Front cover

Solaris to Linux Migration:

A Guide for System Administrators

A comprehensive reference for a quick transition Presents a task-based grouping of differences between the operating system environments Additional content about how to optimize Linux on IBM Eserver platforms

Mark Brown Chuck Davis William Dy Paul Ionescu Jeff Richardson Kurt Taylor Robbie Williamson

ibm.com/redbooks

International Technical Support Organization Solaris to Linux Migration: A Guide for System Administrators February 2006

SG24-7186-00

Note: Before using this information and the product it supports, read the information in Notices on page xiii.

First Edition (February 2006)

Copyright International Business Machines Corporation 2006. All rights reserved. Note to U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.

Contents
Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv The team that wrote this redbook. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviii Become a published author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviii Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix Part 1. Background and planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Chapter 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1 What is Linux?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2 Linux kernels, distributions, and distributors . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2.1 The Linux kernel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.2.2 The Linux operating system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.2.3 Linux distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.3 How this book is organized . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.4 Linux administration overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.4.1 Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.4.2 Utilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.4.3 Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.4.4 Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.5 Reference materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Chapter 2. Planning for migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.1 Assemble the stakeholders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.2 Set objectives and define scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.3 Assess workload and environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.3.1 Workload types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.3.2 ISV applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.3.3 Custom application porting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.3.4 Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.3.5 Integration with existing network services . . . . . . . . . . . . . . . . . . . . . 16 2.3.6 Other important infrastructure requirements . . . . . . . . . . . . . . . . . . . 17 2.3.7 Special hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.4 Assess skill requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.5 Build a plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Copyright IBM Corp. 2006. All rights reserved.

iii

Part 2. System administration differences guide. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Chapter 3. Operating system installation . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.1 Basic system installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.1.1 Graphical or text installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.1.2 Other installation methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.1.3 System bundle options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.1.4 Linux LVM consideration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.1.5 Linux firewall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.2 Advanced installation and automation. . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.2.1 Installing using a serial interface as input/output device . . . . . . . . . . 29 3.2.2 Configuring a serial console device . . . . . . . . . . . . . . . . . . . . . . . . . 30 3.2.3 Installing using remote display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.2.4 Solaris Jumpstart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3.2.5 Red Hat Kickstart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3.2.6 SUSE AutoYaST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 3.2.7 The PXE protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.3 Network-based install: Solaris and Linux heterogeneous environments . . 47 3.3.1 Solaris system serving Linux network installation . . . . . . . . . . . . . . . 47 3.3.2 Linux system serving Solaris network installation . . . . . . . . . . . . . . . 50 Chapter 4. Disks and file systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 4.1 Disks and disk partitioning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 4.1.1 Disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 4.1.2 Disk partitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 4.2 Disk-based file system management . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 4.3 Virtual file systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4.4 Network File System (NFS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 4.5 Swap file systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 4.6 File system journaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4.7 File system organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 4.8 AutoFSCK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 4.9 AutoFS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 4.10 Solaris Volume Manager to Linux LVM . . . . . . . . . . . . . . . . . . . . . . . . . . 65 4.11 VERITAS VxVM and VxFS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Chapter 5. Software management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 5.1 Packages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 5.1.1 Package management in Solaris . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 5.1.2 Package management in Linux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 5.2 Patching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 5.2.1 Patching in Solaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 5.2.2 Patching in Linux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 5.3 Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

iv

Solaris to Linux Migration: A Guide for System Administrators

5.3.1 Dependency management in Solaris . . . . . . . . . . . . . . . . . . . . . . . . 80 5.3.2 Dependency management in Linux. . . . . . . . . . . . . . . . . . . . . . . . . . 80 5.4 Package distribution methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 5.5 Automated software management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 5.5.1 Automated software management in Solaris . . . . . . . . . . . . . . . . . . 81 5.5.2 Automated software management in Linux . . . . . . . . . . . . . . . . . . . . 81 5.6 Activating fixes after updating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 5.6.1 Patch activation in Solaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 5.6.2 Patch activation in Linux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 5.7 Compiling patches in Linux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Chapter 6. Device management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 6.1 Device access and configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 6.1.1 Device naming and access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 6.1.2 Displaying device configuration information . . . . . . . . . . . . . . . . . . . 89 6.1.3 Adding a device. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 6.1.4 Hot-plug devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 6.2 Removable media devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 6.2.1 Supported types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 6.2.2 Managing and accessing removable media . . . . . . . . . . . . . . . . . . . 94 6.2.3 Formatting removable media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 6.2.4 CDs and DVDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 6.2.5 Tape drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 6.3 Terminals and modems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 6.3.1 Terminal setup and initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 6.3.2 Modem setup tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 6.3.3 Serial port management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 6.3.4 Port monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 6.4 Distribution-based device management tools . . . . . . . . . . . . . . . . . . . . . 105 Chapter 7. Network services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 7.1 IPv4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 7.2 IPv6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 7.3 Mixed IPv4 and IPv6 networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 7.4 Static and dynamic routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 7.5 IPSec and IKE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 7.6 Network multipath . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 7.7 Network trunking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 7.8 IP network services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 7.8.1 inetd-based versus xinetd-based network services . . . . . . . . . . . . 117 7.8.2 DHCP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 7.8.3 DNS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 7.8.4 NTP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

