.

Identify les and directories How to Pack Files in the to be loaded in the cache and Cache on page 346 pack them. Packing ensures that current copies of these les are available in the cache. Create a packing list if you do How to Create a Packing not want to specify each List on page 348 individual le that you want packed in the cache. Specify the name of the packing list of the les to be packed in the cache. Remove a le from the cache that is no longer needed. How to Pack Files in the Cache With a Packing List on page 349 How to Unpack Files or Packing Lists From the Cache on page 349How to Unpack Files or Packing Lists From the Cache on page 349

Create a packing list.

Pack les in the cache with a packing list. Unpack les or packing lists from the cache.

Display packed les information.

View information about the How to Display Packed Files Information on page 347 les that youve packed, including their packing status.

Packing a CacheFS File System

For general use, the CacheFS software operates automatically after it is set up, without requiring any action from the user. Files are cached on a most recently used basis. With the packing feature, you can take a more active role in managing your cache by ensuring that certain les or directories are always updated in the cache. You can specify les and directories to be loaded in the cache by using the cachefspack command. This command ensures that current copies of these les are available in the cache.
Chapter 20 Using The CacheFS File System (Tasks) 345

The packing list contains the names of specic les and directories. The packing list can also contain other packing lists. This feature saves you from having to specify individual les and directories when you have many items to pack in your cache. You can print out a brief help summary of all the cachefspack options by using the -h option as follows:
$ cachefspack -h Must select 1 and only 1 of the following 5 options -d Display selected filenames -i Display selected filenames packing status -p Pack selected filenames -u Unpack selected filenames -U Unpack all files in directory dir -f Specify input file containing rules -h Print usage information -r Interpret strings in LIST rules as regular expressions -s Strip ./ from the beginning of a pattern name -v Verbose option files - a list of filenames to be packed/unpacked

M
Step

How to Pack Files in the Cache

G Pack les in the cache. $ cachefspack -p lename

-p lename

Species that you want the le or les to be packed. This option is also the default. Species the name of the le or directory you want packed in the cache. When you specify a directory, all of its subdirectories are also packed. For more information, see cachefspack(1M).

Example 208

ExamplesPacking Files in the Cache

The following example shows the projects le being packed in the cache.
$ cachefspack -p projects

The following example shows three les being packed in the cache.
$ cachefspack -p projects updates master_plan

The following example shows a directory being packed in the cache.

$ cachefspack -p /data/abc/bin

346

System Administration Guide: Devices and File Systems June 2005

M
Step

How to Display Packed Files Information

G Display packed les information. $ cachefspack -i[v] cached-lename-or-directory

-i -v cached-lename-or-directory

Species that you want to view information about your packed les. Is the verbose option. Species the name of the le or directory for which to display information.

Example 209

Displaying Packed Files Information

The following example shows that the doc_file le has been successfully packed.
$ cachefspack -i doc_file cachefspack: file doc_file marked packed YES, packed YES

In the following example, the /data/abc directory contains the bin subdirectory. The bin subdirectory has three les: big, medium, and small. Although the big and small les are specied to be packed, they are not. The medium le is successfully packed.
$ cd /data/abc $ cachefspack -i bin . . . cachefspack: file /bin/big marked packed YES, packed NO cachefspack: file /bin/medium marked packed YES, packed YES cachefspack: file /bin/small marked packed YES, packed NO . . .

If you use the -iv options together, you get additional information as to whether the le or directory specied has been ushed from the cache. For example:
$ cd /data/bin FSCACHEPACK-4$ cachefspack -iv bin . . . cachefspack: file /bin/big marked packed YES, packed NO, nocache YES cachefspack: file /bin/medium marked packed YES, packed YES, nocache NO

Chapter 20 Using The CacheFS File System (Tasks)

347

cachefspack: file /bin/small marked packed YES, packed NO nocache NO . . .

The last line of this example shows that the directory contents have not been ushed from the cache.

Using Packing Lists

One feature of the cachefspack command is the ability to create packing lists. A packing list contains les or directories to be packed in the cache. If a directory is in the packing list, all of its subdirectories and les will also be packed. This feature saves the time of having to specify each individual le that you want packed in the cache.

M
Step

How to Create a Packing List

G Create a packing list le by using vi.

The packing list le format uses the same format as the filesync command. For more information, see filesync(1). Two packing list features are the following:
I

You can identify les in the packing list as regular expressions rather than literal le names so that you dont have to specify each individual le name. You can pack les from a shared directory by ensuring that you pack only those les that you own.

For more information on using these features, see cachefspack(1M).

Example 2010

Creating a Packing List

The following example shows the contents of a packing list le.
BASE /home/ignatz LIST plans LIST docs IGNORE *.ps
I

The path identied with the BASE statement is the directory where you have items you want to pack.

348

System Administration Guide: Devices and File Systems June 2005

I I

The two LIST statements identify specic les within that directory to pack. The IGNORE statement identies the le type of .ps, which you do not want to pack.

M
Step

How to Pack Files in the Cache With a Packing List

G Pack les in the packing list. $ cachefspack -f packing-list

-f packing-list
Example 2011

Species that you want to use a packing list. Species the name of the packing list.

Packing Files in the Cache With a Packing List

This example uses the list.pkg le as the packing list for the cachefspack command.
$ cachefspack -f list.pkg

Unpacking Files or Packing Lists From the Cache

You might need to remove, or unpack, a le from the cache. Perhaps you have some les or directories that have a higher priority than others, so you need to unpack the less critical les. For example, you nished up a project and have archived the les that are associated with that project. You are now working on a new project, and therefore, a new set of les.

How to Unpack Files or Packing Lists From the Cache

G Unpack les or packing lists from the cache. $ cachefspack -u lename | -U cache-directory

Step

-u lename

Species that you want the le or les unpacked. You must specify a le name with this option. Species the name of the le or packing list that you want unpacked in the cache.

Chapter 20 Using The CacheFS File System (Tasks)

349

-U

Species that you want to unpack all les in the cache.

For more information, see cachefspack(1M).

Example 2012

Unpacking Files or Packing Lists From the Cache

The following example shows the le /data/abc/bin/big being unpacked from the cache.
$ cachefspack -u /data/abc/bin/big

The following example shows three les being unpacked from the cache.
$ cd /data/abc/bin/big $ cachefspack -u big small medium

The following example shows how to unpack a packing list. A packing list is a le that contains the path to a directory of les:
$ cachefspack -uf list.pkg

The following example uses the -U option to specify that all les in a cache directory being unpacked.
$ cachefspack -U /local/mycache

You cannot unpack a cache that does not have at least one le system mounted. With the -U option, if you specify a cache that does not contain mounted le systems, output similar to the following is displayed:
$ cachefspack -U /local/mycache cachefspack: Could not unpack cache /local/mycache, no mounted filesystems in the cache.

Troubleshooting cachefspack Errors

You might see the following error messages when you use the cachefspack command.
cachefspack: pathname - cant open directory: permission denied

Cause You might not have the correct permissions to access the le or directory. Action Set the correct permissions.
cachefspack: pathname - cant open directory: no such file or directory

Cause You might not have specied the correct le or directory.

350 System Administration Guide: Devices and File Systems June 2005

Action Check for a possible typo.

cachefspack: pathname - cant open directory: stale NFS file handle

Cause The le or directory might have been moved or deleted from the server at the time you attempted to access it. Action Verify that the le or directory on the server is still accessible.
cachefspack: pathname - cant open directory: interrupted system call

Cause You might have inadvertently pressed Control-C while issuing the command. Action Reissue the command.
cachefspack: pathname - cant open directory: I/O error

Cause You might have a hardware problem. Action Check your hardware connections.
cachefspack: error opening dir

Cause You might not have specied the correct le or directory. The path identied after the BASE command in the le format could be a le and not a directory. The path specied must be a directory. Action Check for a possible typo. Check the path identied after the BASE command in your le format. Ensure that the path identies a directory, not a le.
cachefspack: unable to get shared objects

Cause The executable might be corrupt or in a format that is not recognizable. Action Replace the executable.
cachefspack: lename - cant pack file: permission denied

Cause You might not have the correct permissions to access the le or directory. Action Set the correct permissions.
cachefspack: lename - cant pack file: no such file or directory Chapter 20 Using The CacheFS File System (Tasks) 351

Cause You might not have specied the correct le or directory. Action Check for a possible typo.
cachefspack: lename- cant pack file: stale NFS file handle

Cause The le or directory might have been moved or deleted from the server at the time you attempted to access it. Action Verify that the le or directory on the server is still accessible.
cachefspack: lename- cant pack file: interrupted system call

Cause You might have inadvertently pressed Control-C while issuing the command. Action Reissue the command.
cachefspack: lename- cant pack file: I/O error

Cause You might have a hardware problem. Action Check your hardware connections.
cachefspack: lename- cant pack file: no space left on device.

Cause The cache is out of disk space. Action You need to increase the size of the cache by increasing disk space.
cachefspack: lename - cant unpack file: permission denied

Cause You might not have the correct permissions to access the le or directory. Action Set the correct permissions.
cachefspack: lename - cant unpack file: no such file or directory

Cause You might not have specied the correct le or directory. Action Check for a possible typo.
cachefspack: lename- cant unpack file: stale NFS file handle

352

System Administration Guide: Devices and File Systems June 2005

Cause The le or directory might have been moved or deleted from the server at the time you attempted to access it. Action Verify that the le or directory on the server is still accessible.
cachefspack: lename- cant unpack file: interrupted system call

Cause You might have inadvertently pressed Control-C while issuing the command. Action Reissue the command.
cachefspack: lename- cant unpack file I/O error

Cause You might have a hardware problem. Action Check your hardware connections.
cachefspack: only one d, i, p, or u option allowed

Cause You specied more than one of these options in a command session. Action Select one option for the command session.
cachefspack: cant find environment variable.

Cause You forgot to set a corresponding environment variable to match the $ in your conguration le. Action Dene the environment variable in the proper location.
cachefspack: skipping LIST command - no active base

Cause A LIST command is present in your conguration le but has no corresponding BASE command. Action Dene the BASE command.

Chapter 20 Using The CacheFS File System (Tasks)

353

Collecting CacheFS Statistics (Task Map)

The following task map shows the steps involved in collecting CacheFS statistics. All these procedures are optional.

Task

Description

For Instructions

Set up logging.

Set up logging on a CacheFS le system by using the cachefslog command. Locate the log le by using the cachefslog command. Stop logging by using the cachefslog command. View the cache size by using the cachefswssize command. View the statistics by using the cachefsstat command.

How to Set Up CacheFS Logging on page 355 How to Locate the CacheFS Log File on page 356 How to Stop CacheFS Logging on page 357 How to View the Working Set (Cache) Size on page 357 How to View CacheFS Statistics on page 358

Locate the log le. Stop logging. View the cache size.

View the cache statistics.

Collecting CacheFS Statistics

Collecting CacheFS statistics enables you to do the following:
I I

Determine an appropriate cache size. Observe the performance of the cache.

These statistics help you determine the trade-off between your cache size and the desired performance of the cache. The following table describes the CacheFS statistics commands.

354

System Administration Guide: Devices and File Systems June 2005

Command

Description

Man Page

cachefslog

Species the location of the log le. This command also displays where the statistics are currently being logged, and enables you to stop logging. Interprets the log le to give a recommended cache size.

cachefslog(1M)

cachefswssize

cachefswssize(1M)

cachefsstat

Displays statistical information about cachefsstat(1M) a specic CacheFS le system or all CacheFS le systems. The information provided in the command output is taken directly from the cache.

Note You can issue the CacheFS statistics commands from any directory. You must be superuser to issue the cachefswssize command.

The CacheFS statistics begin accumulating when you create the log le. When the work session is over, stop the logging by using the cachefslog -h command, as described in How to Stop CacheFS Logging on page 357. Before using the CacheFS statistics commands, you must do the following:
I I

Set up your cache by using the cfsadmin command. Decide on an appropriate length of time to allow statistical information to collect in the log le you create. The length of time should equal a typical work session. For example, a day, a week, or a month. Select a location or path for the log le. Ensure that sufficient space to allows for the growth of the log le. The longer you intend to allow statistical information to collect in the log le, the more space you need.

Note The following procedures are presented in a recommended order. This order is not required.

M
Steps

How to Set Up CacheFS Logging

1. Set up logging.
$ cachefslog -f log-le-path /mount-point

-f

Sets up logging.
Chapter 20 Using The CacheFS File System (Tasks) 355

log-le-path /mount-point

Species the location of the log le. The log le is a standard le you create with an editor, such as vi. Designates the mount point (CacheFS le system) for which statistics are being collected.

2. Verify that you correctly set up the log le.

$ cachefslog /mount-point

Example 2013

Setting Up CacheFS Logging

The following example shows how to set up the /var/tmp/samlog log le to collect statistics about the /home/sam directory.
$ cachefslog -f /var/tmp/samlog /home/sam /var/tmp/samlog: /home/sam

M
Step

How to Locate the CacheFS Log File

G Display where CacheFS statistics are being logged. $ cachefslog /mount-point

where /mount-point species the CacheFS le system for which you want to view the statistics. You can also use the cachefslog command with no options to locate a log le for a particular mount point.
Example 2014

Locating the CacheFS Log File

The following example shows what you would see if a log le has been set up. The location of the log le is /var/tmp/stufflog.
$ cachefslog /home/stuff /var/tmp/stufflog: /home/stuff

The following example shows that no log le has been set up for the specied le system.
$ cachefslog /home/zap not logged: /home/zap

356

System Administration Guide: Devices and File Systems June 2005

How to Stop CacheFS Logging

Use the cachefslog -h option to stop logging.
$ cachefslog -h /mount-point

The following example shows how to stop logging on /home/stuff.

$ cachefslog -h /home/stuff not logged: /home/stuff

If you get a system response other than the response specied here, you did not successfully stop logging. Determine if you are using the correct log le name and mount point.

How to View the Working Set (Cache) Size

You might want to check if you need to increase the size of the cache. Or, you might want to determine the ideal cache size based on your activity since you last used the cachefslog command for a particular mount point.

Steps

1. Become superuser on the client system. 2. View the current cache size and highest logged cache size.
# cachefswssize log-le-path

For more information, see cachefswssize(1M).

Example 2015

Viewing the Working Set (Cache) Size

In the following example, the end size is the size of the cache at the time you issued the cachefswssize command. The high water size is the largest size of the cache during the timeframe in which logging occurred.
# cachefswssize /var/tmp/samlog /home/sam end size: high water size: / end size: high water size: /opt end size: high water size: 128k 128k 1736k 1736k

10688k 10704k

Chapter 20 Using The CacheFS File System (Tasks)

357

/nfs/saturn.dist end size: high water size: /data/abc end size: high water size: /nfs/venus.svr4 end size: high water size: /data end size: high water size:

1472k 1472k

7168k 7168k

4688k 5000k

4992k 4992k

total for cache initial size: 110960k end size: 30872k high water size: 30872k

Viewing CacheFS Statistics

The following table explains the terminology that is displayed in the statistics output for CacheFS le systems.
TABLE 202

CacheFS Statistics Terminology

Description

Output Term

cache hit rate

The rate of cache hits versus cache misses, followed by the actual number of hits and misses. A cache hit occurs when the user wants to perform an operation on a le or les, and the le or les are actually in the cache. A cache miss occurs when the le is not in the cache. The load on the server is the sum of cache misses, consistency checks, and modications (modies). The number of consistency checks performed, followed by the number that passed, and the number that failed. The number of modify operations. For example, writes or creates.

consistency checks modifies

How to View CacheFS Statistics

View the statistics with the cachefsstat command. You can view the statistics at any time. For example, you do not have to set up logging in order to view the statistics.

358

System Administration Guide: Devices and File Systems June 2005

Step

G View CacheFS statistics. $ cachefsstat /mount-point

where /mount-point species the CacheFS le system for which you want to view the statistics. If you do not specify the mount point, statistics for all mounted CacheFS le systems will be displayed.
Example 2016

Viewing CacheFS Statistics

This example shows how to view statistics on the cached le system, /home/sam.
$ cachefsstat /home/sam cache hit rate: 73% (1234 hits, 450 misses) consistency checks: 700 (650 pass, 50 fail) modifies: 321 garbage collection: 0

Chapter 20 Using The CacheFS File System (Tasks)

359

360

System Administration Guide: Devices and File Systems June 2005

CHAPTER

21

Conguring Additional Swap Space (Tasks)

This chapter provides guidelines and step-by-step instructions for conguring additional swap space after the Solaris OS is installed. This is a list of the step-by-step instructions in this chapter.
I I

How to Create a Swap File and Make It Available on page 368 How to Remove Unneeded Swap Space on page 369

This is a list of the overview information in this chapter.

I I I I I I

About Swap Space on page 361 How Do I Know If I Need More Swap Space? on page 363 How Swap Space Is Allocated on page 364 Planning for Swap Space on page 365 Monitoring Swap Resources on page 366 Adding More Swap Space on page 367

About Swap Space

You should understand the features of the SunOS swap mechanism to determine the following:
I I I

Swap space requirements The relationship between swap space and the TMPFS le system How to recover from error messages related to swap space

361

Swap Space and Virtual Memory

Solaris software uses some disk slices for temporary storage rather than for le systems. These slices are called swap slices. Swap slices are used as virtual memory storage areas when the system does not have enough physical memory to handle current processes. The virtual memory system maps physical copies of les on disk to virtual addresses in memory. Physical memory pages that contain the data for these mappings can be backed by regular les in the le system, or by swap space. If the memory is backed by swap space it is referred to as anonymous memory because no identity is assigned to the disk space that is backing the memory. The Solaris OS uses the concept of virtual swap space, a layer between anonymous memory pages and the physical storage (or disk-backed swap space) that actually back these pages. A systems virtual swap space is equal to the sum of all its physical (disk-backed) swap space plus a portion of the currently available physical memory. Virtual swap space has these advantages:
I

The need for large amounts of physical swap space is reduced because virtual swap space does not necessarily correspond to physical (disk) storage. A pseudo le system called SWAPFS provides addresses for anonymous memory pages. Because SWAPFS controls the allocation of memory pages, it has greater exibility in deciding what happens to a page. For example, SWAPFS might change the pages requirements for disk-backed swap storage.

Swap Space and the TMPFS File System

The TMPFS le system is activated automatically in the Solaris environment by an entry in the /etc/vfstab le. The TMPFS le system stores les and their associated information in memory (in the /tmp directory) rather than on disk, which speeds access to those les. This feature results in a major performance enhancement for applications such as compilers and DBMS products that use /tmp heavily. The TMPFS le system allocates space in the /tmp directory from the systems swap resources. This feature means that as you use up space in the /tmp directory, you are also using up swap space. So, if your applications use the /tmp directory heavily and you do not monitor swap space usage, your system could run out of swap space. Do use the following if you want to use TMPFS, but your swap resources are limited:
I

Mount the TMPFS le system with the size option (-o size) to control how much swap resources TMPFS can use. Use your compilers TMPDIR environment variable to point to another larger directory. Using your compilers TMPDIR variable only controls whether the compiler is using the /tmp directory. This variable has no effect on other programs use of the /tmp directory.

362

System Administration Guide: Devices and File Systems June 2005

Swap Space as a Dump Device

A dump device is usually disk space that is reserved to store system crash dump information. By default, a systems dump device is congured to be a swap slice. If possible, you should congure an alternate disk partition as a dedicated dump device instead to provide increased reliability for crash dumps and faster reboot time after a system failure. You can congure a dedicated dump device by using the dumpadm command. For more information, see Chapter 24, Managing System Crash Information (Tasks), in System Administration Guide: Advanced Administration. If you are using a volume manager to manage your disks, such as Solaris Volume Manager, do not congure your dedicated dump device to be under its control. You can keep your swap areas under Solaris Volume Managers control, which is a recommended practice. However, for accessibility and performance reasons, congure another disk as a dedicated dump device outside of Solaris Volume Managers control.

Swap Space and Dynamic Reconguration

A good practice is to allocate enough swap space to support a failing CPU or system board during dynamic reconguration. Otherwise, a CPU or system board failure might result in your host or domain rebooting with less memory. Without having this additional swap space available, one or more of your applications might fail to start due to insufficient memory. This problem would require manual intervention either to add additional swap space or to recongure the memory usage of these applications. If you have allocated additional swap space to handle a potential loss of memory on reboot, all of your intensive applications might start as usual. This means the system will be available to the users, perhaps possibly slower due to some additional swapping. For more information, see your hardware dynamic reconguration guide.

How Do I Know If I Need More Swap Space?

Use the swap -l command to determine if your system needs more swap space. For example, the following swap -l output shows that this systems swap space is almost entirely consumed or at 100% allocation.
% swap -l swapfile dev swaplo blocks free 363

Chapter 21 Conguring Additional Swap Space (Tasks)

/dev/dsk/c0t0d0s1

136,1

16 1638608

88

When a systems swap space is at 100% allocation, an applications memory pages become temporarily locked. Application errors might not occur, but system performance will likely suffer. For information on adding more swap space to your system, see How to Create a Swap File and Make It Available on page 368.

Swap-Related Error Messages

These messages indicate that an application was trying to get more anonymous memory. However, no swap space was left to back it.
application is out of memory malloc error O messages.1:Sep 21 20:52:11 mars genunix: [ID 470503 kern.warning] WARNING: Sorry, no swap space to grow stack for pid 100295 (myprog)

TMPFS-Related Error Messages

The following message is displayed if a page could not be allocated when a le was being written. This problem can occur when TMPFS tries to write more than it is allowed or if currently executed programs are using a lot of memory.
directory: File system full, swap space limit exceeded

The following message means that TMPFS ran out of physical memory while attempting to create a new le or directory:
directory: File system full, memory allocation failed

For information on recovering from the TMPFS-related error messages, see tmpfs(7FS).

How Swap Space Is Allocated

Initially, swap space is allocated as part of the Solaris installation process. If you use the installation programs automatic layout of disk slices and do not manually change the size of the swap slice, the Solaris installation program allocates a default swap area of 512 Mbytes.
364 System Administration Guide: Devices and File Systems June 2005

Starting in the Solaris 9 release, the installation program allocates swap space starting at the rst available disk cylinder (typically cylinder 0). This placement provides maximum space for the root (/) le system during the default disk layout and enables the growth of the root (/) le system during an upgrade. For general guidelines on allocating swap space, see Planning for Swap Space on page 365. You can allocate additional swap space to the system by creating a swap le. For information about creating a swap le, see Adding More Swap Space on page 367.

Swap Areas and the /etc/vfstab File

After the system is installed, swap slices and swap les are listed in the /etc/vfstab le. They are activated by the /sbin/swapadd script when the system is booted. An entry for a swap device in the /etc/vfstab le contains the following:
I I

The full path name of the swap slice or swap le File system type of the swap slice or swap le

The le system that contains a swap le must be mounted before the swap le is activated. So, in the /etc/vfstab le, ensure that the entry that mounts the le system comes before the entry that activates the swap le.

Planning for Swap Space

The most important factors in determining swap space size are the requirements of the systems software applications. For example, large applications such as computer-aided design simulators, database management products, transaction monitors, and geologic analysis systems can consume as much as 2001000 Mbytes of swap space. Consult your application vendors for swap space requirements for their applications. If you are unable to determine swap space requirements from your application vendors, use the following general guidelines based on your system type to allocate swap space.

Chapter 21 Conguring Additional Swap Space (Tasks)

365

System Type

Swap Space Size

Dedicated Dump Device Size

Workstation with about 4 Gbytes of physical memory

1 Gbyte

1 Gbyte 2 Gbytes 4 Gbytes

Mid-range server with about 8 2 Gbytes Gbytes of physical memory High-end server with about 16 4 Gbytes to 128 Gbytes of physical memory

In addition to these general guidelines, consider allocating swap space or disk space for the following:
I I

A dedicated dump device. Determine whether large applications (such as compilers) will be using the /tmp directory. Then, allocate additional swap space to be used by TMPFS. For information about TMPFS, see Swap Space and the TMPFS File System on page 362.

Monitoring Swap Resources

The /usr/sbin/swap command is used to manage swap areas. Two options, -l and -s, display information about swap resources. Use the swap -l command to identify a systems swap areas. Activated swap devices or les are listed under the swapfile column.
# swap -l swapfile /dev/dsk/c0t0d0s1 dev swaplo blocks free 136,1 16 1638608 1600528

Use the swap -s command to monitor swap resources.

# swap -s total: 57416k bytes allocated + 10480k reserved = 67896k used, 833128k available

The used value plus the available value equals the total swap space on the system, which includes a portion of physical memory and swap devices (or les). You can use the amount of available and used swap space (in the swap -s output) as a way to monitor swap space usage over time. If a systems performance is good, use swap -s to determine how much swap space is available. When the performance of a system slows down, check the amount of available swap space to determine if it has decreased. Then you can identify what changes to the system might have caused swap space usage to increase.
366 System Administration Guide: Devices and File Systems June 2005

When using this command, keep in mind that the amount of physical memory available for swap usage changes dynamically as the kernel and user processes lock down and release physical memory.
Note The swap -l command displays swap space in 512-byte blocks. The swap -s command displays swap space in 1024-byte blocks. If you add up the blocks from swap -l and convert them to Kbytes, the result is less than used + available (in the swap -s output). The reason is that swap -l does not include physical memory in its calculation of swap space.