Contents

7.8.5 LDAP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 7.8.6 NIS and NIS+ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 7.8.7 NFS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 7.8.8 Web proxy and cache servers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 7.8.9 Mail services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 7.9 TCP wrappers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 7.10 IP stateful firewalling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Chapter 8. Boot and system initialization . . . . . . . . . . . . . . . . . . . . . . . . . 125 8.1 Booting a system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 8.1.1 Booting types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 8.1.2 Booting sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 8.1.3 Booting process overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 8.2 Run levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 8.3 Boot configuration files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 8.3.1 /etc/inittab file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 8.3.2 Run control files (rc files). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 8.3.3 Disabling rc scripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 8.4 Shutting down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 8.5 Network booting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Chapter 9. Managing system resources . . . . . . . . . . . . . . . . . . . . . . . . . . 139 9.1 Displaying system information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 9.2 Resource management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 9.3 Starting and stopping system services . . . . . . . . . . . . . . . . . . . . . . . . . . 141 9.3.1 Solaris system services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 9.3.2 Linux system services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 9.4 Scheduling and cron services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 9.5 Quotas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 9.6 Process accounting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 9.6.1 Solaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 9.6.2 Linux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 9.7 Remote system management services . . . . . . . . . . . . . . . . . . . . . . . . . . 147 9.7.1 Web-Based Enterprise Management (WBEM) . . . . . . . . . . . . . . . . 147 9.7.2 Simple Network Management Protocol (SNMP) . . . . . . . . . . . . . . . 149 Chapter 10. Printing services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 10.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 10.2 Linux CUPS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 10.3 Client and server setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 10.4 Printer management using the lp commands . . . . . . . . . . . . . . . . . . . . 154 10.5 Printer types and CUPS support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 Chapter 11. Users and groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

vi

Solaris to Linux Migration: A Guide for System Administrators

11.1 Basic user administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 11.2 Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 11.2.1 Adding user accounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 11.2.2 Changing user account information . . . . . . . . . . . . . . . . . . . . . . . 161 11.2.3 Removing user accounts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 11.2.4 Home directory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 11.3 Directory services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 11.4 NIS and NIS+ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 11.5 User ID and group ID differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 Chapter 12. Monitoring and performance . . . . . . . . . . . . . . . . . . . . . . . . . 173 12.1 Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 12.2 Real and virtual memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 12.3 Physical media, software RAID, LVM, and file systems . . . . . . . . . . . . 176 12.3.1 Logical volume groups and logical volumes . . . . . . . . . . . . . . . . . 177 12.3.2 File systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 12.4 Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 12.5 System and user processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Chapter 13. Backup and restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 13.1 Common backup tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 13.2 Compression tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 13.3 File system backup and restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 13.3.1 Solaris overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 13.3.2 Linux overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 13.4 File system snapshots. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 13.5 AMANDA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Chapter 14. Security and hardening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 14.1 Patching and security updates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 14.2 Security hardening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 14.2.1 Hardening tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 14.2.2 Auditing: ASET/BSM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 14.3 Securing and removing services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 14.3.1 inetd/xinetd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 14.3.2 TCP wrappers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190 14.3.3 FTP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 14.4 Kernel tuning for security. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 14.5 Logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 14.6 Access control lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 14.7 PAM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 14.7.1 PAM module types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 14.7.2 Limiting superuser login to secure terminals . . . . . . . . . . . . . . . . . 196 14.7.3 Restricting user login. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197

Contents

vii

14.8 umask . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 14.9 SSH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 14.10 IPSec and IKE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 14.11 Kerberos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 14.12 Warning banners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 14.13 Firewalls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 Chapter 15. Linux high availability overview . . . . . . . . . . . . . . . . . . . . . . 205 15.1 Introduction to Linux-HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 15.1.1 Migration scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 15.1.2 Some features of Linux-HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 15.2 Linux-HA services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 15.2.1 Heartbeat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 15.2.2 Cluster Resource Manager (CRM) . . . . . . . . . . . . . . . . . . . . . . . . 207 15.2.3 Consensus Cluster Membership (CCM) . . . . . . . . . . . . . . . . . . . . 208 15.2.4 Local Resource Manager (LRM) . . . . . . . . . . . . . . . . . . . . . . . . . . 208 15.2.5 STONITH daemon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 15.3 Linux migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 15.4 Cluster environment considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 15.4.1 Data sharing arrangements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 15.4.2 IBM ServeRAID. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 Chapter 16. Shell scripting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 16.1 Overview of the shell environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 16.1.1 Solaris shell environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 16.1.2 Linux shell environments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 16.2 Public Domain Korn shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 16.3 Moving from ksh to bash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 Chapter 17. Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 17.1 Troubleshooting the booting process . . . . . . . . . . . . . . . . . . . . . . . . . . 216 17.2 Core files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 17.3 Crash dumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 17.4 Logs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 17.5 Permissions: File access problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 17.5.1 Problem: Command not found . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 17.5.2 Problem: File access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 17.6 Printing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 17.6.1 Troubleshooting remote printer connectivity . . . . . . . . . . . . . . . . . 224 17.6.2 Troubleshooting local printers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 17.7 File systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 17.7.1 Remote file systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 17.7.2 Software RAID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 17.7.3 Logical volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