The output from the swap -s command is summarized in the following table.
TABLE 211 Keyword

Output of the swap -s Command

Description

bytes allocated

The total amount of swap space in 1024-byte blocks that is currently allocated as backing store (disk-backed swap space). The total amount of swap space in 1024-byte blocks that is not currently allocated, but claimed by memory for possible future use. The total amount of swap space in 1024-byte blocks that is either allocated or reserved. The total amount of swap space in 1024-byte blocks that is currently available for future reservation and allocation.

reserved

used

available

Adding More Swap Space

As system congurations change and new software packages are installed, you might need to add more swap space. The easiest way to add more swap space is to use the mkfile and swap commands to designate a part of an existing UFS or NFS le system as a supplementary swap area. These commands, described in the following sections, enable you to add more swap space without repartitioning a disk. Alternative ways to add more swap space are to repartition an existing disk or to add another disk. For information on how to repartition a disk, see Chapter 11.

Chapter 21 Conguring Additional Swap Space (Tasks)

367

Creating a Swap File

The following general steps are involved in creating a swap le:
I I I

Creating a swap le by using the mkfile command. Activating the swap le by using the swap command. Adding an entry for the swap le in the /etc/vfstab le so that the swap le is activated automatically when the system is booted.

mkfile Command
The mkfile command creates a le that is suitable for use as either an NFS-mounted swap area or a local swap area. The sticky bit is set, and the le is lled with zeros. You can specify the size of the swap le in bytes (the default) or in Kbytes, blocks, or Mbytes by using the k, b, or m suffixes, respectively. The following table shows the mkfile command options.
TABLE 212 Option

Options to the mkfile Command

Description

-n

Creates an empty le. The size is noted. However, the disk blocks are not allocated until data is written to them. Reports the names and sizes of created les.

-v

Note Use the -n option only when you create an NFS swap le.

M
Steps

How to Create a Swap File and Make It Available

1. Become superuser. You can create a swap le without root permissions. However, to avoid accidental overwriting, root should be the owner of the swap le. 2. Create a directory for the swap le, if needed. 3. Create the swap le.
# mkfile nnn[k|b|m] lename

The swap le of the size nnn (in Kbytes, bytes, or Mbytes) with the lename you specify is created. 4. Activate the swap le.
# /usr/sbin/swap -a /path/lename 368 System Administration Guide: Devices and File Systems June 2005

You must use the absolute path name to specify the swap le. The swap le is added and available until the le system is unmounted, the system is rebooted, or the swap le is removed. Keep in mind that you cannot unmount a le system while some process or program is swapping to the swap le. 5. Add an entry for the swap le to the /etc/vfstab le that species the full path name of the le, and designates swap as the le system type.
/path/lename swap no -

6. Verify that the swap le is added.

$ /usr/sbin/swap -l

Example 211

Creating a Swap File and Making It Available

The following examples shows how to create a 100-Mbyte swap le called /files/swapfile.
# mkdir /files # mkfile 100m /files/swapfile # swap -a /files/swapfile # vi /etc/vfstab (An entry is added for the swap le): /files/swapfile swap # swap -l swapfile dev swaplo blocks /dev/dsk/c0t0d0s1 136,1 16 1638608 /files/swapfile 16 204784

no

free 1600528 204784

Removing a Swap File From Use

If you have unneeded swap space, you can remove it.

M
Steps

How to Remove Unneeded Swap Space

1. Become superuser. 2. Remove the swap space.
# /usr/sbin/swap -d /path/lename

The swap le name is removed so that it is no longer available for swapping. The le itself is not deleted. 3. Edit the /etc/vfstab le and delete the entry for the swap le.
Chapter 21 Conguring Additional Swap Space (Tasks) 369

4. Recover the disk space so that you can use it for something else.
# rm /path/lename

If the swap space is a le, remove it. Or, if the swap space is on a separate slice and you are sure you will not need it again, make a new le system and mount the le system. For information on mounting a le system, see Chapter 19. 5. Verify that the swap le is no longer available.
# swap -l

Example 212

Removing Unneeded Swap Space

The following examples shows how to delete the /files/swapfile swap le.
# swap -d /files/swapfile # (Remove the swap entry from the /etc/vfstab le) # rm /files/swapfile # swap -l swapfile dev swaplo blocks free /dev/dsk/c0t0d0s1 136,1 16 1638608 1600528

370

System Administration Guide: Devices and File Systems June 2005

CHAPTER

22

Checking UFS File System Consistency (Tasks)

This chapter provides overview information and step-by-step instructions about checking UFS le system consistency. This is a list of step-by-step instructions in this chapter.
I

I I I

How to Check the root (/) or /usr File Systems From an Alternate Boot Device on page 381 How to Check Non-root (/) or Non-/usr File Systems on page 383 How to Preen a UFS File System on page 385 How to Restore a Bad Superblock on page 386

This is a list of the overview information in this chapter.

I I I I I I

File System Consistency on page 372 How the File System State Is Recorded on page 372 What the fsck Command Checks and Tries to Repair on page 373 Interactively Checking and Repairing a UFS File System on page 380 Restoring a Bad Superblock on page 386 Syntax and Options for the fsck Command on page 388

For information about fsck error messages, see Chapter 28, Resolving UFS File System Inconsistencies (Tasks), in System Administration Guide: Advanced Administration. For background information on the UFS le system structures referred to in this chapter, see Chapter 23.

371

File System Consistency

The UFS le system relies on an internal set of tables to keep track of inodes used and available blocks. When these internal tables are not properly synchronized with data on a disk, inconsistencies result and le systems need to be repaired. File systems can be inconsistent because of abrupt termination of the operating system from the following:
I I I I

Power failure Accidental unplugging of the system Turning off the system without proper shutdown procedure A software error in the kernel

File system inconsistencies, while serious, are not common. When a system is booted, a check for le system consistency is automatically performed (with the fsck command). Often, this le system check repairs problems it encounters. The fsck command places les and directories that are allocated but unreferenced in the lost+found directory. An inode number is assigned as the name of unreferenced le and directory. If the lost+found directory does not exist, the fsck command creates it. If there is not enough space in the lost+found directory, the fsck command increases its size. For a description of inodes, see Inodes on page 390.

How the File System State Is Recorded

The fsck command uses a state ag, which is stored in the superblock, to record the condition of the le system. This ag is used by the fsck command to determine whether a le system needs to be checked for consistency. The ag is used by the /sbin/rcS script during booting and by the fsck -m command. If you ignore the result from the fsck -m command, all le systems can be checked regardless of the setting of the state ag. For a description of the superblock, see Superblock on page 390. The possible state ag values are described in the following table.

372

System Administration Guide: Devices and File Systems June 2005

TABLE 221

Values of File System State Flags

Description

State Flag Value

FSACTIVE

Indicates a mounted le system that has modied data in memory. A mounted le system with this state ag indicates that user data or metadata would be lost if power to the system is interrupted. Indicates that the le system contains inconsistent le system data. Indicates an undamaged, cleanly unmounted le system. Indicates that the le system has logging enabled. A le system with this ag set is either mounted or unmounted. If a le system has logging enabled, the only ags that it can have are FSLOG or FSBAD. A le system that has logging disable can have FSACTIVE, FSSTABLE, or FSCLEAN. Indicates an idle mounted le system. A mounted le system with this state ag indicates that neither user data nor metadata would be lost if power to the system is interrupted.

FSBAD FSCLEAN FSLOG

FSSTABLE

What the fsck Command Checks and Tries to Repair

This section describes what happens in the normal operation of a le system, what can go wrong, what problems the fsck command (the checking and repair utility) looks for, and how this command corrects the inconsistencies it nds.

Why UFS File System Inconsistencies Might Occur

Every working day, hundreds of les might be created, modied, and removed. Each time a le is modied, the operating system performs a series of le system updates. These updates, when written to the disk reliably, yield a consistent le system. When a user program does an operation to change the le system, such as a write, the data to be written is rst copied into an in-core buffer in the kernel. Normally, the disk update is handled asynchronously. The user process is allowed to proceed even though the data write might not happen until long after the write system call has returned. Thus, at any given time, the le system, as it resides on the disk, lags behind the state of the le system that is represented by the in-core information. The disk information is updated to reect the in-core information when the buffer is required for another use or when the kernel automatically runs the fsflush daemon (at 30-second intervals). If the system is halted without writing out the in-core information, the le system on the disk might be in an inconsistent state.
Chapter 22 Checking UFS File System Consistency (Tasks) 373

A le system can develop inconsistencies in several ways. The most common causes are operator error and hardware failures. Problems might result from an unclean shutdown, if a system is shut down improperly, or when a mounted le system is taken offline improperly. To prevent unclean shutdowns, the current state of the le systems must be written to disk (that is, synchronized) before you shut down the system, physically take a disk pack out of a drive, or take a disk offline. Inconsistencies can also result from defective hardware or problems with the disk or controller rmware. Blocks can become damaged on a disk drive at any time. Or, a disk controller can stop functioning correctly.

UFS Components That Are Checked for Consistency

This section describes the kinds of consistency checks that the fsck command applies to these UFS le system components: superblock, cylinder group blocks, inodes, indirect blocks, and data blocks. For information about UFS le system structures, see Structure of Cylinder Groups for UFS File Systems on page 389.

Superblock Checks
The superblock stores summary information, which is the most commonly corrupted component in a UFS le system. Each change to le system inodes or data blocks also modies the superblock. If the CPU is halted and the last command is not a sync command, the superblock almost certainly becomes corrupted. The superblock is checked for inconsistencies in the following:
I I I I

File system size Number of inodes Free block count Free inode count

File System Size and Inode List Size Checks

The le system size must be larger than the number of blocks used by the superblock and the list of inodes. The number of inodes must be less than the maximum number allowed for the le system. An inode represents all the information about a le. The le system size and layout information are the most critical pieces of information for the fsck command. There is no way to actually check these sizes because they are statically determined when the le system is created. However, the fsck command
374 System Administration Guide: Devices and File Systems June 2005

can check that the sizes are within reasonable bounds. All other le system checks require that these sizes be correct. If the fsck command detects corruption in the static parameters of the primary superblock, it requests the operator to specify the location of an alternate superblock. For more information about the structure of the UFS le system, see Structure of Cylinder Groups for UFS File Systems on page 389.

Free Block Checks

Free blocks are stored in the cylinder group block maps. The fsck command checks that all the blocks marked as free are not claimed by any les. When all the blocks have been accounted for, the fsck command checks if the number of free blocks plus the number of blocks that are claimed by the inodes equal the total number of blocks in the le system. If anything is wrong with the block maps, the fsck command rebuilds them, leaving out blocks already allocated. The summary information in the superblock includes a count of the total number of free blocks within the le system. The fsck command compares this count to the number of free blocks it nds within the le system. If the counts do not agree, the fsck command replaces the count in the superblock with the actual free block count.

Free Inode Checks

Summary information in the superblock contains a count of the free inodes within the le system. The fsck command compares this count to the number of free inodes it nds within the le system. If the counts do not agree, fsck replaces the count in the superblock with the actual free inode count.

Inodes
The list of inodes is checked sequentially starting with inode 2. (Inode 0 and inode 1 are reserved). Each inode is checked for inconsistencies in the following:
I I I I I

Format and type Link count Duplicate block Bad block numbers Inode size

Format and Type of Inodes

Each inode contains a mode word, which describes the type and state of the inode. Inodes might be one of nine types:
I

Regular
Chapter 22 Checking UFS File System Consistency (Tasks) 375

I I I I I I I I

Directory Block special Character special FIFO (named pipe) Symbolic link Shadow (used for ACLs) Attribute directory Socket

Inodes might be in one of three states:

I I I

Allocated Unallocated Partially allocated

When the le system is created, a xed number of inodes are set aside. However, these inodes are not allocated until they are needed. An allocated inode is one that points to a le. An unallocated inode does not point to a le and, therefore, should be empty. The partially allocated state means that the inode is incorrectly formatted. An inode can get into this state if, for example, bad data is written into the inode list because of a hardware failure. The only corrective action the fsck command can take is to clear the inode.

Link Count Checks

Each inode contains a count of the number of directory entries linked to it. The fsck command veries the link count of each inode by examining the entire directory structure, starting from the root (/) directory, and calculating an actual link count for each inode. Discrepancies between the link count stored in the inode and the actual link count as determined by the fsck command might be one of three types:
I

The stored count is not 0, and the actual count is 0. This condition can occur if no directory entry exists for the inode. In this case, the fsck command puts the disconnected le in the lost+found directory.

The stored count is not 0 and the actual count is not 0. However, the counts are unequal. This condition can occur if a directory entry has been added or removed, but the inode has not been updated. In this case, the fsck command replaces the stored link count with the actual link count.

The stored count is 0, and the actual count is not 0. In this case, the fsck command changes the link count of the inode to the actual count.

376

System Administration Guide: Devices and File Systems June 2005

Duplicate Block Checks

Each inode contains a list, or pointers to lists (indirect blocks), of all the blocks claimed by the inode. Because indirect blocks are owned by an inode, inconsistencies in indirect blocks directly affect the inode that owns the indirect block. The fsck command compares each block number claimed by an inode to a list of allocated blocks. If another inode already claims a block number, the block number is put on a list of duplicate blocks. Otherwise, the list of allocated blocks is updated to include the block number. If duplicate blocks are found, the fsck command makes a second pass of the inode list to nd the other inode that claims each duplicate block. The fsck command cannot determine with certainty which inode is in error. So, the fsck command prompts you to choose which inode should be kept and which inode should be cleared. Note that a large number of duplicate blocks in an inode might be caused by an indirect block not being written to the le system

Bad Block Number Checks

The fsck command checks each block number claimed by an inode to determine whether its value is higher than the value of the rst data block and lower than that of the last data block in the le system. If the block number is outside this range, it is considered a bad block number. Bad block numbers in an inode might be caused by an indirect block not being written to the le system. The fsck command prompts you to clear the inode.

Inode Size Checks

Each inode contains a count of the number of data blocks that it references. The number of actual data blocks is the sum of the allocated data blocks and the indirect blocks. The fsck command computes the number of data blocks and compares that block count against the number of blocks that the inode claims. If an inode contains an incorrect count, the fsck command prompts you to x it. Each inode contains a 64-bit size eld. This eld shows the number of characters (data bytes) in the le associated with the inode. A rough check of the consistency of the size eld of an inode uses the number of characters shown in the size eld to calculate how many blocks should be associated with the inode, and then compares that number to the actual number of blocks claimed by the inode.

Indirect Blocks
Indirect blocks are owned by an inode. Therefore, inconsistencies in an indirect block affect the inode that owns it. Inconsistencies that can be checked are the following:
I

Blocks already claimed by another inode

Chapter 22 Checking UFS File System Consistency (Tasks) 377

Block numbers outside the range of the le system

These consistency checks listed are also performed for direct blocks.

Data Blocks
An inode can directly or indirectly reference three kinds of data blocks. All referenced blocks must be of the same kind. The three types of data blocks are the following:
I I I

Plain data blocks Symbolic-link data blocks Directory data blocks

Plain data blocks contain the information stored in a le. Symbolic-link data blocks contain the path name stored in a symbolic link. Directory data blocks contain directory entries. The fsck command can check only the validity of directory data blocks. Directories are distinguished from regular les by an entry in the mode eld of the inode. Data blocks associated with a directory contain the directory entries. Directory data blocks are checked for inconsistencies involving the following:
I I

Directory inode numbers that point to unallocated inodes Directory inode numbers that are greater than the number of inodes in the le system Incorrect directory inode numbers for . and .. directories Directories that are disconnected from the le system

I I

Directory Unallocated Checks

If the inode number in a directory data block points to an unallocated inode, the fsck command removes the directory entry. This condition can occur if the data blocks that contain a new directory entry are modied and written out, but the inode does not get written out. This condition can occur if the CPU is shut down abruptly.

Bad Inode Number Checks

If a directory entry inode number points beyond the end of the inode list, the fsck command removes the directory entry. This condition can occur when bad data is written into a directory data block.

Incorrect . and .. Entry Checks

The directory inode number entry for . must be the rst entry in the directory data block. The directory inode number must reference itself. That is, its value must be equal to the inode number for the directory data block.
378 System Administration Guide: Devices and File Systems June 2005

The directory inode number entry for .. must be the second entry in the directory data block. The directory inode number value must be equal to the inode number of the parent directory or the inode number of itself if the directory is the root (/) directory). If the directory inode numbers for . and .. are incorrect, the fsck command replaces them with the correct values. If there are multiple hard links to a directory, the rst hard link found is considered the real parent to which .. should point. In this case, the fsck command recommends that you have it delete the other names.

Disconnected Directories
The fsck command checks the general connectivity of the le system. If a directory that is not linked to the le system is found, the fsck command links the directory to the lost+found directory of the le system. This condition can occur when inodes are written to the le system. However, the corresponding directory data blocks are not.

Regular Data Blocks

Data blocks associated with a regular le hold the contents of the le. The fsck command does not attempt to check the validity of the contents of a regular les data blocks.

fsck Summary Message

When you run the fsck command interactively and it completes successfully, a message similar to the following is displayed:
# fsck /dev/rdsk/c0t0d0s7 ** /dev/rdsk/c0t0d0s7 ** Last Mounted on /export/home ** Phase 1 - Check Blocks and Sizes ** Phase 2 - Check Pathnames ** Phase 3 - Check Connectivity ** Phase 4 - Check Reference Counts ** Phase 5 - Check Cyl groups 2 files, 9 used, 2833540 free (20 frags, 354190 blocks, 0.0% fragmentation) #

The last line of fsck output describes the following information about the le system: # files # used # free Number of inodes in use Number of fragments in use Number of unused fragments
Chapter 22 Checking UFS File System Consistency (Tasks) 379

# frags # blocks % fragmentation

Number of unused non-block fragments Number of unused full blocks Percentage of fragmentation, where: free fragments x 100 / total fragments in the le system

For information about fragments, see Fragment Size on page 393.

Interactively Checking and Repairing a UFS File System

You might need to interactively check le systems in the following instances:
I I

When they cannot be mounted When they develop inconsistences while in use

When an in-use le system develops inconsistencies, error messages might be displayed in the console window or the system messages le. Or, the system might crash. For example, the system messages le, /var/adm/messages, might include messages similar to the following:
Sep 5 13:42:40 hostname ufs: [ID 879645 kern.notice] NOTICE: /: unexpected free inode 630916, run fsck(1M)

hostname is the system reporting the error. Before using the fsck command, you might want to refer to these references for information on resolving fsck error messages:
I I

Syntax and Options for the fsck Command on page 388 Chapter 28, Resolving UFS File System Inconsistencies (Tasks), in System Administration Guide: Advanced Administration

Keep the following points in mind when running the fsck command to check UFS le systems:
I

A le system should be inactive when you use fsck to check a le system. File system changes waiting to be ushed to disk or le system changes that occur during the fsck checking process can be interpreted as le system corruption. These issues may not be a reliable indication of a problem. A le system must be inactive when you use fsck to repair that le system. File system changes waiting to be ushed to disk or le system changes that occur during the fsck repairing process might cause the le system to become corrupted. Or, they might cause the system to crash.

380

System Administration Guide: Devices and File Systems June 2005

Unmount a le system before you use fsck on that le system. Doing so ensures that the le system data structures are consistent as possible. The only exceptions are for the active root (/) and /usr le systems because they must be mounted to run fsck. If you need to repair the root (/) or /usr le systems, boot the system from an alternate device, if possible, so that these le systems are unmounted and inactive. For step-by-step instructions on running fsck on the root (/) or /usr le system, see How to Check the root (/) or /usr File Systems From an Alternate Boot Device on page 381.

How to Check the root (/) or /usr File Systems From an Alternate Boot Device
This procedure assumes that a local CD or network boot server is available so that you can boot the system from an alternate device. For information on restoring a bad superblock, see How to Restore a Bad Superblock on page 386.

Steps

1. Become superuser or assume an equivalent role. 2. For systems with mirrored root (/) le systems only: Detach the root (/) mirror before booting from the alternate device, or you risk corrupting the le system. For information on detaching the root (/) mirror, see Working With Submirrors in Solaris Volume Manager Administration Guide. 3. Identify the device, such as /dev/dsk/c0t0d0s0, of the root (/) or /usr le system that needs to be checked. Youll need to supply this device name when booted from an alternate device. Identifying this device when you are already booted from the alternate device is more difficult. 4. Boot the system with the root (/) or /usr le system that needs to be checked from an alternate device, such as a local CD or the network, in single-user mode. Doing so ensures that there is no activity on these le systems. For example:
# init 0 ok boot net -s . . . #

Chapter 22 Checking UFS File System Consistency (Tasks)

381

5. Check the device that contains the root (/) or /usr le system as identied in Step 3. If the hardware for the le system to be checked or repaired has changed, the device names might have changed. Check that the fsck -n message Last Mounted on ... indicates the expected device for the le system. In this example, the root (/) le system to be checked is /dev/dsk/c0t0d0s0.
# fsck -n /dev/rdsk/c0t0d0s0 ** /dev/rdsk/c0t0d0s0 (NO WRITE) ** Last Mounted on / . . . fsck /dev/rdsk/c0t0d0s0 ** /dev/rdsk/c0t0d0s0 ** Last Mounted on / ** Phase 1 - Check Blocks and Sizes ** Phase 2 - Check Pathnames . . .

6. Correct any reported fsck errors. For information on how to respond to the error message prompts while you interactively check one or more UFS le systems, see Chapter 28, Resolving UFS File System Inconsistencies (Tasks), in System Administration Guide: Advanced Administration. 7. If necessary, run the fsck command again if you see messages similar to the following: FILE SYSTEM STATE NOT SET TO OKAY or FILE SYSTEM MODIFIED The fsck command might be unable to x all errors in one execution. If fsck cannot repair all of the problems after running it several times, see Fixing a UFS File System That the fsck Command Cannot Repair on page 385. 8. Mount the repaired le system to determine if any les exist in the lost+found directory. Individual les put in the lost+found directory by the fsck command are renamed with their inode numbers. If possible, rename the les and move them where they belong. Try to use the grep command to match phrases within individual les and the file command to identify le types. Eventually, remove unidentiable les or directories left in the lost+found directory so that it doesnt ll it up unnecessarily. 9. Bring the system back to multiuser mode.
# init 6

382

System Administration Guide: Devices and File Systems June 2005

If you press Control-D when you booted in single-user mode from an alternate device, the system will start the Solaris installation process. 10. For systems with mirrored root (/) le systems only: Reattach the root (/) mirror.

How to Check Non-root (/) or Non-/usr File Systems

This procedure assumes that the le system to be checked is unmounted. For information on restoring a bad superblock, see How to Restore a Bad Superblock on page 386.

Steps

1. Become superuser or assume an equivalent role. 2. Unmount the local le system to ensure that there is no activity on the le system. Specify the mount point directory or /dev/dsk/device-name as arguments to the fsck command. Any inconsistency messages are displayed. For example:
# umount /export/home # fsck /dev/rdsk/c0t0d0s7 ** /dev/dsk/c0t0d0s7 ** Last Mounted on /export/home . . .

3. Correct any reported fsck errors. For information on how to respond to the error message prompts while you interactively check one or more UFS le systems, see Chapter 28, Resolving UFS File System Inconsistencies (Tasks), in System Administration Guide: Advanced Administration. 4. If necessary, run the fsck command again if you see the following messages: FILE SYSTEM STATE NOT SET TO OKAY or FILE SYSTEM MODIFIED The fsck command might be unable to x all errors in one execution. If fsck cannot repair all of the problems after running it several times, see Fixing a UFS File System That the fsck Command Cannot Repair on page 385. 5. Mount the repaired le system to determine if there are any les in the lost+found directory. Individual les put in the lost+found directory by the fsck command are renamed with their inode numbers.
Chapter 22 Checking UFS File System Consistency (Tasks) 383

6. Rename and move any les put in the lost+found directory. If possible, rename the les and move them where they belong. Try to use the grep command to match phrases within individual les and the file command to identify le types. Eventually, remove unidentiable les or directories left in the lost+found directory so that it doesnt ll it up unnecessarily.
Example 221

Interactively Checking Non-root (/) or Non-/usr File Systems

The following example shows how to check the /dev/rdsk/c0t0d0s6 le system and correct the incorrect block count. This example assumes that the le system is unmounted.
# fsck /dev/rdsk/c0t0d0s6 ** Phase 1 - Check Block and Sizes INCORRECT BLOCK COUNT I=2529 (6 should be 2) CORRECT? y ** Phase 2 - Check Pathnames ** Phase 3 - Check Connectivity ** Phase 4 - Check Reference Counts ** Phase 5 - Cylinder Groups 929 files, 8928 used, 2851 free (75 frags, 347 blocks, 0.6% fragmentation) /dev/rdsk/c0t0d0s6 FILE SYSTEM STATE SET TO OKAY ***** FILE SYSTEM WAS MODIFIED *****

Preening UFS File Systems

The fsck -o p command (p is for preen) checks UFS le systems and automatically xes the problems that normally result from an unexpected system shutdown. This command exits immediately if it encounters a problem that requires operator intervention. This command also permits parallel checking of le systems. You can run the fsck -o p command to preen the le systems after an unclean shutdown. In this mode, the fsck command does not look at the clean ag and does a full check. These actions are a subset of the actions that the fsck command takes when it runs interactively.

384

System Administration Guide: Devices and File Systems June 2005

How to Preen a UFS File System

This procedure assumes that the le system is unmounted or inactive. 1. Become superuser or assume an equivalent role. 2. Unmount the UFS le system.
# umount /mount-point

Steps

3. Check the UFS le system with the preen option.

# fsck -o p /dev/rdsk/device-name

You can preen individual le systems by using /mount-point or /dev/rdsk/device-name as arguments to the fsck command.
Example 222

Preening a UFS File System

The following example shows how to preen the /export/home le system.
# fsck -o p /export/home

Fixing a UFS File System That the fsck Command Cannot Repair
The fsck command operates in several passes, and a problem corrected in a later pass can expose other problems that are only detected by earlier passes. Therefore, it is sometimes necessary to run fsck repeatedly until it no longer reports any problems. Doing so ensures that all errors have been found and repaired. The fsck command does not keep running until it comes up clean. So, you must rerun the command manually. Pay attention to the information displayed by the fsck command. This information might help you x the problem. For example, the messages might point to a damaged directory. If you delete the directory, you might nd that the fsck command runs cleanly. If the fsck command still cannot repair the le system, try to use the ff, clri, and ncheck commands to gure out and x what is wrong. For information about how to use these commands, see fsdb(1M), ff(1M), clri(1M), and ncheck(1M). Ultimately, you might need to re-create the le system and restore its contents from backup media. For information about restoring complete le systems, see Chapter 27.

Chapter 22 Checking UFS File System Consistency (Tasks)

385

If you cannot fully repair a le system but you can mount it read-only, try using the cp, tar, or cpio commands to retrieve all or part of the data from the le system. If hardware disk errors are causing the problem, you might need to reformat and repartition the disk again before re-creating and restoring le systems. Check that the device cables and connectors are functional before replacing the disk device. Hardware errors usually display the same error again and again across different commands. The format command tries to work around bad blocks on the disk. However, if the disk is too severely damaged, the problems might persist, even after reformatting. For information about using the format command, see format(1M). For information about installing a new disk, see Chapter 13 or Chapter 14.

Restoring a Bad Superblock

When the superblock of a le system becomes damaged, you must restore it. The fsck command tells you when a superblock is bad. Fortunately, copies of the superblock are stored within a le system. You can use the fsck -o b command to replace the superblock with one of these copies. For more information about the superblock, see Superblock on page 390. If the superblock in the root (/) le system becomes damaged and you cannot restore it, you have two choices:
I I

Reinstall the system. Boot from the network or local CD, and attempt the following steps. If these steps fail, recreate the root (/) le system by using the newfs command and restore it from a backup copy.

M
Steps

How to Restore a Bad Superblock

1. Become superuser or assume an equivalent role. 2. Determine whether the bad superblock is in the root (/) or /usr le system and select one of the following:
I

If the bad superblock is in either the root (/) or /usr le system, then boot from the network or a locally connected CD. From a locally-connected CD, use the following command:
ok boot cdrom -s

386

System Administration Guide: Devices and File Systems June 2005

From the network where a boot or install server is already setup, use the following command:
ok boot net -s

If you need help stopping the system, see Chapter 11, Booting a System (Tasks), in System Administration Guide: Basic Administration or Chapter 12, Booting a System (Tasks), in System Administration Guide: Basic Administration.
I

If the bad superblock is not in either the root (/) or /usr le system, Change to a directory outside the damaged le system and unmount the le system.
# umount /mount-point

Caution Be sure to use the newfs -N in the next step. If you omit the -N

option, you will destroy all of the data in the le system and replace it with an empty le system.

3. Display the superblock values by using the newfs -N command.

# newfs -N /dev/rdsk/device-name

The command output displays the block numbers that were used for the superblock copies when the newfs command created the le system, unless the le system was created with special parameters. For information on creating a customized le system, see Customizing UFS File System Parameters on page 392. 4. Provide an alternate superblock by using the fsck command.
# fsck -F ufs -o b=block-number /dev/rdsk/device-name

The fsck command uses the alternate superblock you specify to restore the primary superblock. You can always try 32 as an alternate block. Or, use any of the alternate blocks shown by the newfs -N command.
Example 223

Restoring a Bad Superblock

The following example shows how to restore the superblock copy 5264.

# newfs -N /dev/rdsk/c0t3d0s7 /dev/rdsk/c0t3d0s7: 163944 sectors in 506 cylinders of 9 tracks, 36 sectors 83.9MB in 32 cyl groups (16 c/g, 2.65MB/g, 1216 i/g) super-block backups (for fsck -b #) at: 32, 5264, 10496, 15728, 20960, 26192, 31424, 36656, 41888, 47120, 52352, 57584, 62816, 68048, 73280, 78512, 82976, 88208, 93440, 98672, 103904, 109136, 114368, 119600, 124832, 130064, 135296, 140528, 145760, 150992, 156224, 161456, # fsck -F ufs -o b=5264 /dev/rdsk/c0t3d0s7 Alternate superblock location: 5264. ** /dev/rdsk/c0t3d0s7 Chapter 22 Checking UFS File System Consistency (Tasks) 387

** Last Mounted on ** Phase 1 - Check Blocks and Sizes ** Phase 2 - Check Pathnames ** Phase 3 - Check Connectivity ** Phase 4 - Check Reference Counts ** Phase 5 - Check Cyl groups 36 files, 867 used, 75712 free (16 frags, 9462 blocks, 0.0% fragmentation) /dev/rdsk/c0t3d0s7 FILE SYSTEM STATE SET TO OKAY ***** FILE SYSTEM WAS MODIFIED ***** #

Syntax and Options for the fsck Command

The fsck command checks and repairs inconsistencies in le systems. If you run the fsck command without any options, it interactively asks for conrmation before making repairs. This command has four options.

Command and Option

Description

fsck -m fsck -y fsck -n fsck -o p

Checks whether a le system can be mounted Assumes a yes response for all repairs Assumes a no response for all repairs Noninteractively preens the le system, xing all expected (innocuous) inconsistencies, but exits when a serious problem is encountered

388

System Administration Guide: Devices and File Systems June 2005

CHAPTER

23

UFS File System (Reference)

This is a list of the reference information in this chapter.
I I

Structure of Cylinder Groups for UFS File Systems on page 389 Customizing UFS File System Parameters on page 392

Structure of Cylinder Groups for UFS File Systems

When you create a UFS le system, the disk slice is divided into cylinder groups. A cylinder group is comprised of one or more consecutive disk cylinders. Cylinder groups are then further divided into addressable blocks to control and organize the structure of the les within the cylinder group. Each type of block has a specic function in the le system. A UFS le system has these four types of blocks.

Block Type

Type of Information Stored

Boot block Superblock Inode Storage or data block

Information used when the system is booted Detailed information about the le system All information about a le Data for each le

The following sections provide additional information about the organization and function of these blocks.

389

Boot Block
The boot block stores objects that are used in booting the system. If a le system is not to be used for booting, the boot block is left blank. The boot block appears only in the rst cylinder group (cylinder group 0) and is the rst 8 Kbytes in a slice.

Superblock
The superblock stores much of the information about the le system, which includes the following:
I I I I I I I I I

Size and status of the le system Label, which includes the le system name and volume name Size of the le system logical block Date and time of the last update Cylinder group size Number of data blocks in a cylinder group Summary data block File system state Path name of the last mount point

Because the superblock contains critical data, multiple superblocks are made when the le system is created. A summary information block is kept within the superblock. The summary information block is not replicated, but is grouped with the primary superblock, usually in cylinder group 0. The summary block records changes that take place as the le system is used. In addition, the summary block lists the number of inodes, directories, fragments, and storage blocks within the le system.

Inodes
An inode contains all the information about a le except its name, which is kept in a directory. An inode is 128 bytes. The inode information is kept in the cylinder information block, and contains the following:
I

The type of the le:

I I I I I I I I

Regular Directory Block special Character special FIFO, also known as named pipe Symbolic link Socket Other inodes Attribute directory and shadow (used for ACLs)

390

System Administration Guide: Devices and File Systems June 2005

I I I I I I I I I

The mode of the le (the set of read-write-execute permissions) The number of hard links to the le The user ID of the owner of the le The group ID to which the le belongs The number of bytes in the le An array of 15 disk-block addresses The date and time the le was last accessed The date and time the le was last modied The date and time the le was created

The array of 15 disk-block addresses (0 to 14) points to the data blocks that store the contents of the le. The rst 12 are direct addresses. That is, they point directly to the rst 12 logical storage blocks of the le contents. If the le is larger than 12 logical blocks, the 13th address points to an indirect block, which contains direct-block addresses instead of le contents. The 14th address points to a double indirect block, which contains addresses of indirect blocks. The 15th address is for triple indirect addresses. The following gure shows this chaining of address blocks starting from the inode.
Inode

Address array

0 11 12 13 14 Double indirect block

Indirect block Storage blocks

Indirect block . . . Indirect block

FIGURE 231

Address Chain for a UFS File System

Data Blocks
Data blocks, also called storage blocks, contain the rest of the space that is allocated to the le system. The size of these data blocks is determined when a le system is created. By default, data blocks are allocated in two sizes: an 8-Kbyte logical block size, and a 1-Kbyte fragment size.
Chapter 23 UFS File System (Reference) 391

For a regular le, the data blocks contain the contents of the le. For a directory, the data blocks contain entries that give the inode number and the le name of the les in the directory.

Free Blocks
Blocks that are not currently being used as inodes, as indirect address blocks, or as storage blocks are marked as free in the cylinder group map. This map also keeps track of fragments to prevent fragmentation from degrading disk performance. To give you an idea of the structure of a typical UFS le system, the following gure shows a series of cylinder groups in a generic UFS le system.
Cylinder Group 0 Bootblock (8 Kbytes) Superblock Cylinder Group Map Inodes Storage Blocks Storage Blocks Superblock Cylinder Group Map Inodes Storage Blocks Storage Blocks Superblock Cylinder Group Map Inodes Storage Blocks Cylinder Group 1 Cylinder Group n

FIGURE 232

A Typical UFS File System

Customizing UFS File System Parameters

Before you alter the default le system parameters that are assigned by the newfs command, you need to understand them. This section describes these parameters:
I I I I

Logical Block Size on page 393 Fragment Size on page 393 Minimum Free Space on page 394 Rotational Delay on page 394 (Obsolete)

392

System Administration Guide: Devices and File Systems June 2005

I I

Optimization Type on page 395 Number of Inodes (Files) on page 395

For a description of the command options that customize these parameters, see newfs(1M) and mkfs_ufs(1M).

Logical Block Size

The logical block size is the size of the blocks that the UNIX kernel uses to read or write les. The logical block size is usually different from the physical block size. The physical block size is usually 512 bytes, which is the size of the smallest block that the disk controller can read or write. Logical block size is set to the page size of the system by default. The default logical block size is 8192 bytes (8 Kbytes) for UFS le systems. The UFS le system supports block sizes of 4096 or 8192 bytes (4 or 8 Kbytes). The recommended logical block size is 8 Kbytes.
SPARC only You can specify only the 8192-byte block size on the sun-4u platform.

To choose the best logical block size for your system, consider both the performance you want and the available space. For most UFS systems, an 8-Kbyte le system provides the best performance, offering a good balance between disk performance and the use of space in primary memory and on disk. As a general rule, to increase efficiency, use a larger logical block size for le systems when most of the les are very large. Use a smaller logical block size for le systems when most of the les are very small. You can use the quot -c lesystem command on a le system to display a complete report on the distribution of les by block size. However, the page size set when the le system is created is probably the best size in most cases.

Fragment Size
As les are created or expanded, they are allocated disk space in either full logical blocks or portions of logical blocks called fragments. When disk space is needed for a le, full blocks are allocated rst, and then one or more fragments of a block are allocated for the remainder. For small les, allocation begins with fragments. The ability to allocate fragments of blocks to les, rather than just whole blocks, saves space by reducing fragmentation of disk space that results from unused holes in blocks.
Chapter 23 UFS File System (Reference) 393

You dene the fragment size when you create a UFS le system. The default fragment size is 1 Kbyte. Each block can be divided into 1, 2, 4, or 8 fragments, which results in fragment sizes from 8192 bytes to 512 bytes (for 4-Kbyte le systems only). The lower bound is actually tied to the disk sector size, typically 512 bytes. For multiterabyte le systems, the fragment size must be equal to the le system block size.
Note The upper bound for the fragment is the logical block size, in which case the fragment is not a fragment at all. This conguration might be optimal for le systems with very large les when you are more concerned with speed than with space.

When choosing a fragment size, consider the trade-off between time and space: A small fragment size saves space, but requires more time to allocate. As a general rule, to increase storage efficiency, use a larger fragment size for le systems when most of the les are large. Use a smaller fragment size for le systems when most of the les are small.

Minimum Free Space

The minimum free space is the percentage of the total disk space that is held in reserve when you create the le system. The default reserve is ((64 Mbytes/partition size) * 100), rounded down to the nearest integer and limited between 1 percent and 10 percent, inclusively. Free space is important because le access becomes less and less efficient as a le system gets full. As long as an adequate amount of free space exists, UFS le systems operate efficiently. When a le system becomes full, using up the available user space, only root can access the reserved free space. Commands such as df report the percentage of space that is available to users, excluding the percentage allocated as the minimum free space. When the command reports that more than 100 percent of the disk space in the le system is in use, some of the reserve has been used by root. If you impose quotas on users, the amount of space available to them does not include the reserved free space. You can change the value of the minimum free space for an existing le system by using the tunefs command.

Rotational Delay
This parameter is obsolete. The value is always set to 0, regardless of the value you specify.
394 System Administration Guide: Devices and File Systems June 2005

Optimization Type
The optimization type parameter is set to either space or time.
I

Space When you select space optimization, disk blocks are allocated to minimize fragmentation and disk use is optimized. Time When you select time optimization, disk blocks are allocated as quickly as possible, with less emphasis on their placement. When sufficient free space exists, allocating disk blocks is relatively easy, without resulting in too much fragmentation. The default is time. You can change the value of the optimization type parameter for an existing le system by using the tunefs command.

For more information, see tunefs(1M).

Number of Inodes (Files)

The number of bytes per inode species the density of inodes in the le system. The number is divided into the total size of the le system to determine the number of inodes to create. Once the inodes are allocated, you cannot change the number without re-creating the le system. The default number of bytes per inode is 2048 bytes (2 Kbytes) if the le system is less than 1 Gbyte. If the le system is larger than 1 Gbyte, the following formula is used:

File System Size

Number of Bytes Per Inode

Less than or equal to 1 Gbyte Less than 2 Gbytes Less than 3 Gbytes 3 Gbytes up to 1 Tbyte

2048 4096 6144 8192

Greater than 1 Tbyte or created with -T option 1048576

If you have a le system with many symbolic links, they can lower the average le size. If your le system is going to have many small les, you can give this parameter a lower value. Note, however, that having too many inodes is much better than running out of inodes. If you have too few inodes, you could reach the maximum number of les on a disk slice that is practically empty.

Chapter 23 UFS File System (Reference)

395

Maximum UFS File and File System Size

The maximum size of a UFS le system is about 16 Tbytes of usable space, minus about one percent overhead. A sparse le can have a logical size of one terabyte. However, the actual amount of data that can be stored in a le is approximately one percent less than 1 Tbyte because of the le system overhead.

Maximum Number of UFS Subdirectories

The maximum number of subdirectories per directory in a UFS le system is 32,767. This limit is predened and cannot be changed.

396

System Administration Guide: Devices and File Systems June 2005

CHAPTER

24

Backing Up and Restoring File Systems (Overview)

This chapter provides guidelines and planning information for backing up and restoring le systems by using the ufsdump and ufsrestore commands. This is a list of the overview information in this chapter.
I I I I I I I

I I

Where to Find Backup and Restore Tasks on page 397 Introduction to Backing Up and Restoring File Systems on page 398 Why You Should Back Up File Systems on page 399 Planning Which File Systems to Back Up on page 399 Choosing the Type of Backup on page 401 Choosing a Tape Device on page 401 High-Level View of Backing Up and Restoring File Systems (Task Map) on page 402 Considerations for Scheduling Backups on page 403 Sample Backup Schedules on page 407

Where to Find Backup and Restore Tasks

Backup or Restore Task For More Information

Back up le systems by using the ufsdump command. Create UFS snapshots by using the fssnap command.

Chapter 25 Chapter 26

Restore le systems by using the ufsrestore Chapter 27 command.

397

Backup or Restore Task

For More Information

Copy les and directories by using the cpio, dd, pax, and cpio commands.

Chapter 29

Introduction to Backing Up and Restoring File Systems

Backing up le systems means copying le systems to removable media, such as tape, to safeguard against loss, damage, or corruption. Restoring le systems means copying reasonably current backup les from removable media to a working directory. This chapter describes the ufsdump and ufsrestore commands for backing up and restoring UFS le systems. Other commands are available for copying les and le systems for the purpose of sharing or transporting les. The following table provides pointers to all commands that copy individual les and le systems to other media.
TABLE 241 Task

Commands for Backing Up and Restoring Files and File Systems

Command For More Information

Back up one or more le systems to a local tape device or a remote tape device. Create read-only copies of le systems. Back up all le systems for systems on a network from a backup server.

ufsdump

Chapter 25 or Chapter 28

fssnap Solstice Backup software

Chapter 26 Solstice Backup 6.1 Administration Guide System Administration Guide: Naming and Directory Services (NIS+) Chapter 29 Chapter 29 Chapter 27

Back up and restore an NIS+ master nisbackup and server. nisrestore Copy, list, and retrieve les on a tape or diskette. Copy the master disk to a clone disk. Restore complete le systems or individual les from removable media to a working directory. tar, cpio, or pax dd

ufsrestore

398

System Administration Guide: Devices and File Systems June 2005

Why You Should Back Up File Systems

Backing up les is one of the most crucial system administration functions. You should perform regularly scheduled backups to prevent loss of data due to the following types of problems:
I I I I I

System crashes Accidental deletion of les Hardware failures Natural disasters such as re, hurricanes, or earthquakes Problems when you reinstall or upgrade a system

Planning Which File Systems to Back Up

You should back up all le systems that are critical to users, including le systems that change frequently. The following tables provide general guidelines on the le systems to back up for stand-alone systems and servers.
TABLE 242

File Systems to Back Up for Stand-alone Systems

Description Back Up Interval

File System to Back Up

root (/) slice 0

This le system contains the At regular intervals such as weekly or daily kernel and possibly the /var directory. The /var directory contains temporary les, logging les, or status les, and possibly contains frequently updated system accounting and mail les. The /usr and /opt le systems contain software and executables. The /opt directory is either part of root (/) or is its own le system. This le system can contain the directories and subdirectories of all users on the stand-alone system. Occasionally

/usr slice 6, /opt

/export/home slice 7

More often than root (/) or /usr, perhaps as often as once a day, depending on your sites needs

Chapter 24 Backing Up and Restoring File Systems (Overview)

399

TABLE 242

File Systems to Back Up for Stand-alone Systems

Description

(Continued)

File System to Back Up

Back Up Interval

/export, /var, or other le systems

The /export le system can contain the kernel and executables for diskless clients. The /var directory contains temporary les, logging les, or status les.

As your site requires

TABLE 243

File Systems to Back Up for Servers

Description Back Up Interval

File System to Back Up

root (/) slice 0

This le system contains the kernel and executables.

Once a day to once a month depending on your sites needs. If you frequently add and remove users and systems on the network, you have to change conguration les in this le system. In this case, you should do a full backup of the root (/) le system at intervals between once a week and once a month. If your site keeps user mail in the /var/mail directory on a mail server, which client systems then mount, you might want to back up root (/) daily. Or, backup the /var directory, if it is a separate le system.

/export slice 3

This le system can contain the kernel and executables for diskless clients.

Once a day to once a month, depending on your sites needs. Because the information in this le system is similar to the servers root directory in slice 0, the le system does not change frequently. You need to back up this le system only occasionally, unless your site delivers mail to client systems. Then, you should back up /export more frequently. Once a day to once a month, depending on your sites needs. These le systems are fairly static unless software is added or removed frequently.

/usr slice 6, /opt

The /usr and /opt le systems contain software and executables. The /opt directory is either part of root (/) or is its own le system.

400

System Administration Guide: Devices and File Systems June 2005

TABLE 243

File Systems to Back Up for Servers

Description

(Continued)
Back Up Interval

File System to Back Up

/export/home slice 7

This le system can contains the home directories of all the users on the system. The les in this le system are volatile.

Once a day to once a week.

Choosing the Type of Backup

You can perform full or incremental backups by using the ufsdump command. You can create a temporary image of a le system by using the fssnap command. The following table lists the differences between these types of backup procedures.
TABLE 244

Differences Between Types of Backups

Result Advantages Disadvantages

Backup Type

Full

Copies a complete le system or directory

All data is in one place

Requires large numbers of backup tapes that take a long time to write. Takes longer to retrieve individual les because the drive has to move sequentially to the point on the tape where the le is located. You might have to search multiple tapes. System performance might degrade while the snapshot is created. Finding which incremental tape contains a le can take time. You might have to go back to the last full backup.

Snapshot

Creates a temporary image of a le system Copies only those les in the specied le system that have changed since a previous backup

System can be in multiuser mode Easier to retrieve small changes in le systems

Incremental

Choosing a Tape Device

The following table shows typical tape devices that are used for storing le systems during the backup process. The storage capacity depends on the type of drive and the data being written to the tape. For more information on tape devices, see Chapter 30.
Chapter 24 Backing Up and Restoring File Systems (Overview) 401

TABLE 245

Typical Media for Backing Up File Systems

Storage Capacity

Backup Media

1/2-inch reel tape 2.5-Gbyte 1/4-inch cartridge (QIC) tape DDS3 4-mm cartridge tape (DAT) 14-Gbyte 8-mm cartridge tape DLT 7000 1/2-inch cartridge tape

140 Mbytes (6250 bpi) 2.5 Gbytes 1224 Gbytes 14 Gbytes 3570 Gbytes

High-Level View of Backing Up and Restoring File Systems (Task Map)

Use this task map to identify all the tasks for backing up and restoring le systems. Each task points to a series of additional tasks, such as determining the type of backup to perform.

Task

Description

For Instructions

1. Identify the le systems to back up. 2. Determine the type of backup. 3. Create the backup.

Identify which le systems need to be backed up on a daily, weekly, or monthly basis. Determine the type of backup you need for the le systems at your site. Use one of the following methods: If you want to have full and incremental backups of your le systems, use the ufsdump command. If you want to create a snapshot of a le system while it is active and mounted, consider using the fssnap command.

Planning Which File Systems to Back Up on page 399 Choosing the Type of Backup on page 401

Chapter 25

Chapter 26

Chapter 29 If you just want to have full backups of your personal home directory or smaller, less-important le systems, use the tar, cpio, or pax commands.

402

System Administration Guide: Devices and File Systems June 2005

Task

Description

For Instructions

4. (Optional) Restore a le system.

Select the restoration method that is based on the command used to back up the les or le system: Restore a le system backup that was created with the ufsdump command. Chapter 27

Restore a le system that was created Chapter 29 with the tar, cpio, or pax command. 5. (Optional) Restore the root (/) or /usr le system. Restoring the root (/) or /usr le system is more complicated than restoring a noncritical le system. You need to boot from a local CD or from the network while these le systems are being restored. How to Restore the root (/) and /usr File Systems on page 446

Considerations for Scheduling Backups

A backup schedule is the schedule that you establish to run the ufsdump command. This section identies considerations to think about when you create a backup schedule. This section also includes sample backup schedules. The backup schedule that you create depends on the following:
I I I I

Your need to minimize the number of tapes that are used for backups The time available for doing backups The time available for doing a full restore of a damaged le system The time available for retrieving individual les that are accidentally deleted

How Often Should You Do Backups?

If you do not need to minimize time requirements and the number of media that is used for backups, you can do full backups every day. However, this backup method is not realistic for most sites, so incremental backups are used most often. In this case, you should back up your site enough to so that you can restore les from the last four weeks. This schedule requires at least four sets of tapes, one set for each week. You would then reuse the tapes each month. In addition, you should archive the monthly backups for at least a year. Then, keep yearly backups for a number of years.

Chapter 24 Backing Up and Restoring File Systems (Overview)

403

Backup Interval Terms and Denitions

The following table describes backup interval terms and denitions.

Term

Denition

Snapshot Full backup Incremental backup

Creates a temporary image of a le system. Copies a complete le system or directory. Copies only those les in the specied le system that have changed since a previous backup. Incremental backup types include the following: I Daily, cumulative Copies a days worth of le changes on Monday. Then, overwrites Mondays backup with le changes from Tuesday, Wednesday, and so on. I Daily, incremental Copies a days worth of le changes so that you have distinct tapes of Mondays changes, Tuesdays changes, and so on. I Weekly cumulative Copies the les that have changed during the week and includes the previous weeks le changes. I Weekly incremental Copies the les that have changed during the week since the previous weekly backup.

Guidelines for Scheduling Backups

The following table provides guidelines for scheduling backups. For additional backup schedule considerations, see Considerations for Scheduling Backups on page 403.

404

System Administration Guide: Devices and File Systems June 2005

TABLE 246

Guidelines for Backup Schedules

Backup Interval Comments

File Restoration Need

To restore different versions Do daily incremental of les (for example, le backups every working systems that are used for day. word processing) Do not reuse the same tape for daily incremental backups.