viii

Solaris to Linux Migration: A Guide for System Administrators

17.8 Packages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 17.9 root password recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 17.10 Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 17.11 System and user processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 17.12 Diagnostic and debugging tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 Part 3. IBM eServer platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 Chapter 18. IBM eServer xSeries hardware platform specifics. . . . . . . . 237 18.1 Installation considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 18.1.1 xSeries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 18.1.2 BladeCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 18.2 Hardware and device differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 18.2.1 xSeries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240 18.2.2 BladeCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 18.3 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 Chapter 19. IBM POWER technology hardware platform specifics . . . . 247 19.1 Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 19.1.1 IBM eServer i5 and eServer p5 . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 19.1.2 IBM eServer BladeCenter JS20 . . . . . . . . . . . . . . . . . . . . . . . . . . 254 19.2 Booting and system initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 19.2.1 IBM eServer i5 and eServer p5 . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 19.2.2 eServer BladeCenter JS20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 19.3 Linux installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 19.3.1 eServer i5 and eServer p5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 19.3.2 eServer BladeCenter JS20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 19.3.3 Red Hat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 19.3.4 Novell SUSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 19.3.5 YaBoot OpenFirmware boot loader. . . . . . . . . . . . . . . . . . . . . . . . 283 19.4 eServer i5 and eServer p5 virtualization . . . . . . . . . . . . . . . . . . . . . . . . 284 19.4.1 Create a virtual I/O server partition profile . . . . . . . . . . . . . . . . . . 285 19.4.2 Create a client partition profile . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 19.4.3 Configure Linux on a virtual I/O server . . . . . . . . . . . . . . . . . . . . . 295 19.4.4 Configure Linux on a virtual client partition . . . . . . . . . . . . . . . . . . 302 19.5 POWER technology platform service and productivity tools . . . . . . . . . 303 19.5.1 Red Hat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 19.5.2 Novell SUSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 19.6 References and further readings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 20.1 Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309 20.1.1 Hardware resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310

Contents

ix

20.1.2 Software resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310 20.1.3 Networking resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 20.1.4 Linux distributions and z/VM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 20.2 S/390 and zSeries overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312 20.2.1 Processing units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312 20.2.2 Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312 20.2.3 The channel subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 20.2.4 Tape drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 20.2.5 Disk drives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 20.2.6 Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316 20.2.7 Printers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 20.2.8 Logical partition concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 20.2.9 Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320 20.3 Installation methods and techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . 320 20.3.1 Preparing the z/VM guest resources . . . . . . . . . . . . . . . . . . . . . . . 321 20.3.2 Server farms or cloned environments . . . . . . . . . . . . . . . . . . . . . . 321 20.4 Booting, system initialization, and shutdown . . . . . . . . . . . . . . . . . . . . . 326 20.5 Device management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 20.5.1 Linux reconfigurability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 20.5.2 DASD hot-plug example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328 20.6 Performance monitoring and tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 20.6.1 Performance monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 20.6.2 Performance tuning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333 20.7 Troubleshooting and diagnostics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334 20.7.1 z/VM troubleshooting and diagnostics . . . . . . . . . . . . . . . . . . . . . 334 20.7.2 Linux troubleshooting and diagnostics . . . . . . . . . . . . . . . . . . . . . 335 20.8 S/390 and zSeries-specific Linux commands . . . . . . . . . . . . . . . . . . . . 335 20.8.1 Packages specifically for the mainframe . . . . . . . . . . . . . . . . . . . . 336 20.8.2 Commands specifically for the mainframe . . . . . . . . . . . . . . . . . . 336 20.9 High availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338 20.9.1 Hardware HA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338 20.9.2 z/VM HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338 20.9.3 Automating Linux guest boot and shutdown . . . . . . . . . . . . . . . . . 339 20.9.4 Linux-HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340 20.9.5 Backup and recovery options . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340 20.10 Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341 20.11 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342 Part 4. Appendixes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343 Appendix A. Tasks reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345 Packaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346 Installing and upgrading tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346

Solaris to Linux Migration: A Guide for System Administrators

Booting and shutting down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348 User management tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351 Device management and configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352 Network management and configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353 NFS management and configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354 Managing system resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354 Managing system services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355 Managing scheduling and cron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356 Managing quota . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356 Managing process accounting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357 Printer management and configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357 Disk and file system management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358 Swap management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359 Logical volume management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359 General troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362 Network troubleshooting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363 Appendix B. Commands and configuration files reference . . . . . . . . . . 365 Configuration and other files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366 Comparable commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367 Common Solaris and Linux commands . . . . . . . . . . . . . . . . . . . . . . . . . . 369 Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter) . . . . . . . . . . . . . . . . . . . . . . . 371 Table of contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372 Chapter 1 Porting Project Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . 373 Software Application Business Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373 The Porting Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374 Defining Project Scope and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378 Estimating. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380 Creating a Porting Project Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385 Porting Process from a Business Perspective . . . . . . . . . . . . . . . . . . . . . . . . 387 Annotated Sample Technical Questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . 387 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394 Appendix D. Example: System information gathering script . . . . . . . . . 397 Appendix E. Additional material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401 Locating the Web material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402 Using the Web material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402 Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403 IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403 Other publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404

Contents

xi

Online resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405 How to get IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411 Help from IBM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413