This schedule saves all les modied that day, as well as those les still on disk that were modied since the last backup of a lower level. However, with this schedule, you should use a different tape each day because you might otherwise be unable to restore the needed version of the le. For example, a le that changed on Tuesday, and again on Thursday, goes onto Fridays lower-level backup appearing as it did Thursday night, not Tuesday night. If a user needs the Tuesday version, you cannot restore it unless you have a Tuesday backup tape (or a Wednesday backup tape). Similarly, a le that is present on Tuesday and Wednesday, but removed on Thursday, does not appear on the Friday lower-level backup.

To quickly restore a complete le system

Do lower-level backups more frequently.

This way youre not doing all level 0 backups on the same day. Only changes from day to day are saved on each daily tape. Only changes from week to week (rather than the entire month) are saved on the weekly tapes.

To back up a number of le Consider staggering the systems on the same server schedule for different le systems. To minimize the number of tapes used Increase the level of incremental backups that are done across the week. Increase the level of backups that are done at the end of the week. Put each days and weeks incremental backups onto the same tape. Put each days and weeks incremental backups onto the same tape.

To do so, use the no rewind option of the ufsdump command, such as specifying /dev/rmt/0n.

Chapter 24 Backing Up and Restoring File Systems (Overview)

405

Using Dump Levels to Create Incremental Backups

The dump level you specify in the ufsdump command (09) determines which les are backed up. Dump level 0 creates a full backup. Levels 19 are used to schedule incremental backups, but have no dened meanings. Levels 19 are just a range of numbers that are used to schedule cumulative or discrete backups. The only meaning levels 19 have is in relationship to each other, as a higher or lower number. A lower dump number always restarts a full or a cumulative backup. The following examples show the exibility of the incremental dump procedure using levels 19.

ExampleDump Levels for Daily, Cumulative Backups

Doing daily, cumulative incremental backups is the most commonly used backup schedule and is recommended for most situations. The following example shows a schedule that uses a level 9 dump Monday through Thursday, and a level 5 dump on Friday restarts process.
Monthly 0 Monday 9 Tuesday 9 Wednesday 9 Thursday 9 Friday 5

FIGURE 241

Incremental Backup: Daily Cumulative

In the preceding example, you could have used other numbers in the 19 range to produce the same results. The key is using the same number Monday through Thursday, with any lower number on Friday. For example, you could have specied levels 4, 4, 4, 4, 2 or 7, 7, 7, 7, 5.

ExampleDump Levels for Daily, Incremental Backups

The following example shows a schedule where you capture only a days work on different tapes. This type of backup is referred to as a daily, incremental backup. In this case, sequential dump level numbers are used during the week (3, 4, 5, 6) with a lower number (2) on Friday. The lower number on Friday restarts the processing.

406

System Administration Guide: Devices and File Systems June 2005

Monthly 0

Monday 3

Tuesday 4

Wednesday 5

Thursday 6

Friday 2

FIGURE 242

Incremental Backup: Daily Incremental

In the preceding example, you could have used the sequence 6, 7, 8, 9 followed by 2, or 5, 6, 7, 8 followed by 3. Remember, the numbers themselves have no dened meaning. You attribute meaning by ordering them in a specied sequence, as described in the examples.

Sample Backup Schedules

This section provides sample backup schedules. All schedules assume that you begin with a full backup (dump level 0), and that you use the -u option to record each backup in the /etc/dumpdates le.

ExampleDaily Cumulative, Weekly Cumulative Backup Schedule

Table 247 shows the most commonly used incremental backup schedule. This schedule is recommended for most situations. With this schedule, the following occurs:
I

Each day, all les that have changed since the lower-level backup at the end of the previous week are saved. For each weekday level 9 backup, the previous level 0 or level 5 backup is the closest backup at a lower level. Therefore, each weekday tape contains all the les that changed since the end of the previous week or the initial level 0 backup for the rst week. For each Friday level 5 backup, the closest lower-level backup is the level 0 backup done at the beginning of the month. Therefore, each Fridays tape contains all the les changed during the month up to that point.

Chapter 24 Backing Up and Restoring File Systems (Overview)

407

TABLE 247

Daily Cumulative/Weekly Cumulative Backup Schedule

Floating Mon Tues Wed Thurs Fri

1st of Month Week 1 Week 2 Week 3 Week 4

0 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 5 5 5 5

The following table shows how the contents of the tapes can change across two weeks with the daily cumulative, weekly cumulative schedule. Each letter represents a different le.
TABLE 248

Contents of Tapes for Daily Cumulative/Weekly Cumulative Backup Schedule

Mon Tues Wed Thurs Fri

Week 1 Week 2

ab g

abc gh

abcd ghi

abcde ghij

abcdef abcdefghijk

Tape Requirements for the Daily Cumulative, Weekly Cumulative Schedule

With this schedule, you need six tapes if you want to reuse daily tapes. However, you need nine tapes if you want to use four different daily tapes:
I I I

One tape for the level 0 backup Four tapes for Fridays One or four daily tapes

If you need to restore a complete le system, you need the following tapes:
I I I

The level 0 tape The most recent Friday tape The most recent daily tape since the last Friday tape, if any

ExampleDaily Cumulative, Weekly Incremental Backup Schedule

The following table shows a schedule where each weekday tape accumulates all les that changed since the beginning of the week, or the initial level 0 backup for the rst week. In addition, each Fridays tape contains all the les that changed that week.

408

System Administration Guide: Devices and File Systems June 2005

TABLE 249

Daily Cumulative, Weekly Incremental Backup Schedule

Floating Mon Tues Wed Thurs Fri

1st of Month Week 1 Week 2 Week 3 Week 4

0 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 3 4 5 6

The following table shows how the contents of the tapes can change across two weeks with the daily cumulative, weekly incremental backup schedule. Each letter represents a different le.
TABLE 2410

Contents of Tapes for Daily Cumulative, Weekly Incremental Backup Schedule

Mon Tues Wed Thurs Fri

Week 1 Week 2

ab g

abc gh

abcd ghi

abcde ghij

abcdef ghijk

Tape Requirements for the Daily Cumulative, Weekly Incremental Backup Schedule
With this schedule, you need six tapes if you want to reuse daily tapes. However, you need nine tapes if you want to use four different daily tapes:
I I I

One tape for the level 0 backup Four tapes for Fridays One or four daily tapes

If you need to restore a complete le system, you need the following tapes:
I I I

The level 0 tape All the Friday tapes The most recent daily tape since the last Friday tape, if any

ExampleDaily Incremental, Weekly Cumulative Backup Schedule

The following table shows a schedule where each weekday tape contains only the les that changed since the previous day. In addition, each Fridays tape contains all les changed since the initial level 0 backup at the beginning of the month.
Chapter 24 Backing Up and Restoring File Systems (Overview) 409

TABLE 2411

Daily Incremental, Weekly Cumulative Backup Schedule

Floating Mon Tues Wed Thurs Fri

1st of Month Week 1 Week 2 Week 3 Week 4

0 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 2 2 2 2

The following table shows how the contents of the tapes can change across two weeks with the daily incremental, weekly cumulative schedule. Each letter represents a different le.
TABLE 2412

Contents of Tapes for Daily Incremental, Weekly Cumulative Backup Schedule

Mon Tues Wed Thurs Fri

Week 1 Week 2

ab jkl

cd m

efg no

hi pq

abcdefghi abcdefghijk lmnopqrs

Tape Requirements for Daily Incremental, Weekly Cumulative Schedule

With this schedule, you need at least 9 tapes if you want to reuse daily tapes, which is not recommended. Preferably, you need 21 tapes if you save weekly tapes for a month: one tape for the level 0, 4 tapes for the Fridays, and 4 or 16 daily tapes.
I I I

1 tape for the level 0 backup 4 tapes for all the Friday backups 4 or 16 daily tapes

If you need to restore the complete le system, you need the following tapes:
I I I

The level 0 tape The most recent Friday tape All the daily tapes since the last Friday tape, if any

ExampleMonthly Backup Schedule for a Server

The following table shows an example backup strategy for a heavily used le server on a small network where users are doing le-intensive work, such as program development or document production. This example assumes that the backup period begins on a Sunday and consists of four seven-day weeks.
410 System Administration Guide: Devices and File Systems June 2005

TABLE 2413 Directory

Example of Monthly Backup Schedule for a Server

Date Dump Level Tape Name

root (/) /usr /export /export/home

1st Sunday 1st Sunday 1st Sunday 1st Sunday 1st Monday 1st Tuesday 1st Wednesday 1st Thursday 1st Friday 1st Saturday

0 0 0 0 9 9 5 9 9 5 0 0 0 0 9 9 5 9 9 5 0 0 0 0 9 9 5 9

n tapes n tapes n tapes n tapes A B C D E F n tapes n tapes n tapes n tapes G H I J K L n tapes n tapes n tapes n tapes M N O P

root (/) /usr /export /export/home

2nd Sunday 2nd Sunday 2nd Sunday 2nd Sunday 2nd Monday 2nd Tuesday 2nd Wednesday 2nd Thursday 2nd Friday 2nd Saturday

root (/) /usr /export /export/home

3rd Sunday 3rd Sunday 3rd Sunday 3rd Sunday 3rd Monday 3rd Tuesday 3rd Wednesday 3rd Thursday

Chapter 24 Backing Up and Restoring File Systems (Overview)

411

TABLE 2413 Directory

Example of Monthly Backup Schedule for a Server

Date Dump Level

(Continued)
Tape Name

3rd Friday 3rd Saturday root (/) /usr /export /export/home 4th Sunday 4th Sunday 4th Sunday 4th Sunday 4th Monday 4th Tuesday 4th Wednesday 4th Thursday 4th Friday 4th Saturday

9 5 0 0 0 0 9 9 5 9 9 5

Q R n tapes n tapes n tapes n tapes S T U V W X

With this schedule, you use 4n tapes, the number of tapes needed for 4 full backups of the root (/), /usr, /export, and /export/home le systems. Also, you need 24 additional tapes for the incremental backups of the /export/home le systems. This schedule assumes that each incremental backup uses one tape and that you save the tapes for a month. Heres how this schedule works: 1. On each Sunday, do a full backup (level 0) of the root (/), /usr, /export, and /export/home le systems. Save the level 0 tapes for at least three months. 2. On the rst Monday of the month, use tape A to do a level 9 backup of the /export/home le system. The ufsdump command copies all les changed since the previous lower-level backup. In this case, the previous lower-level backup is the level 0 backup that you did on Sunday. 3. On the rst Tuesday of the month, use tape B to do a level 9 backup of the /export/home le system. Again, the ufsdump command copies all les changed since the last lower-level backup, which is Sundays level 0 backup. 4. On the rst Wednesday of the month, use tape C to do a level 5 backup of the /export/home le system. The ufsdump command copies all les that changed since Sunday. 5. Do the Thursday and Friday level 9 backups of the /export/home le system on tapes D and E. The ufsdump command copies all les that changed since the last lower-level backup, which is Wednesdays level 5 backup.

412

System Administration Guide: Devices and File Systems June 2005

6. On the rst Saturday of the month, use tape F to do a level 5 backup of /export/home. The ufsdump command copies all les changed since the previous lower-level backup (in this case, the level 0 backup you did on Sunday). Store tapes AF until the rst Monday of the next four-week period, when you use them again. 7. Repeat steps 16 for the next three weeks, using tapes GL and 4n tapes for the level 0 backup on Sunday, and so on. 8. For each four-week period, repeat steps 17, using a new set of tapes for the level 0 backups and reusing tapes AX for the incremental backups. The level 0 tapes could be reused after three months. This schedule lets you save les in their various states for a month. This plan requires many tapes, but ensures that you have a library of tapes to draw upon. To reduce the number of tapes, you could reuse Tapes AF each week.

Chapter 24 Backing Up and Restoring File Systems (Overview)

413

414

System Administration Guide: Devices and File Systems June 2005

CHAPTER

25

Backing Up Files and File Systems (Tasks)

This chapter describes the procedures for backing up le systems by using the ufsdump command. For information on these procedures, see Backing Up Files and File System (Task Map) on page 415. For overview information about performing backups, see Chapter 24. For information about backing up individual les to diskettes, see Chapter 29. For additional information on the ufsdump command, see Chapter 28.

Backing Up Files and File System (Task Map)

Task Description For Instructions

1. Prepare for le system Identify the le systems, the type of Preparing for File System backups. backup, and the tape device to be used Backups on page 416 for the backups. 2. Determine the number Determine the number of tapes that of tapes needed to back are needed for a full backup of a le up a le system. system. How to Determine the Number of Tapes Needed for a Full Backup on page 417

415

Task

Description

For Instructions

3. Back up le systems.

Perform a full backup of le systems to get baseline copies of all les. Perform an incremental backup of le systems based on whether keeping copies of les that have changed on a daily basis is important at your site.

How to Back Up a File System to Tape on page 418

Preparing for File System Backups

The preparation for backing up le systems begins with planning, which is described in Chapter 24 and includes choosing the following:
I I I I

The le systems to back up The type of backup (full or incremental) to perform A backup schedule A tape drive

For more information, see Chapter 24. This section describes two other tasks you might need to perform before you back up le systems:
I I

Finding the names of le systems to back up Determining the number of tapes that are needed for a full backup

M
Steps

How to Find File System Names

1. Display the contents of the /etc/vfstab le.
$ more /etc/vfstab

2. Look in the mount point column for the name of the le system. 3. Use the directory name listed in the mount point column when you back up the le system.
Example 251

Finding File System Names

In this example, the le systems to be backed up are root (/), /usr, /datab, and /export/home.
$ more /etc/vfstab #device device mount FS fsck mount mount

416

System Administration Guide: Devices and File Systems June 2005

#to mount /devices . . . /proc /dev/dsk/c0t0d0s1 /dev/dsk/c0t0d0s0 /dev/dsk/c0t0d0s6 /dev/dsk/c0t0d0s5 /dev/dsk/c0t0d0s7 swap

to fsck -

point /devices

type pass at boot options devfs no -

/dev/rdsk/c0t0d0s0 /dev/rdsk/c0t0d0s6 /dev/rdsk/c0t0d0s5 /dev/rdsk/c0t0d0s7 -

/proc / /usr /datab /export/home /tmp

proc swap ufs ufs ufs ufs tmpfs

1 1 2 2 -

no no no no yes yes yes

How to Determine the Number of Tapes Needed for a Full Backup

1. Become superuser or assume an equivalent role. 2. Estimate the size of the backup in bytes.
# ufsdump [0]S le-system

Steps

Use the S option to display the estimated number of bytes that are needed to do the backup if this is the rst backup of the le system. Use the 0S option to display the estimated number of bytes that are needed to do the backup if this is not the rst backup of the le system. 3. Divide the estimated size by the capacity of the tape to determine how many tapes you need. For a list of tape capacities, see Table 245.
Example 252

Determining the Number of Tapes

In this example, the le system of 489,472 bytes easily ts on a 150-Mbyte tape.
# ufsdump S /export/home 489472

Backing Up a File System

The following are general guidelines for performing backups:
I

Use single-user mode or unmount the le system, unless you are creating a snapshot of a le system. For information about UFS snapshots, see Chapter 26.
Chapter 25 Backing Up Files and File Systems (Tasks) 417

Be aware that backing up le systems when directory-level operations (such as creating, removing, and renaming les) and le-level activity are occurring simultaneously means that some data will not be included in the backup. You can run the ufsdump command from a single system and remotely back up groups of systems across the network through remote shell or remote login. In addition, you can direct the output to the system on which the tape device is located. Typically, the tape device is located on the system from which you run the ufsdump command, but it does not have to be. Another way to back up les to a remote device is to pipe the output from the ufsdump command to the dd command. For information about using the dd command, see Chapter 29.

If you are doing remote backups across the network, the system with the tape device must have entries in its /.rhosts le for each client that will be using the device. Also, the system that initiates the backup must be included in the /.rhosts le on each system that it will back up.

How to Back Up a File System to Tape

The following are general steps for backing up le systems by using the ufsdump command. The examples show specic uses of options and arguments.

Steps

1. Become superuser or assume an equivalent role. 2. Bring the system to run level S (single-user mode). For example:
# shutdown -g30 -y

3. (Optional) Check the le system for consistency. For example:

# fsck -m /dev/rdsk/c0t0d0s7

The fsck -m command checks for the consistency of le systems. For example, power failures can leave les in an inconsistent state. For more information on the fsck command, see Chapter 22. 4. If you need to back up le systems to a remote tape drive, follow these steps: a. On the system to which the tape drive is attached (the tape server), add the following entry to its /.rhosts le:
host root

The host entry species the name of the system on which you will run the ufsdump command to perform the backup. b. On the tape server, verify that the host added to the /.rhosts le is accessible through the name service.
418 System Administration Guide: Devices and File Systems June 2005

5. Identify the device name of the tape drive. The default tape drive is the /dev/rmt/0 device. 6. Insert a tape that is write-enabled into the tape drive. 7. Back up le systems.
# ufsdump options arguments lenames

You can back up le systems or directories, or les within le systems. For information on backing up individual les, see tar(1) or cpio(1). The following examples show how to use the most common ufsdump options and arguments:
I I I I

Example 253 Example 254 Example 255 Example 256

For other ufsdump options and arguments, see Chapter 28. 8. If prompted, remove the tape and insert the next tape volume. 9. Label each tape with the volume number, dump level, date, system name, disk slice, and le system. 10. Bring the system back to run level 3 by pressing Control-D. 11. Verify that the backup was successful.
# ufsrestore tf device-name

Example 253

Performing a Full Backup of root (/)

The following example shows how to do a full backup of the root (/) le system. The system in this example is brought to single-user mode before the backup. The following ufsdump options are included:
I I I I I

0 species a 0 level dump (or a full backup). u species that the /etc/dumpdates le is updated with the date of this backup. c identies a cartridge tape device. f /dev/rmt/0 identies the tape device. / is the le system being backed up.

For example:
# init 0 ok boot -s # ufsdump 0ucf /dev/rmt/0 / DUMP: Date of this level 0 dump: Wed Jul 28 16:13:52 2004 DUMP: Date of last level 0 dump: the epoch DUMP: Dumping /dev/rdsk/c0t0d0s0 (starbug:/) to /dev/rmt/0. Chapter 25 Backing Up Files and File Systems (Tasks) 419

DUMP: Mapping (Pass I) [regular files] DUMP: Mapping (Pass II) [directories] DUMP: Writing 63 Kilobyte records DUMP: Estimated 363468 blocks (177.47MB). DUMP: Dumping (Pass III) [directories] DUMP: Dumping (Pass IV) [regular files] DUMP: Tape rewinding DUMP: 369934 blocks (180.63MB) on 1 volume at 432 KB/sec DUMP: DUMP IS DONE DUMP: Level 0 dump on Wed Jul 28 16:13:52 2004 # ufsrestore tf /dev/rmt/0 2 . 3 ./lost+found 4 ./usr 5 ./export 6 ./export/home 7 ./var 8 ./var/sadm 9 ./var/sadm/install 10 ./var/sadm/install/admin 823 ./var/sadm/install/admin/default 11 ./var/sadm/install/logs 697 ./var/sadm/install/logs/SUNWmpatchmgr 905 ./var/sadm/install/logs/Additional_Software_install... 906 ./var/sadm/install/logs/Additional_Software_install... 13 ./var/sadm/install/.lockfile 14 ./var/sadm/install/install.db 824 ./var/sadm/install/special_contents 838 ./var/sadm/install/contents . . . # (Press Control-D to bring system to run level 3)

Example 254

Performing an Incremental Backup of root (/)

The following example shows how to do an incremental backup of the root (/) le system in single-user mode. The following ufsdump options are included:
I I I I I

9 species a 9 level dump (or an incremental backup). u species that the /etc/dumpdates le is updated with the date of this backup. c identies a cartridge tape device. f /dev/rmt/0 identies the tape device. / is the le system being backed up.

# init 0 ok boot -s # ufsdump 9ucf DUMP: Date of DUMP: Date of DUMP: Dumping DUMP: Mapping 420

/dev/rmt/0 / this level 9 dump: Wed Jul 28 14:26:50 2004 last level 0 dump: Wed Jul 28 11:15:41 2004 /dev/rdsk/c0t0d0s0 (starbug:/) to /dev/rmt/0. (Pass I) [regular files]

System Administration Guide: Devices and File Systems June 2005

DUMP: Mapping (Pass II) [directories] DUMP: Writing 63 Kilobyte records DUMP: Estimated 335844 blocks (163.99MB). DUMP: Dumping (Pass III) [directories] DUMP: Dumping (Pass IV) [regular files] DUMP: 335410 blocks (163.77MB) on 1 volume at 893 KB/sec DUMP: DUMP IS DONE DUMP: Level 9 dump on Wed Jul 28 14:30:50 2004 # ufsrestore tf /dev/rmt/0 2 . 3 ./lost+found 5696 ./usr 11392 ./var 17088 ./export 22784 ./export/home 28480 ./opt 5697 ./etc 11393 ./etc/default 11394 ./etc/default/sys-suspend 11429 ./etc/default/cron 11430 ./etc/default/devfsadm 11431 ./etc/default/dhcpagent 11432 ./etc/default/fs 11433 ./etc/default/inetinit 11434 ./etc/default/kbd 11435 ./etc/default/nfslogd 11436 ./etc/default/passwd 11437 ./etc/default/tar . . .

Example 255

Performing a Full Backup of a Home Directory

The following example shows how to do a full backup of the /export/home/kryten home directory. The following ufsdump options are included:
I I I I I

0 species that this is a 0 level dump (or a full backup). u species that the /etc/dumpdates le is updated with the date of this backup. c identies a cartridge tape device. f /dev/rmt/0 identies the tape device. /export/home/kryten is the directory being backed up.

# ufsdump 0ucf /dev/rmt/0 /export/home/kryten DUMP: Date of this level 0 dump: Wed Jul 28 15:02:48 2004 DUMP: Date of last level 0 dump: the epoch DUMP: Dumping /dev/rdsk/c0t0d0s7 (starbug:/export/home) to /dev/rmt/0. DUMP: Mapping (Pass I) [regular files] DUMP: Mapping (Pass II) [directories] DUMP: Writing 63 Kilobyte records DUMP: Estimated 2412 blocks (1.18MB). Chapter 25 Backing Up Files and File Systems (Tasks) 421

DUMP: Dumping (Pass III) [directories] DUMP: Dumping (Pass IV) [regular files] DUMP: 2392 blocks (1.17MB) on 1 volume at 4241 KB/sec DUMP: DUMP IS DONE # ufsrestore tf /dev/rmt/0 232 ./kryten 233 ./kryten/filea 234 ./kryten/fileb 235 ./kryten/filec 236 ./kryten/letters 237 ./kryten/letters/letter1 238 ./kryten/letters/letter2 239 ./kryten/letters/letter3 240 ./kryten/reports 241 ./kryten/reports/reportA 242 ./kryten/reports/reportB 243 ./kryten/reports/reportC #

Example 256

Performing a Full Backup to a Remote System (Solaris 10 Data to Solaris 10 System)

The following example shows how to do a full backup of a local /export/home le system on a Solaris 10 system (mars) to a tape device on a remote Solaris 10 system (earth) in single-user mode. The following ufsdump options are included:
I I I I I

0 species a 0 level dump (or a full backup). u species that the /etc/dumpdates le is updated with the date of this backup. c identies a cartridge tape device. f earth:/dev/rmt/0 identies the remote system name and tape device. /export/home is the le system being backed up.

# ufsdump 0ucf earth:/dev/rmt/0 /export/home DUMP: Date of this level 0 dump: Wed Jul 28 15:52:59 2004 DUMP: Date of last level 0 dump: the epoch DUMP: Dumping /dev/rdsk/c0t0d0s7 (mars:/export/home) to earth:/dev/rmt/0. DUMP: Mapping (Pass I) [regular files] DUMP: Mapping (Pass II) [directories] DUMP: Writing 63 Kilobyte records DUMP: Estimated 8282 blocks (4.04MB). DUMP: Dumping (Pass III) [directories] DUMP: Dumping (Pass IV) [regular files] DUMP: Tape rewinding DUMP: 8188 blocks (4.00MB) on 1 volume at 67 KB/sec DUMP: DUMP IS DONE DUMP: Level 0 dump on Wed Jul 28 15:52:59 2004 # ufsrestore tf earth:/dev/rmt/0 2 . 3 ./lost+found 4 ./kryten 5 ./kryten/filea 6 ./kryten/fileb 422 System Administration Guide: Devices and File Systems June 2005

7 8 9 10 11 12 . . . #

./kryten/filec ./kryten/letters ./kryten/letters/letter1 ./kryten/letters/letter2 ./kryten/letters/letter3 ./kryten/reports

Chapter 25 Backing Up Files and File Systems (Tasks)

423

424

System Administration Guide: Devices and File Systems June 2005

CHAPTER

26

Using UFS Snapshots (Tasks)

This chapter describes how to create and back up UFS snapshots. For information on the procedures associated with creating UFS snapshots, see Using UFS Snapshots (Task Map) on page 425. For overview information about performing backups, see Chapter 24.

Using UFS Snapshots (Task Map)

Task Description For Instructions

1. Create a UFS snapshot. 2. Display UFS snapshot information. 3. (Optional) Delete a UFS snapshot. 4. (Optional) Back up a UFS snapshot.

Create a read-only copy of a le system by using the fssnap command. Identify UFS snapshot information such as the raw snapshot device. Delete a snapshot that is already backed up or no longer needed. Choose one of the following backup methods:

How to Create a UFS Snapshot on page 429 How to Display UFS Snapshot Information on page 430 How to Delete a UFS Snapshot on page 432

Create a full backup of a UFS snapshot How to Create a Full by using the ufsdump command. Backup of a UFS Snapshot (ufsdump) on page 433

425

Task

Description

For Instructions

Create an incremental backup of a UFS How to Create an snapshot by using the ufsdump Incremental Backup of a command. UFS Snapshot (ufsdump) on page 433 Back up a UFS snapshot by using the tar command. 5. (Optional) Restore data from a UFS snapshot. Restore the UFS snapshot the same way as you would restore data by using the ufsrestore command. How to Back Up a UFS Snapshot (tar) on page 434 How to Restore a Complete File System on page 443

UFS Snapshots Overview

You can use the fssnap command to back up le systems while the le system is mounted. This command to creates a read-only snapshot of a le system. A snapshot is a le systems temporary image that is intended for backup operations. When the fssnap command is run, it creates a virtual device and a backing-store le. You can back up the virtual device, which looks and acts like a real device, with any of the existing Solaris backup commands. The backing-store le is a bitmap le that contains copies of pre snapshot data that has been modied since the snapshot was taken. Keep the following key points in mind when specifying backing-store les:
I

The destination path of the backing store les must have enough free space to hold the le system data. The size of the backing store les vary with the amount of activity on the le system. The backing store le location must be different from the le system that is being captured in a snapshot. The backing-store les can reside on any type of le system, including another UFS le system or an NFS le system. Multiple backing-store les are created when you create a snapshot of a UFS le system that is larger than 512 Gbytes. Backing-store les are sparse les. The logical size of a sparse le, as reported by the ls command, is not the same as the amount of space that has been allocated to the sparse le, as reported by the du command.

For more information about creating snapshots for a UFS le system larger than 512 Gbytes, see Creating a Multiterabyte UFS Snapshot on page 428.
426 System Administration Guide: Devices and File Systems June 2005

Why Use UFS Snapshots?

The UFS snapshots feature provides additional availability and convenience for backing up a le system because the le system remains mounted and the system remains in multiuser mode during backups. Then, you can use the tar or cpio commands to back up a UFS snapshot to tape for more permanent storage. If you use the ufsdump command to perform backups, the system should be in single-user mode to keep the le system inactive when you perform backups. The fssnap command gives administrators of non enterprise-level systems the power of enterprise-level tools, such as Sun StorEdge Instant Image, without the large storage demands. The UFS snapshots feature is similar to the Instant Image product. Although UFS snapshots can make copies of large le systems, Instant Image is better suited for enterprise-level systems. UFS snapshots is better suited for smaller systems. Instant Image allocates space equal to the size of the entire le system that is being captured. However, the backing-store le that is created by UFS snapshots occupies only as much disk space as needed. This table describes specic differences between UFS snapshots and Instant Image.

UFS Snapshots

Sun StorEdge Instant Image

Size of the backing-store le depends on how much data has changed since the snapshot was taken Does not persist across system reboots Works on UFS le systems Available starting with the Solaris 8 1/01 release

Size of the backing-store le equals the size of the entire le system being copied Persists across system reboots Cannot be used with root (/) or /usr le systems Part of Sun StorEdge products

UFS Snapshots Performance Issues

When the UFS snapshot is rst created, users of the le system might notice a slight pause. The length of the pause increases with the size of the le system to be captured. While the snapshot is active, users of the le system might notice a slight performance impact when the le system is written to. However, they see no impact when the le system is read.

Chapter 26 Using UFS Snapshots (Tasks)

427

Creating and Deleting UFS Snapshots

When you use the fssnap command to create a UFS snapshot, observe how much disk space the backing-store le consumes. The backing-store le initially uses no space, and then it grows quickly, especially on heavily used systems. Make sure that the backing-store le has enough space to expand. Or, limit its size with the -o maxsize=n [k,m,g] option, where n [k,m,g] is the maximum size of the backing-store le.
Caution If the backing-store le runs out of space, the snapshot might delete itself, which causes the backup to fail. Check the /var/adm/messages le for possible snapshot errors.

You can also specify a directory for the backing-store path, which means a backing store le is created in the directory specied. For example, if /var/tmp is specied for the backing-store path, the following backing-store le is created.
/var/tmp/snapshot0

If you created one large root (/) le system instead of creating separate le systems for /export/home, /usr, and so on, you will be unable to create a snapshot of those separate le systems. For example, this system does not have a separate le system for /usr as indicated under the Mounted on column:
# df -k /usr Filesystem /dev/dsk/c0t0d0s0 kbytes used avail capacity 3618177 2190002 1391994 62% Mounted on /

If you attempt to create a snapshot for the /usr le system, you will see a message similar to the following:
# fssnap -F ufs -o bs=/snaps/usr.back.file /usr snapshot error: Invalid backing file path

This message indicates that you cannot have the backing store le on the same le system as the le system being snapped, which is the case for the /usr le system, in this example. For more information, see the fssnap_ufs(1M) man page.

Creating a Multiterabyte UFS Snapshot

Creating a multiterabyte UFS snapshot is identical to creating a snapshot for a smaller UFS le system. The only difference is that multiple backing store les are created for each 512 Gbytes of le system space.
428 System Administration Guide: Devices and File Systems June 2005

Keep the following key points in mind when creating a snapshot for a le system that is larger than 512 Gbytes:
I

Multiple backing store les are created.

I

If you specify a backing store le name when the snapshot is created, then the subsequent backing store le names will be interated based on the le name that you specify. The subsequent backing-store les will have the same name, but with the suffixes .2, .3, and so on. If you only specify a backing store le destination (or directory) and not a backing store le name, then multiple backing store le names will be created and iterated with the suffixes .2, .3, and so on.

The fssnap -i command only reports the rst backing store le name even if multiple backing store les have been created. However, the reported backing-store length is the combined sizes of all the backing store les for the snapshot.
Note Backing-store les are sparse les. The logical size of a sparse le, as reported by the ls command, is not the same as the amount of space that has been allocated to the sparse le, as reported by the du command.

After you have backed up the snapshot or you would just like to remove the snapshot, you will have to remove the backing store les manually if you did not use the unlink option when the snapshot was created.

For an example of creating a snapshot for a le system that is larger than 512 Gbytes, see Example 262. For more information, see fssnap_ufs(1M).

M
Steps

How to Create a UFS Snapshot

1. Become superuser or assume an equivalent role. 2. Make sure that the le system has enough disk space for the backing-store le.
# df -k

3. Make sure that a backing-store le of the same name and location does not already exist.
# ls /backing-store-le

4. Create the UFS snapshot.

# fssnap -F ufs -o bs=/backing-store-le /le-system Chapter 26 Using UFS Snapshots (Tasks) 429

Note The backing-store le must reside on a different le system than the le system that is being captured using UFS snapshots.

5. Verify that the snapshot has been created.

# /usr/lib/fs/ufs/fssnap -i /le-system

Example 261

Creating a UFS Snapshot

The following example shows how to create a snapshot of the /usr le system. The backing-store le is /scratch/usr.back.file. The virtual device is /dev/fssnap/1.
# fssnap -F ufs -o bs=/scratch/usr.back.file /usr /dev/fssnap/1

The following example shows how to limit the backing-store le to 500 Mbytes.
# fssnap -F ufs -o maxsize=500m,bs=/scratch/usr.back.file /usr /dev/fssnap/1

Example 262

Creating a Multiterabyte UFS Snapshot

The following example shows how to create a snapshot of a 1.6 Tbyte UFS le system.
# fssnap -F ufs -o bs=/var/tmp /data2 /dev/fssnap/0 # /usr/lib/fs/ufs/fssnap -i Snapshot number : 0 Block Device : /dev/fssnap/0 Raw Device : /dev/rfssnap/0 Mount point : /data2 Device state : idle Backing store path : /var/tmp/snapshot0 Backing store size : 0 KB Maximum backing store size : Unlimited Snapshot create time : Fri Sep 10 13:13:02 2004 Copy-on-write granularity : 32 KB # ls /var/tmp snapshot0 snapshot0.2 snapshot0.3 snapshot0.4

How to Display UFS Snapshot Information

You can display the current snapshots on the system by using the fssnap -i option. If you specify a le system, you see detailed information about that le system snapshot. If you dont specify a le system, you see information about all of the current UFS snapshots and their corresponding virtual devices.

430

System Administration Guide: Devices and File Systems June 2005

Note Use the UFS le system-specic fssnap command to view the extended

snapshot information as shown in the following examples.

Steps

1. Become superuser or assume an equivalent role. 2. List all current snapshots. For example:
# /usr/lib/fs/ufs/fssnap -i Snapshot number Block Device Raw Device Mount point Device state Backing store path Backing store size Maximum backing store size Snapshot create time Copy-on-write granularity : : : : : : : : : : 0 /dev/fssnap/0 /dev/rfssnap/0 /export/home idle /var/tmp/home.snap0 0 KB Unlimited Thu Jul 01 14:50:38 2004 32 KB

3. Display detailed information about a specic snapshot. For example:

# /usr/lib/fs/ufs/fssnap -i /export Snapshot number : 1 Block Device : /dev/fssnap/1 Raw Device : /dev/rfssnap/1 Mount point : /export Device state : idle Backing store path : /var/tmp/export.snap0 Backing store size : 0 KB Maximum backing store size : Unlimited Snapshot create time : Thu Jul 01 15:03:22 2004 Copy-on-write granularity : 32 KB

Deleting a UFS Snapshot

When you create a UFS snapshot, you can specify that the backing-store le is unlinked. An unlinked backing-store le is removed after the snapshot is deleted. If you dont specify the -o unlink option when you create a UFS snapshot, you must manually delete the backing-store le. The backing-store le occupies disk space until the snapshot is deleted, whether you use the -o unlink option to remove the backing-store le or you manually delete the le.

Chapter 26 Using UFS Snapshots (Tasks)

431

How to Delete a UFS Snapshot

You can delete a snapshot either by rebooting the system or by using the fssnap -d command. When you use this command, you must specify the path of the le system that contains the UFS snapshot.

Steps

1. Become superuser or assume an equivalent role. 2. Identify the snapshot to be deleted.

# /usr/lib/fs/ufs/fssnap -i

3. Delete the snapshot.

# fssnap -d /le-system Deleted snapshot 1.

4. If you did not use the -o unlink option when you created the snapshot, manually delete the backing-store le.
# rm /le-system/backing-store-le

Example 263

Deleting a UFS Snapshot

The following example shows how to delete a snapshot and assumes that the -o unlink option was not used.
# fssnap -i 0 /export/home 1 /export # fssnap -d /usr Deleted snapshot 1. # rm /var/tmp/export.snap0

Backing Up a UFS Snapshot

You can create a full backup or an incremental backup of a UFS snapshot. You can use the standard Solaris backup commands to back up a UFS snapshot. The virtual device that contains the UFS snapshot acts as a standard read-only device. So, you can back up the virtual device as if you were backing up a le system device. If you are using the ufsdump command to back up a UFS snapshot, you can specify the snapshot name during the backup. See the following procedure for more information.
432 System Administration Guide: Devices and File Systems June 2005

How to Create a Full Backup of a UFS Snapshot (ufsdump)

1. Become superuser or assume an equivalent role. 2. Identify the UFS snapshot to be backed up.
# /usr/lib/fs/ufs/fssnap -i /le-system

Steps

For example:
# /usr/lib/fs/ufs/fssnap -i /usr Snapshot number : 1 Block Device : /dev/fssnap/1 Raw Device : /dev/rfssnap/1 Mount point : /usr Device state : idle Backing store path : /var/tmp/usr.snap0 Backing store size : 0 KB Maximum backing store size : Unlimited Snapshot create time : Thu Jul 01 15:17:33 2004 Copy-on-write granularity : 32 KB

3. Back up the UFS snapshot.

# ufsdump 0ucf /dev/rmt/0 /snapshot-name

For example:
# ufsdump 0ucf /dev/rmt/0 /dev/rfssnap/1

4. Verify that the snapshot has been backed up. For example:
# ufsrestore tf /dev/rmt/0

How to Create an Incremental Backup of a UFS Snapshot (ufsdump)

Backing up a UFS snapshot incrementally means that only the les that have been modied since the last snapshot are backed up. Use the ufsdump command with the N option. This option species the le system device name to be inserted into the /etc/dumpdates le for tracking incremental dumps. The following ufsdump command species an embedded fssnap command to create an incremental backup of a le system.

Steps

1. Become superuser or assume an equivalent role.

Chapter 26 Using UFS Snapshots (Tasks) 433

2. Create an incremental backup of a UFS snapshot. For example:

# ufsdump 1ufN /dev/rmt/0 /dev/rdsk/c0t1d0s0 fssnap -F ufs -o raw,bs= /export/scratch,unlink /dev/rdsk/c0t1d0s0

In this example, the -o raw option is used to display the name of the raw device instead of the block device. By using this option, you make it easier to embed the fssnap command in commands (such as the ufsdump command) that require the raw device instead. 3. Verify that the snapshot has been backed up.
# ufsrestore ta /dev/rmt/0

How to Back Up a UFS Snapshot (tar)

If you are using the tar command to back up the snapshot, mount the snapshot before backing it up.

Steps

1. Become superuser or assume an equivalent role. 2. Create a mount point for the snapshot. For example:
# mkdir /backups/home.bkup

3. Mount the snapshot.

# mount -F ufs -o ro /dev/fssnap/1 /backups/home.bkup

4. Change to the mounted snapshot directory.

# cd /backups/home.bkup

5. Back up the snapshot with the tar command.

# tar cvf /dev/rmt/0 .

Restoring Data From a UFS Snapshot Backup

The backup created from the virtual device is essentially just a backup of what the original le system looked like when the snapshot was taken. When you restore a le system from the backup, restore as if you had taken the backup directly from the original le system. Such a restore uses the ufsrestore command. For information on using the ufsrestore command to restore a le or le system, see Chapter 27.

434

System Administration Guide: Devices and File Systems June 2005

CHAPTER

27

Restoring Files and File Systems (Tasks)

This chapter describes how to use the ufsrestore command to restore les and le systems that were backed up by using the ufsdump command. For information on the procedures associated with restoring les and le systems, see Restoring Files and File System Backups (Task Map) on page 435. For information about other commands you can use to archive, restore, copy, or move les and le systems, see Chapter 29. For information about backing up and restoring le systems, see Chapter 24.

Restoring Files and File System Backups (Task Map)

The following task map describes the procedures associated with restoring les and le systems.

Task

Description

For Instructions

Prepare to restore les and le systems. Determine which tapes to use.

Identify the le systems or les to be Preparing to Restore Files restored, the tape device, and how you and File Systems on page will restore them. 436 Refer to your backup tapes to nd the date of the last backup that contains the le or le system that you need to restore. How to Determine Which Tapes to Use on page 438

435

Task

Description

For Instructions

Restore les.

Choose one of the following restore methods: Restore les interactively Use this How to Restore Files method when you are unsure of the Interactively on page 439 le names because you can browse the media contents and select individual les and directories. Restore les noninteractively Use this method when you already know the few le names to be restored. Restore a le system Use this method when you get a new disk drive or as part of a recovery procedure. How to Restore Specic Files Noninteractively on page 441 How to Restore a Complete File System on page 443 How to Restore the root (/) and /usr File Systems on page 446

Restore the root (/) or /usr le systems.

Restoring the root (/) or /usr le systems involves booting the system from a local CD or the network.

Preparing to Restore Files and File Systems

The ufsrestore command copies les to disk, relative to the current working directory, from backups that were created by using the ufsdump command. You can use the ufsrestore command to reload an entire le system hierarchy from a level 0 dump and incremental dumps that follow it. You can also use this command to restore one or more single les from any backup tape. If you run the ufsrestore command as superuser, les are restored with their original owner, last modication time, and mode (permissions). Before you start to restore les or le systems, you need to know the following:
I I I I

The tapes (or diskettes) you need to restore from The raw device name on which you want to restore the le system The type of tape device you will use The device name (local or remote) for the tape device

436

System Administration Guide: Devices and File Systems June 2005

Determining the File System Name

If you have properly labeled your backup tapes, you should be able to use the le system name (/dev/rdsk/device-name) from the tape label. For more information, see How to Find File System Names on page 416.

Determining the Type of Tape Device You Need

You must use a tape device that is compatible with the backup media to restore the les. The format of the backup media determines which drive you must use to restore les. For example, if your backup media is 8-mm tape, you must use an 8-mm tape device to restore the les.

Determining the Tape Device Name

You might have specied the tape device name (/dev/rmt/n) as part of the backup tape label information. If you are using the same drive to restore a backup tape, you can use the device name from the label. For more information on media devices and device names, see Chapter 30.

Restoring Files and File Systems

When you back up les and directories, you save them relative to the le system in which they belong. When you restore les and directories, the ufsrestore command re-creates the le hierarchy in the current working directory. For example, les backed up from the /export/doc/books directory (where /export is the le system) are saved relative to /export. In other words, the book1 le in the books directory is saved as ./doc/books/book1 on the tape. Later on, if you restored the ./doc/books/book1 le to the /var/tmp directory, the le would be restored to /var/tmp/doc/books/book1. When you restore individual les and directories, you should restore them to a temporary location, such as the /var/tmp directory. After you verify the les, you can move them to their proper locations. However, you can restore individual les and directories to their original locations. If you do so, be sure you are not overwriting newer les with older versions from the backup tape. To avoid conicts with other users, you might want to create and change to a subdirectory, such as the/var/tmp/restore le, in which to restore the les.
Chapter 27 Restoring Files and File Systems (Tasks) 437

If you are restoring a hierarchy, you should restore the les to a temporary directory on the same le system where the les will reside. Then, you can use the mv command to move the entire hierarchy where it belongs after it is restored.
Note Do not restore les in the /tmp directory even temporarily. The /tmp directory is usually mounted as a TMPFS le system. TMPFS does not support UFS le system attributes such as ACLs.

M
Steps

How to Determine Which Tapes to Use

1. Ask the user for the approximate date the les to be restored were last modied. 2. Refer to your backup plan to nd the date of the last backup that contains the le or le system. To retrieve the most recent version of a le, work backward through the incremental backups from highest to lowest dump level and from most recent to least recent date, unless the user requests otherwise. 3. If you have online archive les, identify the correct media.
# ufsrestore ta archive-name ./path/lename ./path/lename

t a archive-name ./path/lename

Lists each le on the tape. Reads the table of contents from the online archive le instead of from the tape. Identies the online archive le name. Identies the le name or le names you are looking for on the online archive. If successful, the ufsrestore command prints out the inode number and le name. If unsuccessful, ufsrestore prints an error message.

For more information, see the ufsrestore(1M) man page. 4. Insert the media that contains the les to be restored in the drive and verify the correct media.
# ufsrestore tf /dev/rmt/n ./path/lename ./path/lename

Be sure to use the complete path for each lename. If a le is in the backup, its name and inode number are listed. Otherwise, a message states that the le is not on the volume. 5. If you have multiple backup les on the same tape, position the tape at the backup le you want to use.
# ufsrestore xfs /dev/rmt/n tape-number 438 System Administration Guide: Devices and File Systems June 2005

Example 271

Determining Which Tapes to Use

The following example shows how to check if the /etc/passwd le is in the online archive.
# ufsrestore ta /var/tmp/root.archive ./etc/passwd

The following example shows how to verify that the /etc/passwd le is on the backup tape.
# ufsrestore tf /dev/rmt/0 ./etc/passwd

M
Steps

How to Restore Files Interactively

1. Become superuser or assume an equivalent role. 2. (Optional) Write-protect the tapes for safety. 3. Insert the volume 1 tape into the tape drive. 4. Change to a directory that will be used to restore the les to temporarily.
# cd /var/tmp

5. Start the interactive restoration.

# ufsrestore if /dev/rmt/n

Some informational messages and the ufsrestore> prompt are displayed. 6. Create a list of les to be restored. a. List the contents of a directory.
ufsrestore> ls [directory-name]

b. Change to a directory.
ufsrestore> cd directory-name

c. Create a list of les and directories that you want to restore.

ufsrestore> add lenames

d. (Optional) Remove any directory or le from the list of les to be restored, if necessary.
ufsrestore> delete lename

7. (Optional) Display the le names as they are being restored.

ufsrestore> verbose Chapter 27 Restoring Files and File Systems (Tasks) 439

8. Restore the les.

ufsrestore> extract

The ufsrestore command asks you which volume number to use. 9. Type the volume number and press Return. If you have only one volume, type 1 and press Return.
Specify next volume #: 1

The les and directories in the list are extracted and restored to the current working directory. 10. To maintain the mode of the current directory, enter n at the set owner/mode prompt.
set owner/mode for .? [yn] n

You must wait while the ufsrestore command performs its nal cleanup. 11. Quit the ufsrestore program.
ufsrestore> quit

You then see the shell prompt. 12. Verify the restored les. a. List the restored les and directories.
# ls -l

A list of les and directories is displayed. b. Check the list to be sure that all the les and directories you specied in the list have been restored. 13. Move the les to the proper directories.
Example 272

Restoring Files Interactively

The following example shows how to extract the /etc/passwd and /etc/shadow les from the backup tape.
# cd /var/tmp # ufsrestore if /dev/rmt/0 ufsrestore> ls .: .: .sunw/ export/ Sources/ etools/ b/ home/ bin kernel/ dev/ lib/ devices/ lost+found/

net/ opt/ ptools/ pkg/ platform/ proc/

sbin/ scde/ set/ share/ shared/ src/

usr/ var/ vol/

440

System Administration Guide: Devices and File Systems June 2005

etc/ mnt/ rtools/ tmp/ ufsrestore> cd etc ufsrestore> add passwd shadow ufsrestore> verbose verbose mode on ufsrestore> extract Extract requested files You have not read any volumes yet. Unless you know which volume your file(s) are on you should start with the last volume and work towards the first. Specify next volume #: 1 extract file ./etc/shadow extract file ./etc/passwd Add links Set directory mode, owner, and times. set owner/mode for .? [yn] n ufsrestore> quit # cd etc # mv passwd /etc # mv shadow /etc # ls -l /etc

M
Steps

How to Restore Specic Files Noninteractively

1. Become superuser or assume an equivalent role. 2. (Optional) Write-protect the tape for safety. 3. Insert the volume 1 tape into the tape drive. 4. Change to a directory that will be used to restore les to temporarily.
# cd /var/tmp

5. Restore the le or les.

# ufsrestore xvf /dev/rmt/n lename

x v f /dev/rmt/n lename

Tells ufsrestore to copy specic les or directories in the lename argument. Displays the le names as they are restored. Identies the tape device name. Species one or more le names or directory names, separated by spaces. For example: ./export/home/user1/mail ./export/home/user2/mail.

Chapter 27 Restoring Files and File Systems (Tasks)

441

6. Type the volume number where les are located. Press Return.
Specify next volume #: 1

The le or les are restored to the current working directory. 7. To maintain the mode of the current directory, type n and press Return at the set owner/mode prompt.
set owner/mode for .? [yn] n

8. Verify the restored les. a. List the restored les and directories.
# ls -l

A list of les and directories is displayed. b. Check the list to be sure that all the les and directories you specied in the list have been restored. 9. Move the les to the proper directories.
Example 273

Restoring Specic Files Noninteractively

The following example shows how to noninteractively restore the passwd and shadow les to the /var/tmp directory.
# cd /var/tmp # ufsrestore xvf /dev/rmt/0 ./etc/passwd ./etc/shadow Verify volume and initialize maps Media block size is 126 Dump date: Wed Jul 28 16:13:52 2004 Dumped from: the epoch Level 0 dump of / on starbug:/dev/dsk/c0t0d0s0 Label: none Extract directories from tape Initialize symbol table. Extract requested files You have not read any volumes yet. Unless you know which volume your file(s) are on you should start with the last volume and work towards the first. Specify next volume #: 1 extract file ./etc/passwd extract file ./etc/shadow Add links Set directory mode, owner, and times. Specify next volume #:1 extract file ./etc/passwd extract file ./etc/shadow Add links Set directory mode, owner, and times. set owner/mode for .? [yn] n

442

System Administration Guide: Devices and File Systems June 2005

# # # #

cd mv mv ls

etc passwd /etc shadow /etc -l /etc

Example 274

Restoring Files From a Remote Tape Device

You can restore les from a remote tape drive by adding remote-host: to the front of the tape device name, when using the ufsrestore command. The following example shows how to restore les by using a remote tape drive /dev/rmt/0 on the system venus.
# ufsrestore xf venus:/dev/rmt/0 ./etc/hosts

How to Restore a Complete File System

Occasionally, a le system becomes so damaged that you must completely restore it. Typically, you need to restore a complete le system after a disk failure. You might need to replace the hardware before you can restore the software. For information on how to replace a disk, see SPARC: Adding a System Disk or a Secondary Disk (Task Map) on page 217 or x86: Adding a System Disk or a Secondary Disk (Task Map) on page 227. Full restoration of a le system such as /export/home can take a lot of time. If you have consistently backed up le systems, you can restore them to their state from the time of the last incremental backup.
Note You cannot use this procedure to restore the root (/) or /usr le systems. For instructions on restoring these le systems, see How to Restore the root (/) and /usr File Systems on page 446.