xii

Solaris to Linux Migration: A Guide for System Administrators

Notices
This information was developed for products and services offered in the U.S.A. IBM may not offer the products, services, or features discussed in this document in other countries. Consult your local IBM representative for information on the products and services currently available in your area. Any reference to an IBM product, program, or service is not intended to state or imply that only that IBM product, program, or service may be used. Any functionally equivalent product, program, or service that does not infringe any IBM intellectual property right may be used instead. However, it is the user's responsibility to evaluate and verify the operation of any non-IBM product, program, or service. IBM may have patents or pending patent applications covering subject matter described in this document. The furnishing of this document does not give you any license to these patents. You can send license inquiries, in writing, to: IBM Director of Licensing, IBM Corporation, North Castle Drive Armonk, NY 10504-1785 U.S.A. The following paragraph does not apply to the United Kingdom or any other country where such provisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS PUBLICATION "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer of express or implied warranties in certain transactions, therefore, this statement may not apply to you. This information could include technical inaccuracies or typographical errors. Changes are periodically made to the information herein; these changes will be incorporated in new editions of the publication. IBM may make improvements and/or changes in the product(s) and/or the program(s) described in this publication at any time without notice. Any references in this information to non-IBM Web sites are provided for convenience only and do not in any manner serve as an endorsement of those Web sites. The materials at those Web sites are not part of the materials for this IBM product and use of those Web sites is at your own risk. IBM may use or distribute any of the information you supply in any way it believes appropriate without incurring any obligation to you. Information concerning non-IBM products was obtained from the suppliers of those products, their published announcements or other publicly available sources. IBM has not tested those products and cannot confirm the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. This information contains examples of data and reports used in daily business operations. To illustrate them as completely as possible, the examples include the names of individuals, companies, brands, and products. All of these names are fictitious and any similarity to the names and addresses used by an actual business enterprise is entirely coincidental. COPYRIGHT LICENSE: This information contains sample application programs in source language, which illustrates programming techniques on various operating platforms. You may copy, modify, and distribute these sample programs in any form without payment to IBM, for the purposes of developing, using, marketing or distributing application programs conforming to the application programming interface for the operating platform for which the sample programs are written. These examples have not been thoroughly tested under all conditions. IBM, therefore, cannot guarantee or imply reliability, serviceability, or function of these programs. You may copy, modify, and distribute these sample programs in any form without payment to IBM for the purposes of developing, using, marketing, or distributing application programs conforming to IBM's application programming interfaces.

Copyright IBM Corp. 2006. All rights reserved.

xiii

Trademarks
The following terms are trademarks of the International Business Machines Corporation in the United States, other countries, or both: AIX 5L AIX BladeCenter DB2 developerWorks DFSMSdss Domino Enterprise Storage Server ESCON Eserver FICON HiperSockets i5/OS IBM iSeries Lotus Micro-Partitioning OpenPower OS/390 POWER5 POWER PR/SM pSeries RACF RDN Redbooks (logo) Redbooks RMF RS/6000 S/390 ServeRAID System p5 System z9 System z Tivoli TotalStorage VSE/ESA WebSphere Workplace xSeries z/OS z/VM z9 zSeries

The following terms are trademarks of other companies: iPlanet, CacheFS, Java, JavaScript, Portmapper, Solaris, Solstice, Solstice DiskSuite, Sun, Sun Microsystems, SunOS, SunScreen, and all Java-based trademarks are trademarks of Sun Microsystems, Inc. in the United States, other countries, or both. Microsoft, MS-DOS, Windows, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both. Intel, Intel logo, Intel Inside logo, and Intel Centrino logo are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States, other countries, or both. UNIX is a registered trademark of The Open Group in the United States and other countries. Linux is a trademark of Linus Torvalds in the United States, other countries, or both. Other company, product, or service names may be trademarks or service marks of others.

xiv

Solaris to Linux Migration: A Guide for System Administrators

Preface
The goal of this IBM Redbook is to provide a technical reference for IT system administrators in organizations that are considering a migration from Sun Solaris to Linux-based systems. We present a system administrator view of the technical differences and methods necessary to complete a successful migration to Linux-based systems, including covering the differences between major Linux distributions and IBM server and storage platforms. This book is designed primarily to be a reference work for the experienced Sun Solaris 8 or 9 system administrator who will be working with Linux. It is not a Linux administration how-to book for beginning system administrators, but rather a guide for experienced administrators who need to translate a given Solaris system administration task to Linux. We organize this book into four main parts. In Part 1, Background and planning on page 1, we provide an introduction the topic and a general purpose planning guide for migrations. Part 2, System administration differences guide on page 21 covers of a broad set of system administration topics. Chapters in this part focus on specific configuration management tasks and system functions for which a system administrator is typically responsible. In each chapter, we present the topics with the goal of identifying the major differences between how the tasks and functions are managed in Solaris and Linux. Wherever possible, we also highlight any differences in how those tasks are accomplished between Red Hat Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SLES). Because it would be impossible to provide a comprehensive reference about each topic in a single volume, we provide links to find more detailed information about the topics presented in each chapter. In Part 3, IBM eServer platforms on page 235, we provide technical implementation details for installing and running Linux on IBM Eserver platforms. We do not attempt to provide a complete reference for running Linux on each IBM platform. Instead, we focus on the Linux administration tasks that change due to hardware differences, such as boot configuration or virtualization functions. To accomplish this, we identify and explain how to leverage the additional industry-leading technologies in IBM Eserver xSeries, IBM POWER technology (iSeries/pSeries), and IBM Eserver zSeries systems

Copyright IBM Corp. 2006. All rights reserved.

xv

and IBM System z that make them very powerful and flexible platforms for hosting Linux-based solutions. Part 4, Appendixes on page 343 provides quick reference tables for looking up Solaris to Linux task equivalencies, comparing system configuration files and locations, and using comparable commands (different name, comparable function) and common commands (same name and purpose). In Part 4, we also include a sample chapter from the following related publication: Mendoza, et al., UNIX to Linux Porting: A Comprehensive Reference, Prentice Hall, 2006, ISBN 0131871099 Along with the sample chapter, we include link information so that you can order the UNIX to Linux Porting book at a discounted price.