Steps

1. Become superuser or assume an equivalent role. 2. If necessary, unmount the le system.

# umount /dev/rdsk/device-name

Or:
# umount /le-system

3. Create the new le system.

# newfs /dev/rdsk/device-name

Chapter 27 Restoring Files and File Systems (Tasks)

443

You are asked if you want to construct a new le system on the raw device. Verify that the device-name is correct so that you dont destroy the wrong le system. For more information, see the newfs(1M) man page. 4. Conrm that the new le system should be created.
newfs: construct a new file system /dev/rdsk/cwtxdysz:(y/n)? y

The new le system is created. 5. Mount the new le system on a temporary mount point.
# mount /dev/dsk/device-name /mnt

6. Change to the mount point directory.

# cd /mnt

7. (Optional) Write-protect the tapes for safety. 8. Insert the rst volume of the level 0 tape into the tape drive. 9. Restore the les.
# ufsrestore rvf /dev/rmt/n

The dump level 0 backup is restored. If the backup required multiple tapes, you are prompted to load each tape in numeric order. 10. Remove the tape and load the next level tape in the drive. Always restore tapes starting with dump level 0 and continuing until you reach the highest dump level. 11. Repeat Step 8 through Step 10 for each dump level, from the lowest to the highest level. 12. Verify that the le system has been restored.
# ls

13. Remove the restoresymtable le.

# rm restoresymtable

The restoresymtable le that is created and used by the ufsrestore command to check-point the restore is removed. 14. Change to another directory.
# cd /

15. Unmount the newly restored le system.

# umount /mnt

444

System Administration Guide: Devices and File Systems June 2005

16. Remove the last tape and insert a new tape that is not write-protected in the tape drive. 17. Make a level 0 backup of the newly restored le system.
# ufsdump 0ucf /dev/rmt/n /dev/rdsk/device-name

A level 0 backup is performed. Always immediately do a full backup of a newly created le system because the ufsrestore command repositions the les and changes the inode allocation. 18. Mount the restored le system.
# mount /dev/dsk/device-name mount-point

The restored le system is mounted and available for use. 19. Verify that the restored and mounted le system is available.
# ls mount-point

Example 275

Restoring a Complete File System

The following example shows how to restore the /export/home le system.
# newfs /dev/rdsk/c0t0d0s7 newfs: /dev/rdsk/c0t0d0s7 last mounted as /export/home newfs: construct a new file system /dev/rdsk/c0t0d0s7: (y/n)? y 819314 sectors in 867 cylinders of 15 tracks, 63 sectors 400.1MB in 55 cyl groups (16 c/g, 7.38MB/g, 3584 i/g) super-block backups (for fsck -F ufs -o b=#) at: 32, 15216, 30400, 45584, 60768, 75952, 91136, 106320, 121504, 136688, 681264, 696448, 711632, 725792, 740976, 756160, 771344, 786528, 801712, 816896, # mount /dev/dsk/c0t0d0s7 /mnt # cd /mnt # ufsrestore rvf /dev/rmt/0 Verify volume and initialize maps Media block size is 126 Dump date: Thu Jul 29 10:14:00 2004 Dumped from: the epoch Level 0 dump of /export/home on starbug:/dev/dsk/c0t0d0s7 Label: none Begin level 0 restore Initialize symbol table. Extract directories from tape Calculate extraction list. Warning: ./lost+found: File exists Make node ./rimmer Make node ./rimmer/wdir Make node ./lister Make node ./pmorph Make node ./inquisitor Make node ./kryten Make node ./kryten/letters Chapter 27 Restoring Files and File Systems (Tasks) 445

Make node ./kryten/reports Extract new leaves. Check pointing the restore extract file ./rimmer/words extract file ./rimmer/words1 extract file ./rimmer/words2 extract file ./rimmer/words3 extract file ./rimmer/wdir/words extract file ./rimmer/wdir/words1 extract file ./rimmer/wdir/words2 extract file ./rimmer/wdir/words3 . . . Add links Set directory mode, owner, and times. Check the symbol table. Check pointing the restore # rm restoresymtable # cd / # umount /mnt # ufsdump 0ucf /dev/rmt/0 /export/home . . . # mount /dev/dsk/c0t0d0s7 /export/home # ls /export/home

M
Steps

How to Restore the root (/) and /usr File Systems

1. Become superuser or assume an equivalent role. 2. Add a new system disk to the system where the root (/) and /usr le systems will be restored. For a detailed description about adding a system disk, refer to SPARC: How to Connect a System Disk and Boot on page 218 or x86: How to Connect a System Disk and Boot on page 229. 3. Mount the new le system on a temporary mount point.
# mount /dev/dsk/device-name /mnt

4. Change to the /mnt directory.

# cd /mnt

5. (Optional) Write-protect the tapes for safety.

446

System Administration Guide: Devices and File Systems June 2005

6. Create the links for the tape device.

# tapes

7. Restore the root (/) le system.

# ufsrestore rvf /dev/rmt/n

The dump level 0 tape is restored. 8. Remove the tape and load the next level tape in the drive. Always restore tapes starting with dump level 0 and continuing from the lowest to highest dump level. 9. Continue restoring as needed.
# ufsrestore rvf /dev/rmt/n

The next level tape is restored. 10. Repeat Step 8 and Step 9 for each additional tape. 11. Verify that the le system has been restored.
# ls

12. Remove the restoresymtable le.

# rm restoresymtable

The restoresymtable le that is created and used by the ufsrestore command to check-point the restore is removed. 13. Change to the root (/) directory.
# cd /

14. Unmount the newly created le system.

# umount /mnt

15. Check the new le system.

# fsck /dev/rdsk/device-name

The restored le system is checked for consistency. 16. Create the boot blocks on the root partition.
# installboot /usr/platform/uname-i/lib/fs/ufs/bootblk /dev/rdsk/device-name

For more information, see the installboot(1M) man page. For an example of using the installboot command on a SPARC based system, see Example 276. For an example of using the installboot command on an x86 based system, see Example 277.

Chapter 27 Restoring Files and File Systems (Tasks)

447

17. Insert a new tape in the tape drive. 18. Back up the new le system.
# ufsdump 0uf /dev/rmt/n /dev/rdsk/device-name

A dump level 0 backup is performed. Always immediately do a full backup of a newly created le system because the ufsrestore command repositions the les and changes the inode allocation. 19. Repeat steps 5 through 16 for the /usr le system, if necessary. 20. Reboot the system.
# init 6

The system is rebooted.

Example 276

SPARC: Restoring the root (/) File System

This example shows how to restore the root (/) le system on a SPARC system. This example assumes that the system is booted from a local CD or from the network.
# mount /dev/dsk/c0t3d0s0 /mnt # cd /mnt # tapes # ufsrestore rvf /dev/rmt/0 # ls # rm restoresymtable # cd / # umount /mnt # fsck /dev/rdsk/c0t3d0s0 # installboot /usr/platform/sun4u/lib/fs/ufs/bootblk /dev/rdsk/c0t3d0s0 # ufsdump 0uf /dev/rmt/0 /dev/rdsk/c0t3d0s0 # init 6

Example 277

x86: Restoring the root (/) File System

This example shows how to restore the root (/) le system on an x86 system. This example assumes that the system is booted from a local CD or from the network.
# mount /dev/dsk/c0t3d0s0 /mnt # cd /mnt # tapes # ufsrestore rvf /dev/rmt/0 # ls # rm restoresymtable # cd / # umount /mnt # fsck /dev/rdsk/c0t3d0s0 # installboot /usr/platform/uname -i/lib/fs/ufs/pboot /usr/platform/ uname -i /lib/fs/ufs/bootblk /dev/rdsk/c0t3d0s2

448

System Administration Guide: Devices and File Systems June 2005

# ufsdump 0uf /dev/rmt/0 /dev/rdsk/c0t3d0s0 # init 6

Chapter 27 Restoring Files and File Systems (Tasks)

449

450

System Administration Guide: Devices and File Systems June 2005

CHAPTER

28

UFS Backup and Restore Commands (Reference)

This chapter contains reference information on the ufsdump and ufsrestore commands. This is a list of the information in this chapter.
I I I I

How the ufsdump Command Works on page 451 Specifying ufsdump Command Options and Arguments on page 456 The ufsdump Command and Security Issues on page 456 Specifying ufsrestore Options and Arguments on page 457

For overview information about performing backups, see Chapter 24. For information about backup tasks, see Chapter 25.

How the ufsdump Command Works

The ufsdump command makes two passes when it backs up a le system. On the rst pass, this command scans the raw device le for the le system and builds a table of directories and les in memory. Then, this command writes the table to the backup media. In the second pass, the ufsdump command goes through the inodes in numerical order, reading the le contents and writing the data to the backup media.

Determining Device Characteristics

The ufsdump command needs to know only an appropriate tape block size and how to detect the end of media.

451

Detecting the End of Media

The ufsdump command writes a sequence of xed-size records. When the ufsdump command receives notication that a record was only partially written, it assumes that it has reached the physical end of the media. This method works for most devices. If a device is not able to notify the ufsdump command that only a partial record has been written, a media error occurs as the ufsdump command tries to write another record.
Note DAT devices and 8-mm tape devices detect end-of-media. Cartridge tape devices and 1/2-inch tape devices do not detect end-of-media.

The ufsdump command automatically detects the end-of-media for most devices. Therefore, you do not usually need to use the -c, -d, -s, and -t options to perform multivolume backups. You need to use the end-of-media options when the ufsdump command does not understand the way the device detects the end-of-media. To ensure compatibility with the restore command, the size option can still force the ufsdump command to go to the next tape or diskette before reaching the end of the current tape or diskette.

Copying Data With the ufsdump Command

The ufsdump command copies data only from the raw disk slice. If the le system is still active, any data in memory buffers is probably not copied. The backup done by the ufsdump command does not copy free blocks, nor does it make an image of the disk slice. If symbolic links point to les on other slices, the link itself is copied.

Purpose of the /etc/dumpdates File

The ufsdump command, when used with the -u option, maintains and updates the /etc/dumpdates le. Each line in the /etc/dumpdates le shows the following information:
I I I

The le system backed up The dump level of the last backup The day, date, and time of the backup

For example:
# cat /etc/dumpdates /dev/rdsk/c0t0d0s0 /dev/rdsk/c0t0d0s7 452 0 Wed Jul 28 16:13:52 2004 0 Thu Jul 29 10:36:13 2004

System Administration Guide: Devices and File Systems June 2005

/dev/rdsk/c0t0d0s7

9 Thu Jul 29 10:37:12 2004

When you do an incremental backup, the ufsdump command checks the /etc/dumpdates le to nd the date of the most recent backup of the next lower dump level. Then, this command copies to the media all les that were modied since the date of that lower-level backup. After the backup is complete, a new information line, which describes the backup you just completed, replaces the information line for the previous backup at that level. Use the /etc/dumpdates le to verify that backups are being done. This verication is particularly important if you are having equipment problems. If a backup cannot be completed because of equipment failure, the backup is not recorded in the /etc/dumpdates le. If you need to restore an entire disk, check the /etc/dumpdates le for a list of the most recent dates and levels of backups so that you can determine which tapes you need to restore the entire le system.
Note The /etc/dumpdates le is a text le that can be edited. However, edit it only at your own risk. If you make changes to the le that do not match your archive tapes, you might be unable to nd the tapes (or les) you need.

Backup Device (dump-le) Argument

The dump-le argument (to the -f option) species the destination of the backup. The destination can be one of the following:
I I I I I

Local tape drive Local diskette drive Remote tape drive Remote diskette drive Standard output

Use this argument when the destination is not the default local tape drive /dev/rmt/0. If you use the -f option, then you must specify a value for the dump-le argument.
Note The dump-le argument can also point to a le on a local disk or on a remote disk. If done by mistake, this usage can ll up a le system.

Local Tape or Diskette Drive

Typically, the dump-le argument species a raw device le for a tape device or diskette. When the ufsdump command writes to an output device, it creates a single backup le that might span multiple tapes or diskettes.
Chapter 28 UFS Backup and Restore Commands (Reference) 453

You specify a tape device or a diskette on your system by using a device abbreviation. The rst device is always 0. For example, if you have a SCSI tape controller and one QIC-24 tape drive that uses medium-density formatting, use this device name: /dev/rmt/0m When you specify a tape device name, you can also type the letter n at the end of the name to indicate that the tape drive should not rewind after the backup is completed. For example: /dev/rmt/0mn Use the no-rewind option if you want to put more than one le onto the tape. If you run out of space during a backup, the tape does not rewind before the ufsdump command asks for a new tape. For a complete description of device-naming conventions, see Backup Device Names on page 484.

Remote Tape or Diskette Drive

You specify a remote tape device or a remote diskette by using the syntax host:device. The ufsdump command writes to the remote device when superuser on the local system has access to the remote system. If you usually run the ufsdump command as superuser, the name of the local system must be included in the /.rhosts le on the remote system. If you specify the device as user@host:device, the ufsdump command tries to access the device on the remote system as the specied user. In this case, the specied user must be included in the /.rhosts le on the remote system. Use the naming convention for the device that matches the operating system for the system on which the device resides, not the system from which you run the ufsdump command. If the drive is on a system that is running a previous SunOS release (for example, 4.1.1), use the SunOS 4.1 device name (for example, /dev/rst0). If the system is running Solaris software, use the SunOS 5.9 convention (for example, /dev/rmt/0).

Using Standard Output With the ufsdump Command

When you specify a dash (-) as the dump-le argument, the ufsdump command writes to standard output.
Note The -v option (verify) does not work when the dump-le argument is standard

output.

You can use the ufsdump and ufsrestore commands in a pipeline to copy a le system by writing to standard output with the ufsdump command and reading from standard input with the ufsrestore command. For example:
454 System Administration Guide: Devices and File Systems June 2005

# ufsdump 0f - /dev/rdsk/c0t0d0s7 | (cd /home; ufsrestore xf -)

Specifying Files to Back Up

You must always include lenames as the last argument on the command line. This argument species the source or contents of the backup. For a le system, specify the raw device le as follows: /dev/rdsk/c0t0d0s7 You can specify the le system by its mount point directory (for example, /export/home), as long as an entry for it exists in the /etc/vfstab le. For a complete description of device-naming conventions, see Backup Device Names on page 484. For individual les or directories, type one or more names separated by spaces.
Note When you use the ufsdump command to back up one or more directories or

les (rather than a complete le system), a level 0 backup is done. Incremental backups do not apply.

Specifying Tape Characteristics

If you do not specify any tape characteristics, the ufsdump command uses a set of defaults. You can specify the tape cartridge (c), density (d), size (s), and number of tracks (t). Note that you can specify the options in any order, as long as the arguments that follow match the order of the options.

Limitations of the ufsdump Command

The ufsdump command cannot do the following:
I

Automatically calculate the number of tapes or diskettes that are needed for backing up le systems. You can use the dry run mode (S option) to determine how much space is needed before actually backing up le systems. Provide built-in error checking to minimize problems when it backs up an active le system. Back up les that are remotely mounted from a server. Files on the server must be backed up on the server itself. Users are denied permission to run the ufsdump command on les they own that are located on a server.
Chapter 28 UFS Backup and Restore Commands (Reference) 455

Specifying ufsdump Command Options and Arguments

This section describes how to specify options and arguments for the ufsdump command. The syntax for the ufsdump command is as follows:
/usr/sbin/ufsdump options arguments lenames

options arguments lenames

Is a single string of one-letter option names. Identies option arguments and might consist of multiple strings. The option letters and their associated arguments must be in the same order. Identies the les to back up. These arguments must always come last, each separated by a space.

Default ufsdump Options

If you run the ufsdump command without any options, use this syntax:
# ufsdump lenames

The ufsdump command uses these options and arguments, by default:

ufsdump 9uf /dev/rmt/0 lenames

These options do a level 9 incremental backup to the default tape drive at its preferred density. For a description of the ufsdump options, see ufsdump(1M).

The ufsdump Command and Security Issues

If you are concerned about security, you should do the following:
I I

Require superuser access for the ufsdump command. Ensure superuser access entries are removed from /.rhosts les on clients and servers if you are doing centralized backups.

456

System Administration Guide: Devices and File Systems June 2005

For general information on security, see System Administration Guide: Security Services.

Specifying ufsrestore Options and Arguments

The syntax of the ufsrestore command is as follows:
/usr/sbin/ufsrestore options arguments lenames

options

Is a single string of one-letter option names. You must choose one and only one of these options: i, r, R, t, or x. For a description of the ufsrestore options, see ufsrestore(1M). Follows the option string with the arguments that match the options. The option letters and their associated arguments must be in the same order. Species the le or les to be restored as arguments to the x or t options. These arguments must always come last, separated by spaces.

arguments

lenames

Chapter 28 UFS Backup and Restore Commands (Reference)

457

458

System Administration Guide: Devices and File Systems June 2005

CHAPTER

29

Copying UFS Files and File Systems (Tasks)

This chapter describes how to copy UFS les and le systems to disk, tape, and diskettes by using various backup commands. This is a list of the step-by-step instructions in this chapter.
I I I I I I I I I I I I I I I

How to Copy a Disk (dd) on page 463 How to Copy Directories Between File Systems (cpio) on page 466 How to Copy Files to a Tape (tar) on page 468 How to List the Files on a Tape (tar) on page 469 How to Retrieve Files From a Tape (tar) on page 470 Copying Files to a Tape With the pax Command on page 471 How to Copy All Files in a Directory to a Tape (cpio) on page 472 How to List the Files on a Tape (cpio) on page 473 How to Retrieve All Files From a Tape (cpio) on page 474 How to Retrieve Specic Files From a Tape (cpio) on page 475 How to Copy Files to a Remote Tape Device (tar and dd) on page 476 How to Extract Files From a Remote Tape Device on page 477 How to Copy Files to a Single Formatted Diskette (tar) on page 479 How to List the Files on a Diskette (tar) on page 480 How to Retrieve Files From a Diskette (tar) on page 480

Commands for Copying File Systems

When you need to back up and restore complete le systems, use the ufsdump and ufsrestore commands described in Chapter 28. When you want to copy or move individual les, portions of le systems, or complete le systems, you can use the procedures described in this chapter instead of the ufsdump and ufsrestore commands. The following table describes when to use the various backup commands.
459

TABLE 291 Task

When to Use Various Backup Commands

Command For More Information

Back up le systems to tape. Create a le system snapshot. Restore le systems from tape. Transport les to other systems. Copy les or le systems between disks. Copy les to diskette.

ufsdump

How to Back Up a File System to Tape on page 418 Chapter 26 How to Restore a Complete File System on page 443 Copying Files and File Systems to Tape on page 467 How to Copy a Disk (dd) on page 463 How to Copy Files to a Single Formatted Diskette (tar) on page 479

fssnap ufsrestore pax, tar, or cpio dd

tar

The following table describes various backup and restore commands.

TABLE 292

Summary of Various Backup Commands

Aware of File System Boundaries? Supports Multiple Volume Backups? Physical or Logical Copy?

Command Name

volcopy tar cpio pax dd ufsdump/ufsrestore fssnap

Yes No No Yes Yes Yes N/A

Yes No Yes Yes No Yes N/A

Physical Logical Logical Logical Physical Logical Logical

The following table describes the advantages and disadvantages of some of these commands.

460

System Administration Guide: Devices and File Systems June 2005

TABLE 293 Command

Advantages and Disadvantages of tar, pax, and cpio Commands

Function Advantages
I

Disadvantages
I

tar

Use to copy les and directory subtrees to a single tape.

Available on most UNIX operating systems Public domain versions are readily available

Is not aware of le system boundaries Length of full path name cannot exceed 255 characters Cannot be used to create multiple tape volumes

pax

Use to copy les, special les, or le systems that require multiple tape volumes. Or, use when you want to copy les to and from POSIX-compliant systems. Use to copy les, special les, or le systems that require multiple tape volumes. Or, use when you want to copy les from systems running current Solaris releases systems to systems running SunOS 4.0/4.1 releases.

Better portability than the tar or cpio commands for POSIX-compliant systems Multiple vendor support

Same disadvantages as the tar command, except that the pax command can create multiple tape volumes.

cpio

Packs data onto tape The command syntax is more efficiently than more difficult than the tar or pax commands. the tar command Skips over any bad spots in a tape when restoring Provides options for writing les with different header formats, such as ( tar, ustar, crc, odc, bar), for portability between different system types Creates multiple tape volumes

The following sections describes step-by-step instructions and examples of how to use these commands.

Chapter 29 Copying UFS Files and File Systems (Tasks)

461

Copying File Systems Between Disks

Two commands are used to copy le systems between disks:
I I

volcopy dd

For more information about volcopy, see volcopy(1M). The next section describes how to use the dd command to copy le systems between disks.

Making a Literal File System Copy

The dd command makes a literal (block-level) copy of a complete UFS le system to another le system or to a tape. By default, the dd command copies standard input to standard output.
Note Do not use the dd command with variable-length tape drives without rst specifying an appropriate block size.

You can specify a device name in place of standard input or standard output, or both. In this example, the contents of the diskette are copied to a le in the /tmp directory:
$ dd < /floppy/floppy0 > /tmp/output.file 2400+0 records in 2400+0 records out

The dd command reports on the number of blocks it reads and writes. The number after the + is a count of the partial blocks that were copied. The default block size is 512 bytes. The dd command syntax is different from most other commands. Options are specied as keyword=value pairs, where keyword is the option you want to set and value is the argument for that option. For example, you can replace standard input and standard output with this syntax:
$ dd if=input-le of=output-le

To use the keyword=value pairs instead of the redirect symbols, you would type the following:
$ dd if=/floppy/floppy0 of=/tmp/output.file

462

System Administration Guide: Devices and File Systems June 2005

How to Copy a Disk (dd)

Keep the following key points in mind when you consider copying a disk:
I

Do not use this procedure to copy a disk that is under the control of a volume manager. The primary methods for copying UFS le system data from one disk or system to another disk or system is by using the ufsdump and ufsrestore commands. For more information on using these commands, see Chapter 24. You can clone systems by creating a ash archive and copying it to destination systems. For more information about creating a ash archive, see Solaris 10 Installation Guide: Solaris Flash Archives (Creation and Installation). If you are copying a disk with an EFI disk label, see Example 292.

If you are still considering copying a disk with the dd command keep the following cautions in mind:
I I I

Make sure that the source disk and destination disk have the same disk geometry. Check the UFS le systems on the disk to be copied with the fsck utility. Make sure the system is in single-user mode when copying a disk with the dd command.

Steps

1. Become superuser or assume an equivalent role. 2. (Optional) Create the /reconfigure le so that the system will recognize the destination disk to be added when it reboots, if necessary.
# touch /reconfigure

3. Shut down the system.

# init 0

4. Attach the destination disk to the system. 5. Boot the system.

ok boot -s

6. Copy the source disk to the destination disk.

# dd if=/dev/rdsk/device-name of=/dev/rdsk/device-name bs=block-size

if=/dev/rdsk/device-name of=/dev/rdsk/device-name bs=blocksize

Represents the overlap slice of the master disk device, usually slice 2. Represents the overlap slice of the destination disk device, usually slice 2. Identies the block size, such as 128 Kbytes or 256 Kbytes. A large block size decreases the time it
Chapter 29 Copying UFS Files and File Systems (Tasks) 463

takes to copy the disk. For more information, see dd(1M). 7. Check the new le system.
# fsck /dev/rdsk/device-name

8. Mount the destination disks root (/) le system.

# mount /dev/dsk/device-name /mnt

9. Change to the directory where the /etc/vfstab le is located.

# cd /mnt/etc

10. Using a text editor, edit the destination disks /etc/vfstab le to reference the correct device names. For example, change all instances of c0t3d0 to c0t1d0. 11. Change to the destination disks root (/) directory.
# cd /

12. Unmount the destination disks root (/) le system.

# umount /mnt

13. Shut down the system.

# init 0

14. Boot from the destination disk to single-user mode.

# boot diskn -s

Note The installboot command is not needed for the destination disk because the boot blocks are copied as part of the overlap slice.

15. Uncongure the destination disk.

# sys-unconfig

The system is shut down after it is uncongured. 16. Boot from the destination disk again and provide its system information, such as host name, time zone, and so forth.
# boot diskn

17. After the system is booted, log in as superuser to verify the system information.
hostname console login:

464

System Administration Guide: Devices and File Systems June 2005

Example 291

Copying a Disk With a VTOC Label (dd)

This example shows how to copy the master disk (with a VTOC label) /dev/rdsk/c0t0d0s2 to the destination disk /dev/rdsk/c0t2d0s2.
# touch /reconfigure # init 0 ok boot # dd if=/dev/rdsk/c0t0d0s2 of=/dev/rdsk/c0t2d0s2 bs=128k # fsck /dev/rdsk/c0t2d0s2 # mount /dev/dsk/c0t2d0s2 /mnt # cd /mnt/etc # vi vfstab (Modify entries for the new disk) # cd / # umount /mnt # init 0 # boot disk2 -s # sys-unconfig # boot disk2

Example 292

Copying a Disk with an EFI Label (dd)

In previous Solaris releases, slice 2 (s2) was used to represent the entire disk. On a disk with an EFI label, you must use a slightly different procedure to clone or copy disks larger than 1 terabyte so that the UUID of cloned disks is unique. If you do not create a new label for the cloned disk, other software products might corrupt data on EFI-labeled disks if they encounter duplicate UUIDs. For example: 1. Clone the disk with an EFI label. For example:
# dd if=/dev/rdsk/c0t0d0 of=/dev/rdsk/c0t2d0 bs=128k

2. Pipe the prtvtoc output of the disk to be copied to the fmthard command to create a new label for the cloned disk. For example:
# prtvtoc /dev/rdsk/c0t0d0 | fmthard -s - /dev/rdsk/c0t2d0

For more information about EFI disk labels, see Multiterabyte Disk Support With EFI Disk Label on page 175.

Chapter 29 Copying UFS Files and File Systems (Tasks)

465

Copying Directories Between File Systems (cpio Command)

You can use the cpio (copy in and out) command to copy individual les, groups of les, or complete le systems. This section describes how to use the cpio command to copy complete le systems. The cpio command is an archiving program that copies a list of les into a single, large output le. This command inserts headers between the individual les to facilitate recovery. You can use the cpio command to copy complete le systems to another slice, another system, or to a media device, such as a tape or diskette. Because the cpio command recognizes end-of-media and prompts you to insert another volume, it is the most effective command, other than ufsdump, to use to create archives that require multiple tapes or diskettes. With the cpio command, you frequently use the ls and find commands to list and select the les you want to copy, and then to pipe the output to the cpio command.

How to Copy Directories Between File Systems (cpio)

1. Become superuser or assume an equivalent role. 2. Change to the appropriate directory.
# cd lesystem1

Steps

3. Copy the directory tree from lesystem1 to lesystem2 by using a combination of the find and cpio commands.
# find . -print -depth | cpio -pdm lesystem2

. -print -depth -p -d

Starts in the current working directory. Prints the le names. Descends the directory hierarchy and prints le names from the bottom up. Creates a list of les. Creates directories as needed.

466

System Administration Guide: Devices and File Systems June 2005

-m

Sets the correct modication times on directories.

For more information, see cpio(1). The les from the directory name you specify are copied. The symbolic links are preserved. You might also specify the -u option. This option forces an unconditional copy. Otherwise, older les do not replace newer les. This option might be useful if you want an exact copy of a directory, and some of the les being copied might already exist in the target directory. 4. Verify that the copy was successful by displaying the contents of the destination directory.
# cd lesystem2 # ls

5. If appropriate, remove the source directory.

# rm -rf lesystem1

Example 293

Copying Directories Between File Systems (cpio)

# cd /data1 # find . -print -depth | cpio -pdm /data2 19013 blocks # cd /data2 # ls # rm -rf /data1

Copying Files and File Systems to Tape

You can use the tar, pax, and cpio commands to copy les and le systems to tape. The command that you choose depends on how much exibility and precision you require for the copy. Because all three commands use the raw device, you do not need to format or make a le system on tapes before you use them. The tape drive and device name that you use depend on the hardware conguration for each system. For more information about tape device names, see Choosing Which Media to Use on page 483.

Chapter 29 Copying UFS Files and File Systems (Tasks)

467

Copying Files to Tape (tar Command)

Here is information that you should know before you copy les to tape with the tar command:
I

Copying les to a tape with the -c option to the tar command destroys any les already on the tape at or beyond the current tape position. You can use le name substitution wildcards (? and *) as part of the le names that you specify when copying les. For example, to copy all documents with a .doc suffix, type *.doc as the le name argument. You cannot use le name substitution wildcards when you extract les from a tar archive.

M
Steps

How to Copy Files to a Tape (tar)

1. Change to the directory that contains the les you want to copy. 2. Insert a write-enabled tape into the tape drive. 3. Copy the les to tape.
$ tar cvf /dev/rmt/n lenames

c v f /dev/rmt/n lenames

Indicates that you want to create an archive. Displays the name of each le as it is archived. Indicates that the archive should be written to the specied device or le. Indicates the les and directories that you want to copy. Separate multiple les with spaces.

The le names that you specify are copied to the tape, overwriting any existing les on the tape. 4. Remove the tape from the drive. Write the names of the les on the tape label. 5. Verify that the les you copied are on the tape.
$ tar tvf /dev/rmt/n

For more information on listing les on a tar tape, see How to List the Files on a Tape (tar) on page 469.

468

System Administration Guide: Devices and File Systems June 2005

Example 294

Copying Files to a Tape (tar)

The following example shows how to copy three les to the tape in tape drive 0.
$ cd /export/home/kryten $ ls reports reportA reportB reportC $ tar cvf /dev/rmt/0 reports a reports/ 0 tape blocks a reports/reportA 59 tape blocks a reports/reportB 61 tape blocks a reports/reportC 63 tape blocks $ tar tvf /dev/rmt/0

M
Steps

How to List the Files on a Tape (tar)

1. Insert a tape into the tape drive. 2. Display the tape contents.
$ tar tvf /dev/rmt/n

t v f /dev/rmt/n
Example 295

Lists the table of contents for the les on the tape. Used with the t option, and provides detailed information about the les on the tape. Indicates the tape device.

Listing the Files on a Tape (tar)

The following example shows a listing of les on the tape in drive 0.
$ tar tvf /dev/rmt/0 drwxr-xr-x 0/1 0 -r--r--r-0/1 206663 -r--r--r-0/1 206663 -r--r--r-0/1 206663 Jul Jul Jul Jul 28 28 28 28 15:00 15:00 15:00 15:00 2004 2004 2004 2004 reports/ reports/reportA reports/reportB reports/reportC

Chapter 29 Copying UFS Files and File Systems (Tasks)

469

M
Steps

How to Retrieve Files From a Tape (tar)

1. Change to the directory where you want to put the les. 2. Insert the tape into the tape drive. 3. Retrieve the les from the tape.
$ tar xvf /dev/rmt/n [lenames]

Indicates that the les should be extracted from the specied archive le. All les on the tape in the specied drive are copied to the current directory. Displays the name of each le as it is retrieved. Indicates the tape device that contains the archive. Species a le to retrieve. Separate multiple les with spaces.

v f /dev/rmt/n lenames

For more information, see the tar(1) man page. 4. Verify that the les have been copied.
$ ls -l

Example 296

Retrieving Files on a Tape (tar)

The following example shows how to retrieve all the les from the tape in drive 0.
$ $ x x x x x $ cd /var/tmp tar xvf /dev/rmt/0 reports/, 0 bytes, reports/reportA, 0 reports/reportB, 0 reports/reportC, 0 reports/reportD, 0 ls -l

0 tape bytes, bytes, bytes, bytes,

blocks 0 tape 0 tape 0 tape 0 tape

blocks blocks blocks blocks

Troubleshooting The names of the les extracted from the tape must exactly match the names of the

les that are stored on the archive. If you have any doubts about the names or paths of the les, rst list the les on the tape. For instructions on listing the les on the tape, see How to List the Files on a Tape (tar) on page 469.

470

System Administration Guide: Devices and File Systems June 2005

Copying Files to a Tape With the pax Command

M
Steps

How to Copy Files to a Tape (pax)

1. Change to the directory that contains the les you want to copy. 2. Insert a write-enabled tape into the tape drive. 3. Copy the les to tape.
$ pax -w -f /dev/rmt/n lenames

-w -f /dev/rmt/n lenames

Enables the write mode. Identies the tape drive. Indicates the les and directories that you want to copy. Separate multiple les with spaces.

For more information, see the pax(1) man page. 4. Verify that the les have been copied to tape.
$ pax -f /dev/rmt/n

5. Remove the tape from the drive. Write the names of the les on the tape label.
Example 297

Copying Files to a Tape (pax)

The following example shows how to use the pax command to copy all the les in the current directory.
$ pax -w -f /dev/rmt/0 . $ pax -f /dev/rmt/0 filea fileb filec

Chapter 29 Copying UFS Files and File Systems (Tasks)

471

Copying Files to Tape With the cpio Command

M

How to Copy All Files in a Directory to a Tape (cpio)

1. Change to the directory that contains the les you want to copy. 2. Insert a write-enabled tape into the tape drive. 3. Copy the les to tape.
$ ls | cpio -oc > /dev/rmt/n

Steps

ls cpio -oc

Provides the cpio command with a list of le names. Species that the cpio command should operate in copy-out mode (-o) and write header information in ASCII character format (-c). These options ensure portability to other vendors systems. Species the output le.

> /dev/rmt/n

All les in the directory are copied to the tape in the drive you specify, overwriting any existing les on the tape. The total number of blocks that are copied is shown. 4. Verify that the les have been copied to tape.
$ cpio -civt < /dev/rmt/n

-c -i

Species that the cpio command should read les in ASCII character format. Species that the cpio command should operate in copy-in mode, even though the command is only listing les at this point. Displays the output in a format that is similar to the output from the ls -l command. Lists the table of contents for the les on the tape in the tape drive that you specify. Species the input le of an existing cpio archive.

-v -t < /dev/rmt/n

5. Remove the tape from the drive. Write the names of the les on the tape label.
472 System Administration Guide: Devices and File Systems June 2005

Example 298

Copying All Files in a Directory to a Tape (cpio)

The following example shows how to copy all of the les in the /export/home/kryten directory to the tape in tape drive 0.
$ cd /export/home/kryten $ ls | cpio -oc > /dev/rmt/0 16 blocks $ cpio -civt < /dev/rmt/0 -rw-r--r-1 root other -rw-r--r-1 root other -rw-r--r-1 root other drwxr-xr-x 2 root other drwxr-xr-x 2 root other 16 blocks $

0 0 0 0 0

Jul Jul Jul Jul Jul

28 28 28 28 28

14:59 14:59 14:59 14:59 15:00

2004, 2004, 2004, 2004, 2004,

filea fileb filec letters reports

How to List the Files on a Tape (cpio)

Note Listing the table of contents on a tape takes a long time because the cpio command must process the entire archive.

Steps

1. Insert an archive tape into the tape drive. 2. List the les on the tape.
$ cpio -civt < /dev/rmt/n

Example 299

Listing the Files on a Tape (cpio)

The following example shows how to list the les on the tape in drive 0.
$ cpio -civt < /dev/rmt/0 -rw-r--r-1 root other -rw-r--r-1 root other -rw-r--r-1 root other drwxr-xr-x 2 root other drwxr-xr-x 2 root other 16 blocks $ 0 0 0 0 0 Jul Jul Jul Jul Jul 28 28 28 28 28 14:59 14:59 14:59 14:59 15:00 2004, 2004, 2004, 2004, 2004, filea fileb filec letters reports

Chapter 29 Copying UFS Files and File Systems (Tasks)

473

How to Retrieve All Files From a Tape (cpio)

If the archive was created using relative path names, the input les are built as a directory within the current directory when you retrieve the les. If, however, the archive was created with absolute path names, the same absolute paths are used to re-create the le on your system.
Caution The use of absolute path names can be dangerous because you might overwrite existing les on your system.

Steps

1. Change to the directory where you want to put the les. 2. Insert the tape into the tape drive. 3. Extract all les from the tape.
$ cpio -icvd < /dev/rmt/n

-i -c -v -d < /dev/rmt/n

Extracts les from standard input. Species that the cpio command should read les in ASCII character format. Displays the les as they are retrieved in a format that is similar to the output from the ls command. Creates directories as needed. Species the output le.

4. Verify that the les were copied.

$ ls -l

Example 2910

Retrieving All Files From a Tape (cpio)

The following example shows how to retrieve all les from the tape in drive 0.
$ cd /var/tmp cpio -icvd < /dev/rmt/0 answers sc.directives tests 8 blocks $ ls -l

474

System Administration Guide: Devices and File Systems June 2005

M
Steps

How to Retrieve Specic Files From a Tape (cpio)

1. Change to the directory where you want to put the les. 2. Insert the tape into the tape drive. 3. Retrieve a subset of les from the tape.
$ cpio -icv "*le" < /dev/rmt/n

-i -c -v "*le"

Extracts les from standard input. Species that the cpio command should read headers in ASCII character format. Displays the les as they are retrieved in a format that is similar to the output from the ls command. Species that all les that match the pattern are copied to the current directory. You can specify multiple patterns, but each pattern must be enclosed in double quotation marks. Species the input le.

< /dev/rmt/n

For more information, see the cpio(1) man page. 4. Verify that the les were copied.
$ ls -l

Example 2911

Retrieving Specic Files From a Tape (cpio)

The following example shows how to retrieve all les with the chapter suffix from the tape in drive 0.
$ cd /home/smith/Book $ cpio -icv "*chapter" < /dev/rmt/0 Boot.chapter Directory.chapter Install.chapter Intro.chapter 31 blocks $ ls -l

Chapter 29 Copying UFS Files and File Systems (Tasks)

475

Copying Files to a Remote Tape Device

M

How to Copy Files to a Remote Tape Device (tar and dd)

1. The following prerequisites must be met to use a remote tape drive: a. The local host name and optionally, the user name of the user doing the copy, must appear in the remote systems /etc/hosts.equiv le. Or, the user doing the copy must have his or her home directory accessible on the remote machine, and have the local machine name in $HOME/.rhosts. For more information, see the hosts.equiv(4) man page. b. An entry for the remote system must be in the local systems /etc/inet/hosts le or in the name service hosts le. 2. To test whether you have the appropriate permission to execute a remote command, try the following:
$ rsh remotehost echo test

Steps

If test is echoed back to you, you have permission to execute remote commands. If Permission denied is echoed back to you, check your setup as described in Step 1. 3. Change to the directory where you want to put the les. 4. Insert the tape into the tape drive. 5. Copy the les to a remote tape drive.
$ tar cvf - lenames | rsh remote-host dd of=/dev/rmt/n obs=block-size

tar cf v - (Hyphen) lenames rsh | remote-host

Creates a tape archive, lists the les as they are archived, and species the tape device. Provides additional information about the tar le entries. Represents a placeholder for the tape device. Identies the les to be copied. Separate multiple les with spaces. Pipes the tar commands output to a remote shell.

476

System Administration Guide: Devices and File Systems June 2005

dd of= /dev/rmt/n obs=block-size

Represents the output device. Represents the blocking factor.

6. Remove the tape from the drive. Write the names of the les on the tape label.
Example 2912

Copying Files to a Remote Tape Drive (tar and dd)

# tar cvf - * | rsh mercury dd of=/dev/rmt/0 obs=126b a answers/ 0 tape blocks a answers/test129 1 tape blocks a sc.directives/ 0 tape blocks a sc.directives/sc.190089 1 tape blocks a tests/ 0 tape blocks a tests/test131 1 tape blocks 6+9 records in 0+1 records out

M
Steps

How to Extract Files From a Remote Tape Device

1. Insert the tape into the tape drive. 2. Change to a temporary directory.
$ cd /var/tmp

3. Extract the les from a remote tape device.

$ rsh remote-host dd if=/dev/rmt/n | tar xvBpf -

rsh remote-host dd if=/dev/rmt/n | tar xvBpf -

Indicates a remote shell that is started to extract the les from the tape device by using the dd command. Indicates the input device. Pipes the output of the dd command to the tar command, which is used to restore the les.

4. Verify that the les have been extracted.

$ ls -l

Example 2913

Extracting Files From a Remote Tape Drive

$ $ x x cd /var/tmp rsh mercury dd if=/dev/rmt/0 | tar xvBpf answers/, 0 bytes, 0 tape blocks answers/test129, 48 bytes, 1 tape blocks

Chapter 29 Copying UFS Files and File Systems (Tasks)

477

20+0 records in 20+0 records out x sc.directives/, 0 bytes, 0 tape blocks x sc.directives/sc.190089, 77 bytes, 1 tape blocks x tests/, 0 bytes, 0 tape blocks x tests/test131, 84 bytes, 1 tape blocks $ ls -l

Copying Files and File Systems to Diskette

Before you can copy les or le systems to diskette, you must format the diskette. For information on how to format a diskette, see Chapter 3. Use the tar command to copy UFS les to a single formatted diskette. Use the cpio command if you need to copy UFS les to multiple formatted diskettes. The cpio command recognizes end-of-media and prompts you to insert the next diskette.

What You Should Know When Copying Files to Diskettes

I

Copying les to a formatted diskette by using the tar -c command destroys any les that are already on the diskette. A diskette that contains a tar image is not mountable. If you need a multiple-volume interchange utility, use the cpio command. The tar command is only a single-volume utility.

I I

For more information, see tar(1).

478

System Administration Guide: Devices and File Systems June 2005

How to Copy Files to a Single Formatted Diskette (tar)

1. Change to the directory that contains the les you want to copy. 2. Insert a formatted diskette that is not write-protected into the drive. 3. Make the diskette available.
$ volcheck

Steps

4. Reformat the diskette, if necessary.

$ rmformat -U /dev/rdiskette Formatting will erase all the data on disk. Do you want to continue? (y/n)y

5. Copy the les to diskette.

$ tar cvf /vol/dev/aliases/floppy0 lenames

The le names that you specify are copied to the diskette, overwriting any existing les on the diskette. 6. Verify that the les were copied.
$ tar tvf /vol/dev/aliases/floppy0

For more information on listing les, see How to List the Files on a Diskette (tar) on page 480. 7. Remove the diskette from the drive. 8. Write the names of the les on the diskette label.
Example 2914

Copying Files to a Single Formatted Diskette (tar)

The following example shows how to copy les named evaluation* to a diskette.
$ cd /home/smith $ volcheck $ ls evaluation* evaluation.doc evaluation.doc.backup $ tar cvf /vol/dev/aliases/floppy0 evaluation* a evaluation.doc 86 blocks a evaluation.doc.backup 84 blocks $ tar tvf /vol/dev/aliases/floppy0

Chapter 29 Copying UFS Files and File Systems (Tasks)

479

M
Steps

How to List the Files on a Diskette (tar)

1. Insert a diskette into the drive. 2. Make the diskette available.
$ volcheck

3. List the les on a diskette.

$ tar tvf /vol/dev/aliases/floppy0

Example 2915

Listing the Files on a Diskette (tar)

The following example shows how to list the les on a diskette.
$ volcheck $ tar tvf /vol/dev/aliases/floppy0 rw-rw-rw-6693/10 44032 Jun 9 15:45 evaluation.doc rw-rw-rw-6693/10 43008 Jun 9 15:55 evaluation.doc.backup $

M
Steps

How to Retrieve Files From a Diskette (tar)

1. Change to the directory where you want to put the les. 2. Insert the diskette into the drive. 3. Make the diskette available.
$ volcheck

4. Retrieve les from the diskette.

$ tar xvf /vol/dev/aliases/floppy0

All les on the diskette are copied to the current directory. 5. Verify that the les have been retrieved.
$ ls -l

6. Remove the diskette from the drive.

Example 2916

Retrieving Files From a Diskette (tar)

The following example shows how to retrieve all the les from a diskette.

480

System Administration Guide: Devices and File Systems June 2005

$ $ $ x x $

cd /home/smith/Evaluations volcheck tar xvf /vol/dev/aliases/floppy0 evaluation.doc, 44032 bytes, 86 tape blocks evaluation.doc.backup, 43008 bytes, 84 tape blocks ls -l

The following example shows how to retrieve an individual le from a diskette. The le is extracted from the diskette and placed in the current working directory.
$ $ x $ volcheck tar xvf /vol/dev/aliases/floppy0 evaluation.doc evaluation.doc, 44032 bytes, 86 tape blocks ls -l

Archiving Files to Multiple Diskettes

If you are copying large les onto diskettes, you want to be prompted to replace a full diskette with another formatted diskette. The cpio command provides this capability. The cpio commands you use are the same that you would use to copy les to tape, except you would specify /vol/dev/aliases/floppy0 as the device instead of the tape device name. For information on how to use the cpio command, see How to Copy All Files in a Directory to a Tape (cpio) on page 472.

Chapter 29 Copying UFS Files and File Systems (Tasks)

481

482

System Administration Guide: Devices and File Systems June 2005

CHAPTER

30

Managing Tape Drives (Tasks)

This chapter describes how to manage tape drives in the Solaris Operating System (Solaris OS). This is a list of the step-by-step instructions in this chapter.
I I I

How to Display Tape Drive Status on page 486 Retensioning a Magnetic Tape Cartridge on page 487 Rewinding a Magnetic Tape Cartridge on page 488

This is a list of overview information in this chapter.

I I I I

Choosing Which Media to Use on page 483 Backup Device Names on page 484 Displaying Tape Drive Status on page 486 Guidelines for Drive Maintenance and Media Handling on page 488

Choosing Which Media to Use

You typically back up Solaris systems by using the following tape media:
I I I I

1/2-inch reel tape 1/4-inch streaming cartridge tape 8-mm cartridge tape 4-mm cartridge tape (DAT)

You can perform backups with diskettes, but doing so is time-consuming and cumbersome. The media that you choose depends on the availability of the equipment that supports it and of the media (usually tape) that you use to store the les. Although you must do the backup from a local system, you can write the les to a remote device.
483

The following table shows typical tape devices that are used for backing up le systems. The storage capacity for each device depends on the type of drive and the data being written to the tape.
TABLE 301

Media Storage Capacities

Storage Capacity

Backup Media

1/2-inch reel tape 2.5-Gbyte 1/4-inch cartridge (QIC) tape DDS3 4-mm cartridge tape (DAT) 14-Gbyte 8-mm cartridge tape DLT 7000 1/2-inch cartridge tape

140 Mbytes (6250 bpi) 2.5 Gbytes 1224 Gbytes 14 Gbytes 3570 Gbytes

Backup Device Names

You specify a tape or diskette to use for backup by supplying a logical device name. This name points to the subdirectory that contains the raw device le and includes the logical unit number of the drive. Tape drive naming conventions use a logical, not a physical, device name. The following table shows this naming convention.
TABLE 302

Basic Device Names for Backup Devices

Name

Device Type

Tape Diskette

/dev/rmt/n /vol/dev/rdiskette0/unlabeled

In general, you specify a tape device as shown in the following gure.

484

System Administration Guide: Devices and File Systems June 2005

/dev/rmt/XAbn Optional no-rewind n no-rewind omit for re-wind Berkeley compatability Optional density l low m medium h high u ultra c compressed Drive number (0-n) Raw magnetic tape device directory Devices directory

FIGURE 301

Tape Drive Device Names

If you dont specify the density, a tape drive typically writes at its preferred density. The preferred density usually means the highest density the tape drive supports. Most SCSI drives can automatically detect the density or format on the tape and read it accordingly. To determine the different densities that are supported for a drive, look at the /dev/rmt subdirectory. This subdirectory includes the set of tape device les that support different output densities for each tape. Also, a SCSI controller can have a maximum of seven SCSI tape drives.

Specifying the Rewind Option for a Tape Drive

Normally, you specify a tape drive by its logical unit number, which can run from 0 to n. The following table describes how to specify tape device names with a rewind or a no-rewind option.
TABLE 303

Specifying Rewind or No-Rewind for a Tape Drive

Use This Option

Drive and Rewind Value

First drive, rewind First drive, no rewind Second drive, rewind Second drive, no rewind

/dev/rmt/0 /dev/rmt/0n /dev/rmt/1 /dev/rmt/1n

Chapter 30 Managing Tape Drives (Tasks)

485

Specifying Different Densities for a Tape Drive

By default, the drive writes at its preferred density, which is usually the highest density the tape drive supports. If you do not specify a tape device, the command writes to drive number 0 at the default density the device supports. To transport a tape to a system whose tape drive supports only a certain density, specify a device name that writes at the desired density. The following table describes how to specify different densities for a tape drive.
TABLE 304

Specifying Different Densities for a Tape Drive

Use This Option

Drive, Density, and Rewind Value

First drive, low density, rewind First drive, low density, no rewind Second drive, medium density, rewind Second drive, medium density, no rewind

/dev/rmt/0l /dev/rmt/0ln /dev/rmt/1m /dev/rmt/1mn

The additional density values are shown in Backup Device Names on page 484.

Displaying Tape Drive Status

You can use the status option with the mt command to get status information about tape drives. The mt command reports information about any tape drives that are described in the /kernel/drv/st.conf le.

M
Steps

How to Display Tape Drive Status

1. Load a tape into the drive you want information about. 2. Display the tape drive status.
# mt -f /dev/rmt/n status

3. Repeat steps 12, substituting tape drive numbers 0, 1, 2, 3, and so on to display information about all available tape drives.
Example 301

Displaying Tape Drive Status

The following example shows the status for a QIC-150 tape drive (/dev/rmt/0):

486

System Administration Guide: Devices and File Systems June 2005

$ mt -f /dev/rmt/0 status Archive QIC-150 tape drive: sense key(0x0)= No Additional Sense file no= 0 block no= 0

residual= 0

retries= 0

The following example shows the status for an Exabyte tape drive (/dev/rmt/1):
$ mt -f /dev/rmt/1 status Exabyte EXB-8200 8mm tape drive: sense key(0x0)= NO Additional Sense residual= 0 file no= 0 block no= 0

retries= 0

The following example shows a quick way to poll a system and locate all of its tape drives:
$ for drive in 0 1 2 3 4 5 6 7 > do > mt -f /dev/rmt/$drive status > done Archive QIC-150 tape drive: sense key(0x0)= No Additional Sense file no= 0 block no= 0 /dev/rmt/1: No such file or directory /dev/rmt/2: No such file or directory /dev/rmt/3: No such file or directory /dev/rmt/4: No such file or directory /dev/rmt/5: No such file or directory /dev/rmt/6: No such file or directory /dev/rmt/7: No such file or directory $

residual= 0

retries= 0

Handling Magnetic Tape Cartridges

If errors occur when a tape is being read, you can retension the tape, clean the tape drive, and then try again.

Retensioning a Magnetic Tape Cartridge

Retension a magnetic tape cartridge with the mt command. For example:
$ mt -f /dev/rmt/1 retension $

Chapter 30 Managing Tape Drives (Tasks)

487

Note Do not retension non-QIC tape drives.