The team that wrote this redbook

This redbook was produced by a team of specialists that came from around the world to work together at the IBM International Technical Support Organization in Austin, Texas, in October and November of 2005. Mark Brown is a Senior Technical Staff Member at IBM Linux Technology Center in Austin, Texas. He is currently the technical lead for UNIX to Linux porting efforts within the Linux Technology Center, and the IBM corporate representative to the IEEE POSIX and UNIX trademark standards working groups. Mark is a coauthor of the IEEE POSIX 1003.1 and 1003.2 specifications, as well as the Single UNIX Specification. His experience includes more than 20 years in UNIX and Linux operating system development, with an emphasis on C language libraries and ABIs. Chuck Davis (CISSP-ISSAP) is a Security Architect for IBM Global Services. Chucks responsibilities are to author and support the IBM global e-business UNIX security policies, write vulnerability tests for the IBM internal penetration testing application, and support the IBM internal host-based intrusion detection solution. Chuck has been a speaker at information security conferences worldwide. Prior to working for IBM, Chuck spent six years working in the Internet industry as a systems security administrator, webmaster, and Internet engineer. Chuck holds certifications from Sun Microsystems, Cisco Systems, and (ISC)2. William Dy has been a member of the Canadian IBM Information Technology Services Americas (formerly known as IBM Global Services) since late 1996. He has worked on various platforms, including Sun SPARC, HP RISC, IBM RS/6000 (now known as pSeries), and other smaller UNIX-based systems. His current main concentration is the Sun Solaris operating system, and he provides

xvi

Solaris to Linux Migration: A Guide for System Administrators

day-to-day operational support for various commercial outsourcing accounts in Canada, as well as some IBM internal servers. His interests include writing scripts and programs to automate specific tasks. He currently resides in Calgary, Alberta, Canada. Paul Ionescu is a Senior IT Specialist working for IBM Global Services Romania. He has more than 10 years of Linux and 7 years of Sun Solaris systems administration experience focusing on networking and security. He is a Sun Certified System Administrator and Network Administrator for Solaris 9, a Certified Information Systems Security Professional (CISSP), a Cisco Certified Internetworking Expert (CCIE), a Cisco Certified System Instructor, and a Red Hat Certified Engineer (RHCE), and holds several other Linux, networking, and security certifications. He is currently focused on network engineering projects for the ISP, telco, and banking industries, and is also helping customers migrate from other operating systems to the Linux OS. Paul joined IBM in 2000. Jeff Richardson is an Advisory Software Engineer in the IBM Software Group Linux Integration Center. Jeff has a bachelor of arts degree in Computer Science from Saint Edward's University and is an IBM Certified Database Administrator for IBM DB2 UDB V8.1 for Linux, UNIX, and Microsoft Windows. Jeff has worked for IBM for more than 23 years. His first assignment was an OS/MVS and z/VM system operator and analyst. Since then, he has worked with IBM Middleware products on different operating systems and hardware platforms in functional and system test teams. Jeff currently develops educational material and provides pre-sales technical support for IBM Middleware products running on xSeries, pSeries, and zSeries for Linux. Kurt Taylor is a Senior Software Engineer at the IBM Linux Technology Center, located in Austin, Texas. Kurt is currently a Technical Team Leader for the Linux System Management Team. He has worked at IBM for 10 years, and has more than 17 years of industry experience architecting, designing, and developing distributed network and system management products. He is currently working with the OpenPegasus project and the OSDL Desktop Linux work group. Kurt graduated from the University of North Texas with a bachelor of science degree in Computer Science. Robbie Williamson is an Advisory Software Engineer working in the IBM Linux Technology Center (LTC). He is currently the IBM Eserver Red Hat Linux Test Architect. Robbie has a bachelors degree in Computer Science and is completing his masters degree in Engineering Management from the University of Texas at Austin. He has more than seven years of UNIX/Linux experience, including IBM AIX 5L technical support and testing, Hewlett-Packard HP-UX development, and Linux development and testing. He is also an open source project maintainer for the Linux Test Project.

Preface

xvii

Production of this IBM Redbook was managed by: Chris Almond, an ITSO Project Leader and IT Architect based at the ITSO Center in Austin, Texas. In his current role, he specializes in managing content development projects focused on Linux and AIX 5L systems engineering. He has a total of 15 years of IT industry experience, including the last six with IBM.

Acknowledgements
A complete and detailed Redbook about a topic such as this would not have been possible without the support of the IBM Worldwide Linux team, the IBM Linux Technology Center, and the IBM Systems and Technology Group. The primary sponsors supporting development of this book include: Terese Johnson, Susanne Libischer, Mark Brown, Susan Greenlee, Connie Blauwkamp, and John Milford Additional content for this book was provided by Alan Robertson, an IBM employee and a lead developer in the High Availability Linux project. The Redbook team would like to acknowledge the following people for their contribution of draft review feedback in support of this project: Roland Caris, Shane Kilmon, Cliff Laking, Michael MacIsaac, Stephen Wehr, Elton Rajaniemi, and Jeroen van Hoof The team would also like to acknowledge the support efforts of Greg Geiselhart and Stephen Hochstetler from IBM ITSO, Lucy Ringland from Red Hat Inc., Steve Hansen from Novell Inc., and our editor for this book, Elizabeth Barnes, from our ITSO Authoring Services team.

Become a published author

Join us for a two- to six-week residency program! Help write an IBM Redbook dealing with specific products or solutions, while getting hands-on experience with leading-edge technologies. You'll team with IBM technical professionals, Business Partners and/or customers. Your efforts will help increase product acceptance and customer satisfaction. As a bonus, you'll develop a network of contacts in IBM development labs, and increase your productivity and marketability.

xviii

Solaris to Linux Migration: A Guide for System Administrators

Find out more about the residency program, browse the residency index, and apply online at:
ibm.com/redbooks/residencies.html

Comments welcome
Your comments are important to us! We want our Redbooks to be as helpful as possible. Send us your comments about this or other Redbooks in one of the following ways: Use the online Contact us review redbook form found at:
ibm.com/redbooks

Send your comments in an e-mail to:

[email protected]

Mail your comments to: IBM Corporation, International Technical Support Organization Dept. JN9B Building 905 11501 Burnet Road Austin, Texas 78758-3493

Preface

xix

xx

Solaris to Linux Migration: A Guide for System Administrators

Part 1

Part

Background and planning

This part provides an introduction to the topic and a general-purpose planning guide for migrations.