Rewinding a Magnetic Tape Cartridge

To rewind a magnetic tape cartridge, use the mt command. For example:
$ mt -f /dev/rmt/1 rewind $

Guidelines for Drive Maintenance and Media Handling

A backup tape that cannot be read is useless. So, periodically clean and check your tape drives to ensure correct operation. See your hardware manuals for instructions on procedures for cleaning a tape drive. You can check your tape hardware by doing either of the following:
I

Copying some les to the tape, reading the les back, and then comparing the original les with the copied les. Using the -v option of the ufsdump command to verify the contents of the media with the source le system. The le system must be unmounted or completely idle for the -v option to be effective.

Be aware that hardware can fail in ways that the system does not report. Always label your tapes after a backup. If you are using a backup strategy similar to the strategies suggested in Chapter 24, you should indicate on the label Tape A, Tape B, and so forth. This label should never change. Every time you do a backup, make another tape label that contains the following information:
I I I I I

The backup date The name of the machine and le system that is backed up The backup level The tape number (1 of n, if the backup spans multiple volumes) Any information specic to your site

Store your tapes in a dust-free safe location, away from magnetic equipment. Some sites store archived tapes in reproof cabinets at remote locations. You should create and maintain a log that tracks which media (tape volume) stores each job (backup) and the location of each backed-up le.
488 System Administration Guide: Devices and File Systems June 2005

Index
Numbers and Symbols
/export/home directory, 292 4.3 Tahoe le system, 285 9660 CD format, 34

B
backing up and restoring le systems commands for, 398 denition, 398 choosing le systems to, 399 full and incremental, denition, 401 preparing for (overview), 416-417 reasons for, 399 types of, 401 backup device names, 484-486 record of incremental, 453 backup schedules daily cumulative, weekly cumulative backups, 407 daily cumulative, weekly incremental backups, 408 daily incremental, weekly cumulative backups, 409 examples, 407, 413 for a server, 410-413 guidelines, 404 guidelines for, 403 using dump levels for, 406 bad block numbers, 377 bad inode number, 378 bad superblock, 386 block disk device interface denition, 171 when to use, 172 blocks bad, 377
489

A
accessing disk devices, 171 removable media (how to), 36 tape devices, 174 adding a disk (overview) SPARC, 218 x86, 228-240 a SCSI device to a SCSI bus (how to), 94 a USB camera (how to), 131 a USB mass storage device with vold running (how to), 130 a USB mass storage device without vold running (how to), 131 entry to /etc/vfstab le (how to), 321 PCI adapter card (how to), 102 swap to vfstab, 365 USB audio devices (how to), 143 VPPA communication service (how to), 164 allocated inodes, 376 archiving, les to multiple diskettes with cpio command (how to), 481 autoconguration process, 75 autofs, 300 automounting, and /home, 300

blocks (Continued) boot, 390 data, 391-392 directory data, 378 duplicate, 377 free, 392 indirect, 377 logical size, 393 regular data, 379 special inodes, 376 boot block, 390 BSD Fat Fast File system, 285 bus-oriented disk controllers, 173 bytes (number per inode), 395

C
CacheFS le systems (overview), 333 checking with fsck command (example of), 344 collecting CacheFS statistics (overview), 354 creating (how to), 335 creating a packing list (how to), 348 deleting (how to), 343 displaying information about (how to), 341 displaying packed les (example of), 347 displaying packed les (how to), 347 locating CacheFS log le, 356 mounting (how to), 336 packing with cachefspack command (how to), 346 packing with cachefspack command (overview), 345 parameters, 333 setting up CacheFS logging (how to), 355 stopping CacheFS logging, 357 troubleshooting cachefspack errors, 350 viewing CacheFS statistics, 358 cachefspack command how to use, 346 overview, 345 causes of le system damage, 372 cdrw command description, 59 restricting access to (how to), 62
490

cdrw command (Continued) writing data CDs and DVDs and audio CDs (overview), 61 CDs ISO 9660 format, 34 names, 32 UFS CDs SPARC vs. x86 format, 34 cfgadm PCI hot-plugging (overview), 86 SCSI hot-plugging (overview), 86 cfsadmin command, 335, 343 changing, primary USB audio device (how to), 144 character special inodes, 376 checking and repairing le systems, 380 CacheFS le systems (example of), 344 le system size, 374 format and type of inodes, 375 free blocks, 375 free inodes, 375 inode list for consistency, 374 clri command, 290 collecting, CacheFS statistics (overview), 354 conguring a SCSI controller (how to), 92 a SCSI device (how to), 92 a USB device (how to), 148 IB Port, HCA_SVC, or a VPPA device (how to), 160 IB pseudo device (how to), 161 IOC device (how to), 159 unidirectional or bidirectional CHAP authentication for iSCSI (how to), 246 connecting a SCSI controller (how to), 94 a USB device, logically (how to), 149 copying complete le systems (dd), 462 directories between le systems with cpio command (overview), 466 les to diskette (overview), 478 groups of les with cpio command (overview), 466 individual les with cpio command (overview), 466

System Administration Guide: Devices and File Systems June 2005

cpio command (overview), 466 copying directories between le systems (how to), 466 extract all les from tape (how to), 474 listing les on tape (how to), 473 creating a data CD or DVD le system (how to), 64 a full backup of UFS snapshot information (how to), 433 a packing list (how to), 348 a UFS snapshot (example of), 430 a UFS snapshot (how to), 429 le systems (overview), 304 loopback le system (overview), 311 swap le, 367 CTFS le system, 288 custom parameters for le systems, 392-396 cylinder group, 389-392

D
daily discrete backups, 406 damage to le systems, 372 data block, 379, 391-392 data directory blocks, 378 datadm command, 167 dd command (overview), 462 cloning disks (how to), 463 default le system for /tmp (TMPFS), 287 SunOS le system, 291 deleting CacheFS le systems (how to), 343 UFS snapshot information (example of), 432 detecting end of media cpio command, 466 ufsdump command, 452 determining le system types, 300 mounted le systems, 320 tape device name, 437 type of tape drive, 437 /dev/dsk directory, 171

/dev/rdsk directory, 171 devfsadm command, 170 device driver adding, 83 denition, 75 device names backup, 484-486 nding a le system name, 437 nding tape, 437 devices, accessing, 169 df command, 172, 290 dfstab le, conguring for shared local removable media (how to), 40 direct I/O, 295 directories copying between le systems with cpio command (overview), 466 inodes, 376 /proc, 288 /tmp, 287 unallocated blocks, 378 disconnect, a USB device, logically (how to), 149 disconnecting a SCSI controller (how to), 93 a USB device subtree, logically (how to), 150 disk adding to a (overview) x86, 228-240 automatic conguration of SCSI drives, 209 formatting a (overview), 188 repairing defective sectors, 211, 213 when to format (overview), 196 disk-based le systems, 284 disk controllers, 172 disk label creating (overview), 201 description, 189 disk slices denition, 182 determining which slices to use, 185 displaying information about (overview), 199-201 requirements for system congurations, 185 diskettes archiving les to multiple with cpio command (how to), 481
491

diskettes (Continued) loading with volume management (how to), 50 disks adding to a (overview) SPARC, 218 connecting a secondary disk (example of) SPARC, 223 creating disk slices and labeling a disk (example of) SPARC, 222 creating disk slices and labeling a disk (how to) SPARC, 220 determining if formatted (how to), 197 recovering a corrupted disk label (how to), 205 recovering a corrupted disk label (overview), 204 displaying disk slice information (overview), 199 IB communication services (how to), 163 InniBand device information (how to), 157 information about SCSI devices, 90 kernel IB clients of an HCA (how to), 161 packed les (example of), 347 packed les (how to), 347 PCI slot conguration information (how to), 100 removable media user (how to), 38 swap space, 366-367 system conguration information, 77, 80 USB bus information (how to), 147 USB device information (how to), 135 DOS, le system, 285 driver not attached message, 77 dump levels daily, incremental backups, 406 denition, 406 duplicate blocks, 377 DVD-ROM, 286 dynamic reconguration, InniBand devices, 156 dynamic reconguration (overview), 86

E
EFI label (overview), 176 comparison with VTOC label, 176 installing a system with, 178 restrictions of, 177 troubleshooting problems, 179 eject command, removable media (how to), 39 ejecting, removable media (how to), 39 enabling, uDAPL, 166 end-of-media detection cpio command, 466 ufsdump command, 452 /etc/dfs/dfstab le, conguring for shared removable media (how to), 40 /etc/dumpdates le, 452-453 /etc/rmmount.conf le, sharing removable media drives (how to), 41 extended fundamental types (UFS le system), 293

F
FDFS le system, 288 ff command, 290 FIFO inodes, 376 FIFOFS le system, 288 le system name, 437 le system table, virtual, 298 le systems /, 292 4.3 Tahoe, 285 BSD Fat Fast, 285 cached (overview), 333 checking and repairing, 380 checking size, 374 copying complete (dd), 462 creating (overview) loopback (LOFS), 311 CTFS, 288 custom parameters, 392-396 cylinder group struct, 389-392 damage to, 372 default SunOS, 291 description of administration commands, 290

492

System Administration Guide: Devices and File Systems June 2005

le systems (Continued) disk-based, 284 DOS, 285 /export/home, 292 FDFS, 288 FIFOFS, 288 nding types, 300 xing, 385 High Sierra, 285 ISO 9660, 285 large, 318 making available (overview), 315-320 manual pages for, 291 MNTFS, 292 mount table, 297 NAMEFS, 288 network-based, 286 OBJFS, 289 /opt, 292 PCFS, 285 preening, 384, 385 /proc, 292 process, (overview), 288 PROCFS, (overview), 288 pseudo, (overview), 286 reasons for inconsistencies, 373 sharing, 299 SPECFS, 289 stopping all processes accessing (how to), 328 SWAPFS, 289 TMPFS, 287 types of, 284 UFS, 285 UNIX, 285 /usr, 292 /var, 292 which to back up, 399 why you back up, 399 les archiving to multiple diskettes with cpio command (how to), 481 commands for copying to media (overview), 459 /etc/default/fs, 300 /etc/dfs/fstypes, 300 in the /proc directory, 288

les (Continued) retrieving from tape with tar command (how to), 470 sharing, 299 nding le system name, 437 tape device name, 437 type of le system, 300 xing inconsistent le systems, 385 format.dat le creating an entry (how to), 208 creating an entry (overview), 208 keywords, 271, 274 syntax rules, 271 format of inodes, 375 format utility (overview), 185 analyze menu, 268 automatic conguration of SCSI disk drives (how to), 211 automatic conguration of SCSI disk drives (overview), 209 creating a Solaris fdisk partition (how to), 233 creating disk slices and labeling disk (how to) SPARC, 220 x86, 238 defect menu, 269-270 determining if a disk is formatted (how to), 196 displaying disk slice information (example of), 200 entering command names (how to), 276 fdisk menu, 267 features and benets, 186 formatting a disk (example of), 198 guidelines for using, 187-188 identifying disks on a system (examples of), 196 identifying disks on a system (how to), 194 input to, 275, 277 labeling a disk example of, 202 main menu, 264 partition menu, 266, 267 recovering corrupted disk label (how to), 205
493

format utility (Continued) specifying block numbers (how to), 276 using help facility, 277 when to use, 186 formatting, a USB mass storage device without vold running (how to), 135 formatting a disk, (overview), 188 fragment size, 393-394 free blocks, 375, 392 free hog slice, See donor slice free inodes, 375 free space (minimum), 394 fsck command, 172, 290 checking free blocks, 375 free inodes, 375 inode list size, 374 superblock, 374 conditions to repair, 373 FSACTIVE state ag, 372 FSBAD state ag, 372 FSCLEAN state ag, 372 FSSTABLE state ag, 372 preening, 384 state ags, 372 syntax and options, 388 using interactively, 380 fsdb command, 290 fssnap command, creating a UFS snapshot (how to), 429 fstyp command, 290 fstypes le, 300 full backup (example of), 419, 421 denition, 401 fuser command nding if removable media is in use (how to), 38 killing processes accessing removable media (how to), 38

H
High Sierra le system, 285 /home (automounted), 300 hot-plugging (overview), 86 adding a SCSI device to a SCSI bus (how to), 94 adding PCI adapter card (how to), 102 conguring a SCSI controller (how to), 92 conguring a SCSI device (how to), 92 conguring a USB device (how to), 148 connecting a SCSI controller (how to), 94 disconnecting a SCSI controller with cfgadm command (how to), 93 logically connecting a USB device (how to), 149 logically disconnecting a USB device (how to), 149 logically disconnecting a USB device subtree (how to), 150 PCI devices (overview), 100 removing a SCSI device (how to), 97 removing PCI adapter card (how to), 101 replacing an identical SCSI device on a SCSI controller (how to), 95 unconguring a SCSI device (how to), 91 unconguring a USB device (how to), 148 HSFS, See High Sierra le system

I
I/O, direct, 295 identifying devices, 77 disks on a system (how to), 195 primary USB audio device (how to), 143 inconsistencies in le systems, 373 incorrect . and .. entries, 378 incremental backup, 401, 453 (example of), 420 indirect blocks, 377 InniBand devices adding a VPPA communication service (how to), 164 conguring an IB Port, HCA_SVC, or a VPPA device (how to), 160

G
grep command, 300

494

System Administration Guide: Devices and File Systems June 2005

InniBand devices (Continued) conguring an IB pseudo device (how to), 161 conguring an IOC device (how to), 159 displaying (how to), 157 displaying IB communication services (how to), 163 displaying kernel IB clients of an HCA (how to), 161 dynamic reconguration (overview), 156 overview, 153 removing an existing IB port, HCA_SVC, or a VPPA communication service (how to), 164 unconguring an IB Port, HCA_SVC, or a VPPA (how to), 160 unconguring an IB pseudo device (how to), 161 unconguring an IOC device (how to), 159 unconguring IB devices connected an HCA (how to), 162 updating an IOC conguration (how to), 165 updating the IP p_key tables (how to), 163 inode list size, 374 inode states, 376 inodes, 390-391 bad number, 378 block special, 376 character special, 376 checking format and type, 375 directory, 376 FIFO, 376 link count, 376 number of bytes per, 395 regular, 375 size, 377 symbolic link, 376 installboot command, 226, 240 installing a boot block (how to), SPARC, 226 iSCSI (overview), 241 conguring unidirectional or bidirectional CHAP authentication for (how to), 246 setting up (how to), 245 software and hardware requirements, 242 static and dynamic target discovery, 244

iscsiadm add command adding static or dynamic targets (example of), 249, 250 iscsiadm list, displaying ISCSI conguration information (example of), 251 iscsiadm modify command enabling CHAP (example of), 247 enabling or disabling static or dynamic targets (example of), 249 ISO 9660 le system, 285 ISO standards, 9660 CD format, 34

K
/kernel/drv directory, 76 killing all processes accessing a le system (how to), 328 processes accessing removable media (how to), 38

L
labelit command, 290 large les option, 318 level 0 backup, 406 link count of inodes, 376 loading diskettes with volume management (how to), 50 locating, CacheFS log le, 356 log (record of dumps), 452-453 logical block size, 393 logical device name denition, 170 disk, 171 tape, 174 logical device names, removable media, 174 loopback le system (LOFS) creating (overview), 311 mounting, 322 lost+found directory, 372

M
maintaining tape drives, 488
495

manual pages, for le systems, 291 media was found message, 50 memory storage (virtual), denition, 362 minimum free space, 394 mkfile command, 367, 368 mkfs command, 290, 304 mkisofs command, create a data CD or DVD le system (how to), 64 MNTFS le system, 292 mnttab le, 297 mount command, 172 mount point, denition, 296 mount table, 297 mountall command, 290 mounting a le system with /etc/vfstab, 322 a USB mass storage device with vold running (how to), 137 a USB mass storage device without vold running (how to), 138 all les in vfstab le, 322 le systems automatically, 300 loopback le systems (LOFS), 322 NFS le systems, 321 PCMCIA memory cards on other systems (example of), 44 remote removable media manually (example of), 44 removable media automatic mounting compared to, 28, 29 UFS le systems, 321 UFS le systems (how to) without large les, 324 mt command, 487

nfsd daemon (Continued) verifying if running, 40 no media was found message, 50

O
OBJFS le system, 289 /opt directory, 292 options, for ufsdump command, 456

P
parameters (le system), 392-396 partition (swap), denition, 362 passwd le, restoring from tape (example of), 442 PCFS le system, 285 PCI devices adding PCI adapter card (how to), 102 displaying PCI slot conguration information (how to), 100 removing PCI adapter card (how to), 101 troubleshooting PCI conguration problems, 103 PCMCIA memory cards accessing on other systems (example of), 44 mounting remotely (example of), 44 physical device name denition, 170 preening le systems, 384, 385 preparing for backing up (overview), 416-417 to restore les (overview), 436-437 to use a USB mass storage device without vold running (how to), 134 /proc directory, 288, 292 process le system (PROCFS), 288 PROCFS le system, (overview), 288 prtvtoc command, 172 (example of), 203 pseudo le systems, (overview), 286

N
NAMEFS le system, 288 ncheck command, 290 network-based le systems, 286 newfs command, 172, 304 NFS description, 299 server description, 299 vfstab entry for, 321 nfsd daemon starting, 41
496

R
raw disk device interface, 171, 172

System Administration Guide: Devices and File Systems June 2005

RCM script commands for, 105 overview, 104 reconguration boot, 209 SPARC example, 219 x86 example, 231 record of dumps, 452-453 incremental backup, 453 registering, service provider in the DAT static registry (how to), 167 regular inodes, 375 removable media accessing (examples of), 37 accessing (how to), 36 accessing media on other systems (example of), 44 ejecting (how to), 39 nding out if media is in use (how to), 38 killing processes accessing (how to), 38 mounting manual compared to automatic, 28, 29 mounting remote media (example of), 44 names, 32 removing a SCSI device (how to), 97 a swap le from use, 369 a USB mass storage device with vold running (how to), 132 a USB mass storage device without vold running (how to), 133 existing IB port, HCA_SVC, or a VPPA communication service (how to), 164 PCI adapter card (how to), 101 replacing, an identical SCSI device on a SCSI controller (how to), 95 resetting, a USB device (how to), 150 resolving, a failed SCSI uncongure operation (how to), 99 restoring bad superblock, 386 restoring le systems complete (example of), 445 complete (how to), 443 determining which tapes to use (how to), 438 preparing to (overview), 436-437 root and /usr (SPARC) (example of), 448 root or /usr (x86) (example of), 448

restoring le systems (Continued) type of tape drive, 437 restoring les interactive restore (example of), 440 non-interactive restore (example of), 442 restricting, removable media access (how to), 62 retrieving, les from tape with tar command (how to), 470 rmmount.conf le, sharing removable media drives (how to), 41 Rock Ridge extension (HSFS le system), 285 root (/) le system, 292

S
scheduling backups, 403 SCSI devices adding a SCSI device to a SCSI bus (how to), 94 conguring a SCSI controller (how to), 92 conguring a SCSI device (how to), 92 connecting a SCSI controller (how to), 94 disconnecting with cfgadm command (how to), 93 displaying information about (how to), 90 removing a SCSI device (how to), 97 replacing an identical SCSI device on a SCSI controller (how to), 95 resolving a failed SCSI uncongure operation (how to), 99 troubleshooting SCSI conguration problem, 97 unconguring a SCSI controller (how to), 91 SCSI disk drives, 209 SCSI tape drives, 485 secondary disk connecting to the system (how to) SPARC, 220 x86, 231 description, 184 setting up CacheFS logging, 355 iSCSI devices (how to), 245 share command, 299 making removable media available to other systems (how to), 40
497

shareall command, 299 sharing, les, 299 size checking le system, 374 fragment, 393-394 inode, 377 slice (denition), 182 Solaris fdisk partition, guidelines, 232-233 space optimization type, 395 SPARC based systems, UFS format, 34 SPECFS le system, 289 specifying a disk slice, 172 starting nfsd daemon, 41 volume management (how to), 36 state ag fsck, 372 UFS le systems, 293 stopping all processes for a le system (how to), 328 CacheFS logging, 357 killing processes accessing removable media (how to), 38 volume management (how to), 36 storage (virtual memory), denition, 362 storage capacities (media), 401, 484 structure of cylinder groups, 389-392 SunOS default le system, 291 superblock, 374, 386, 390 swap command, 367 swap le adding to vfstab, 365 creating, 367 displaying, 366-367 removing from use, 369 swap partition, denition, 362 swapadd command, 365 SWAPFS le system, 289 symbolic links, 376 syntax, fsck command, 388 sysdef command, 78 system disk connecting (how to) x86, 229 description, 184

T
tape capacity, 455 characteristics, 455 retrieving les from with tar command (how to), 470 sizes, 401, 484 storage capacities, 401, 484 tape devices (naming), 174 tape drive determining type for restore, 437 maintaining, 488 maximum SCSI, 485 rewind, 485-486 tar command (overview), 468 copying les to remote tape with dd command (how to), 476 listing les on diskette (how to), 480 listing les on tape (how to), 469 retrieving les from diskette (how to), 480 retrieving les from remote tape with dd command (how to), 477 retrieving les from tape (how to), 470 temporary le system (TMPFS), overview, 287 time (optimization type), 395 /tmp directory, 287, 292 TMPFS le system, overview, 287 troubleshooting a failed SCSI uncongure operation, 99 cachefspack errors, 350 EFI disk labels, 179 PCI conguration problems, 103 SCSI conguration problems, 97 USB audio device problems, 144 USB mass storage devices, 139 type of le systems, 284 type of inodes, 375

U
uDAPL (overview), 165 enabling (how to), 166 registering a service provider in the DAT static registry (how to), 167

498

System Administration Guide: Devices and File Systems June 2005

uDAPL (Continued) unregistering a service provider in the DAT static registry (how to), 168 updating the DAT static registry (how to), 167 UDF le system, 285 UFS CDs, SPARC compared to x86 formats, 34 UFS le system, 285, 293 extended fundamental types, 293 large le systems, 293 logging, 293 mounting, 321 mounting with /etc/vfstab, 322 mounting without large les (how to), 324 multiterabyte le systems, 293 state ags, 293 UFS logging, overview, 294 UFS snapshot creating (how to), 429 creating a full backup of (howto), 433 description, 427 ufsdump command end-of-media detection, 452 full backup (example of), 419, 421 how data is copied with, 452 how it works, 451-455 incremental backup (example of), 420 limitations, 455 options and arguments, 456 ufsdump command (overview), 418 ufsrestore command, 457 determining which tapes to use (how to), 438 preparing to use (overview), 436 umount command, 290 umountall command, 290 unallocated directory blocks, 378 unallocated inodes, 376 unconguring a SCSI controller (how to), 91 a USB device (how to), 148 IB devices connected an HCA (how to), 162 IB Port, HCA_SVC, or a VPPA Device (how to), 160 IB pseudo device (how to), 161 IOC device (how to), 159 UNIX le system, 285

unmounting a USB mass storage device with vold running (how to), 137 a USB mass storage device without vold running (how to), 138 unregistering, service provider in the DAT static registry (how to), 168 unsupported devices, 76 updating DAT static registry (how to), 167 IOC conguration (how to), 165 IP p_key tables (how to), 163 USB devices (overview), 114 acronyms, 114 adding a mass storage device with vold running (how to), 130 adding a mass storage device without vold running (how to), 131 adding a USB camera (how to), 131 adding audio devices (how to), 143 audio (overview), 141 changing the primary device (how to), 144 device ownership, 144 bus description, 115 bus-powered devices, 121 cables for, 124 cables for 2.0 devices, 121 composite device, 116 compound device, 116 conguring a USB device (how to), 148 connect a USB device (how to), 149 device classes, 116 device nodes, 118 diskette devices (overview), 129 displaying bus information (how to), 147 displaying USB device information (how to), 135 drivers, 116 formatting a mass storage device without vold running (how to), 135 host controller and root hub, 123 hot-plugging (overview), 130 identifying primary audio device (how to), 143 keyboards and mouse devices, 121
499

USB devices (Continued) logically disconnecting a USB device (how to), 149 logically disconnecting a USB device subtree (how to), 150 mounting mass storage with vold running (how to), 137 mounting mass storage without vold running (how to), 138 names of, 116 overview of 2.0 devices, 119 physical device hierarchy, 115 power management, 124 preparing to use a mass storage device without vold running (how to), 134 removable mass storage (overview), 127 removing a mass storage device with vold running (how to), 132 removing a mass storage device without vold running (how to), 133 resetting a USB device (how to), 150 Solaris USB Architecture (USBA), 118 troubleshooting audio device problems, 144 troubleshooting tips for mass storage devices, 139 unconguring a device (how to), 148 unmounting mass storage with vold running (how to), 137 unmounting mass storage without vold running (how to), 138 using non-compliant mass storage devices (overview), 129 wheel mouse support, 122 /usr le system, 292

virtual memory storage, denition, 362 volcopy command, 290 volmgt start command, 36 volume management benets, 27 loading diskettes (how to), 50 manual compared to automatic mounting, 28, 29 restarting (how to), 36 stopping (how to), 36

W
writing, data CDs and DVDs and audio CDs (overview), 61

X
x86 based systems, UFS format, 34

V
/var directory, 292 verifying, nfsd daemon is running, 40 vfstab le, 300, 365 adding entries to (how to), 321 adding swap to, 365 default, 298 entry for LOFS, 313 mounting all les, 322 viewing, CacheFS statistics, 358 virtual le system table, 298
500 System Administration Guide: Devices and File Systems June 2005