Copyright IBM Corp. 2006. All rights reserved.

Solaris to Linux Migration: A Guide for System Administrators

Chapter 1.

Introduction
This chapter introduces Linux and the topic of the system administration of the Linux operating system from the perspective of administrators familiar with the Solaris operating system versions 8 or 9 and who are planning their first move to a Linux system. This chapter includes the following topics: What is Linux? Linux kernels, distributions, and distributors How this book is organized Linux administration overview Reference materials

Copyright IBM Corp. 2006. All rights reserved.

1.1 What is Linux?

In 1991 Linus Torvalds, back then a university student of computer science at the University of Helsinki, sent a historic post to the MINIX1 news group announcing that he was working on a new free operating system and was interested in getting feedback about features that should be included in Linux. A few weeks later, Linux version 0.01 was released on the Internet, and its development became a shared effort as developers downloaded code, tested it, tweaked it, and sent it back to Linus, who then integrated them into his project. New versions came rapidly, and the amount of followers grew as the code went worldwide through FTP sites. Soon Linux became popular among students and developers and the project expanded. Linux was licensed under the GNU General Public License (GPL), ensuring that the source code had to be included and it would be free for everyone to copy, study, and change. You can read the GPL:
http://www.gnu.org/copyleft/gpl.html

Although Linux itself was free, commercial vendors moved in soon in order to make money by gathering and compiling software and thus creating a package easy to install and distribute, more like the other, more familiar commercial operating systems. Today, Linux has been ported to many other platforms, from hand-held devices to mainframes. It is the fastest growing OS in the server market and is used in enterprise as well as single-purpose server installations. The open nature of its development process has led to rapid and frequent additions of features and technologies. Added support for clustering, for example, has enabled highly available and scalable solutions using Linux. Linuxs roots in the UNIX operating system philosophy make it stable, reliable, and expandable, while at the same time, familiar to experienced UNIX users and administrators.

1.2 Linux kernels, distributions, and distributors

The term Linux (which is a trademarked term, held by Linus Torvalds) is often used in different ways: the operating system kernel itself, the entire operating system built around this kernel, or a distribution, in other words, a packaged Linux OS product.
1

Small UNIX clone designed for IBM PCs and compatibles. It is ideal for educational purposes due to its small size, microkernel-base design, and full documentation. Although MINIX comes with the source code, it is copyrighted software.

Solaris to Linux Migration: A Guide for System Administrators

1.2.1 The Linux kernel

The term kernel in this sense is meant in the same sense as in UNIX, that is, the core layer of the OS that controls hardware, manages files, creates processes, and so forth. Also included in this sense are the set of loadable modules and drivers that come with the basic kernel. The kernel is maintained by a team of volunteers, headed by Linus Torvalds, who work on maintenance and new features. They are aided in this by developers from all over the world (including IBM) who contribute fixes (patches) and new feature code as candidates for inclusion. Although a person or organization is free to make a copy of the Linux kernel source code and make their own version or variation of it, Linus kernel project is considered the one official Linux kernel.

1.2.2 The Linux operating system

Used in this sense, the term Linux includes all of the drivers, modules, libraries, tools, utilities, and applications that make up an operating system based on the Linux kernel. These are almost always entirely made up of open source products: The base C and C++ libraries, compiler, and associated development tools, base utilities (ls, ps, mv, etc), and many other packages of libraries and utilities come from Free Software Foundation projects (GNU). Other tools, from the shells (bash, pdksh) to scripting languages (Perl, PHP) to Internet daemons (xinetd, wu-ftp) come from individual open projects. Some parts come from large, organized open projects (X, GNOME desktop, KDE desktop, Apache Web Server). These packages and applications are almost exclusively distributed for use under open licenses (source code must be made freely available) by anyone who wishes to use them. The most widely known and used license is the Free Software Foundations General Public License (GPL). Under the GPL or the other licensing terms of open software packages, anyone can produce a Linux OS or variant thereof. There is no one official Linux OS.

1.2.3 Linux distributions

Distributions are Linux operating systems, made up of a kernel and a collection
of the necessary tools and applications to make a functioning, useful, stable operating system for general use. A distribution has taken this collection and produced a packed, installable product that can be easily distributed. Anyone can create a distribution, just as anyone can create a Linux OS or use the Linux kernel. There is no one official Linux distribution.

Chapter 1. Introduction

Linux distributions all have one thing in common: They provide Linux in source code and binary form. The base code is freely redistributable under the terms of GNU General Public License (GPL) and any non-free and contributed software has to be clearly marked as such. Distributions divide into two main categories: Non-commercial distributions tend to focus on creating technical excellence and follow a design philosophy or manifesto. These distributions do not offer any direct support other than mailing lists and bug reports. They are freely available on the Internet for download, and for the cost of materials and shipping, some can be purchased as a shrink-wrapped product with manuals. Commercial distributions offer training, certification, and support services in addition to their software, which is commercially distributed through official dealer channels or downloadable from the company's Web sites for payment. Most commercial distributions base themselves on non-commercial ones and are available for free download. However, any commercial software bundled with the product is only available on a separate CD for purchase. Some of the best known and widely-used distributions include Red Hat, SUSE, Mandriva, and Debian. This book describes Linux using the operating system sense and in terms of two major distribution vendors: Red Hat (Red Hat Enterprise Linux 4) Red Hat is a well-known Linux and open-source provider. It was founded in 1993. Red Hat Enterprise Linux runs on multiple architectures and is certified by enterprise software and hardware vendors. Red Hat is based on the Linux kernel version 2.6. Novell SUSE (SUSE Linux Enterprise Server 9) The first distribution of SUSE Linux began in 1993. The company was acquired by Novell in 2004. SUSE Linux Enterprise Server 9 (SLES9) is a secure, reliable platform for open-source computing, supporting a range of hardware platforms and software packages. It also provides open APIs and other development tools to help simplify Linux integration and customization. SLES9 includes the Linux kernel version 2.6.

1.3 How this book is organized

This book is designed primarily to be a reference work for the Solaris 8 or 9 system administrator who will be working with Linux. It is not a Linux administration how-to book for beginning system administrators, but rather a

Solaris to Linux Migration: A Guide for System Administrators

guide for experienced administrators that need to translate a given Solaris administration task to Linux. We organize this book into four main parts: Part 1, Background and planning Part 2, System administration differences guide Part 3, IBM eServer platforms Part 4, Appendixes Each section of the book attempts to outline the considerations for the experienced Solaris system administrator to keep in mind, the new and different tools and features, and where to go for further in-depth information about the topic. The system administration differences guide section provides task-based technical topics in terms of two major Enterprise Linux server distributions (Red Hat Enterprise Linux and SUSE Linux Enterprise Server) where a generic hardware platform is assumed. The IBM Eserver platforms part outlines the administration tasks that change due to hardware differences, such as boot configuration or virtualization functions. This book also includes appendixes, containing direct side-by-side task, command, and configuration file comparisons, for use as a quick reference.

1.4 Linux administration overview

As mentioned earlier, the Linux kernel and operating system are based on the UNIX philosophy and practice. Also, Linux contains almost all of the features described by the IEEE POSIX 1003.1-2001 standard for interfaces, commands, and utilities. These two facts will make Linux seem very familiar to a Solaris system administrator. Some of the many basic features in common between Solaris and Linux include: Users, groups, and a root superuser File system and directory hierarchies init, boot configuration/rc.* scripts Processes and POSIX threads Networking stacks and services ISO C/C++ library, basic utilities, and basic shells vi, emacs, and other editing tools Accounting and log files X Window System

Chapter 1. Introduction

This is just a short list of commonalities. Almost all of a Solaris administrators skills and experience should transfer easily to Linux, with a short learning curve to pick up the differences in the common tools (in most cases small) or learn the equivalent new tools (in some cases).

1.4.1 Documentation
Documentation for Linux comes in several forms, the most common of which are manual pages (man) and HTML (either on the system or over the network). Almost all commands can be referenced through the man command, which works in the same way as on Solaris and is organized in the same manner (sections 1,2, and so on). System guides are usually provided in HTML or PDF format. Some documentation, particularly for parts of the system that come from Free Software Foundation projects, is documented using the info tool. The C library (GNU libc, or glibc) is one example of this. This is a command line tool that operates as a full-screen, text-based GUI. It has an extensive built-in help system, but requires some time to learn. For basic information about the info tool, use the command man info.

1.4.2 Utilities
Some of the most common software packages that are used to manage your Linux servers include: YaST and YaST2 for SUSE Linux YaST stands for yet another system tool and is a system management GUI. YaST can be started by typing the following command as root on the Linux command line:
~> yast

It contains modules for managing the installation and configuration of the system features. Red Hat setup and system-* utilities setup is the Red Hat implementation of the Linux configuration tool. It is used to configure and manage your Linux server. As system administrator (root) on the Linux command line, start setup by typing the following command on the command line (it will ask for the root password):
~> setup

There are also individual GUI-based utilities for configuration of various system features in the /usr/bin path, as shown in Table 1-1 on page 9.

Solaris to Linux Migration: A Guide for System Administrators

Table 1-1 Red Hat system configuration GUI tools system-cdinstall-helper system-config-authentication system-config-boot system-config-date system-config-display system-config-httpd system-config-keyboard system-config-language system-config-lvm system-config-mouse system-config-netboot system-config-network system-config-network-cmd system-config-network-druid system-config-nfs system-config-packages system-config-printer system-config-printer-gui system-config-printer-tui system-config-rootpassword system-config-samba system-config-securitylevel system-config-securitylevel-tui system-config-services system-config-soundcard system-config-time system-config-users system-control-network system-install-packages system-logviewer system-switch-im system-switch-mail system-switch-mail-nox

VNC Virtual Network Computing (VNC) is free software that enables you to view and interact with another computer, which might or might not have the same operating system. It gives you a remote desktop view of the other system. VNC or its equivalent is shipped with most Linux distributions. However, the default mode in VNC on Linux provides the connecting party with a new desktop screen, not the console screen as under Microsoft Windows.

1.4.3 Standards
Much like the Solaris OS, there are specifications that cover operations in the Linux space as well. The Linux Standard Base (LSB, http://www.linuxbase.org/) covers the implementation of a conforming Linux environment in much the same way that the Single UNIX Specification (SUS, Open Group) covers the UNIX environment. It covers application programming interfaces (APIs), utilities, software packaging, and installation, and additionally

Chapter 1. Introduction

specifies ABIs for the most common Linux platforms. We recommend that you investigate LSB support when choosing a Linux distribution for use. Linux offers IEEE POSIX compliance where possible in both the API and utilities sets. There is also significant overlap in conformance with the Single UNIX Specification, though Linux cannot be called UNIX.

1.4.4 Support
Because the distributions themselves are made up of free software, the business model of Enterprise Linux distribution vendors typically emphasizes the selling of support and services related to their offerings. These services range from self-help advice forums (for the lower-end or desktop systems) to full 24x7x365 enterprise-level maintenance with support level contracts. Additionally, third parties (including IBM) offer support services related to Linux. Consider support and updates when choosing a distribution. Some of the larger distributors also offer Linux system administration courses.

1.5 Reference materials

There are many sources for technical information about Linux. Some additional sources related to topics discussed so far in this book include: Linux and Linux solutions in general: Siever, E., et al., Linux in a Nutshell, O'Reilly Media, Inc., 5 Edition, 2005, ISBN 0596009305 IBM Redbook Linux Handbook A Guide to IBM Linux Solutions and Resources, SG24-7000 The Linux Documentation Project, online guides to almost any Linux topic:
http://www.tldp.org/

Open source licenses and how they work: The Free Software Foundation:
http://www.fsf.org/

The Open Source Initiative:

http://www.opensource.org/

Application development: Linux Standard Base Team, Building Applications with the Linux Standard Base, IBM Press, 2004, ISBN 0131456954

10

Solaris to Linux Migration: A Guide for System Administrators

Chapter 2.

Planning for migration

This chapter provides some useful information regarding the non-technical and technical issues involved with a migration to a new platform in an established IT environment. Migrations, whether in a data center or on client workstations, involve considerations beyond simple OS operation. In many cases, the move to (or addition of) a new platform involves organizational and strategic changes as well. In this chapter, we cover some of the important planning steps involved in a migration to a Linux platform. This chapter includes the following topics: Assemble the stakeholders Set objectives and define scope Assess workload and environment Assess skill requirements Build a plan

Copyright IBM Corp. 2006. All rights reserved.

11

2.1 Assemble the stakeholders

As the name implies, these are the people with a stake in the migration. This group is not limited just to the people responsible for directly maintaining the systems. End users of the services (especially if it is a client workstation) and business managers have an interest in a successful move that meets the goals of the organization. A meeting of stakeholders (or representatives of large groups of stakeholders) is a good way to set expectations and to perform some of the other planning considerations put forward in this chapter. A meeting of this group of people will be a good way to identify if any IT environment or administrator, manager, and user skills changes are needed, for example. These will also be the people to whom status and results milestones are reported. Many times as a system administrator, you will not be the one proposing or funding the migration; instead, you are usually the implementor of a migration drive decided by others. In order to make sure that all interests are taken into account, it is still a good idea to request a meeting of the key people asking for and affected by the move.

Communicating the change

Such meetings are also good ways to open communication channels. The effect of a good communications plan will be to flatten out the negative aspects of the acceptance curve, as shown in Figure 2-1.

Figure 2-1 Acceptance of new technologies

A communications plan coupled with proper training on the new system should minimize the number of users that fall into the rejection/opposition mode, cause

12

Solaris to Linux Migration: A Guide for System Administrators

users to start out with acceptance instead of dissatisfaction as the initial response, and lead to a quick transition into exploration and productive use. Regarding the IT support team, these same issues have even more importance. A strategic decision to switch the OS can cause the impression of disapproval of the work they have done so far. It might give staff the impression that their current skills are being devalued. It probably will not be easy to convince an administrator of Solaris systems that migrating to Linux is a good strategy unless you can also convince them that the organization is prepared to make a significant investment in upgrading their skills as well.

2.2 Set objectives and define scope

You should be able to articulate the objectives for your Linux migration and relate them to your key business drivers. Whether you are trying to gain efficiencies by reducing costs, increasing your flexibility, improving your ability to support and roll out new application workloads, or some other key business drivers, set up objectives that line up with these. Even the smallest of migrations should be able to do this, and it will help guide your planning. Defining metrics (increased performance? more uptime?) up-front during the project will help your team stay focused. Be sure that you will have a means of tracking the metrics. Getting stakeholder agreement on your metrics early in the project will help ensure executive and user support. Often, the migration to Linux will be accompanied by other objectives. For instance, some customers upgrade their database at the same time to get the latest features and performance enhancements and to obtain support that lines up well with the latest distributions of Linux. As with any project, the scope must be well defined to prevent project overrun, but it is also important that you have a means to manage additions to the plan as business needs dictate. Because cost is often a key motivator for migrating to Linux, give careful consideration to identifying where cost reduction is targeted. Identify metrics for defining return on investment prior to beginning migration activities. Questions to ask in this area include: What are the tactical goals of the migration? What measurements of success will validate achieving those goals? What are the financial aspects of the migration (budget, return on investment, justification, and so on)? What are the strategic goals for the migration? How does the migration fit into the overall IT strategy for the organization?

Chapter 2. Planning for migration

13

The system administrators perspective

While the objectives and scope of the migration will determine the size of the work involved, it is meeting the metrics (goals) that will guide everything from the hardware choices to the configuration of the various subsytems in the OS. Most of the considerations can be tackled the same way as in a move to another flavor of commercial UNIX. However, for many areas (security and firewalls, mail service, and databases are a few), there are multiple tools and packages available to achieve your goals. Even though this book outlines the differences between Solaris and Linux (from a system administrator point of view) and lists many of the choices (in those areas that have them) as well, we strongly advise that you continue your research beyond this book as necessary to determine which of the variety of tools for a given task best fits your metrics. For example, methods for evaluation of PostgreSQL, MySQL, or a commercial database such as IBM DB2 to determine which is best for your use is beyond the scope of this book. But as administrator of the system running the database, such a choice can strongly affect how you initially configure the system and what types of tools you use for ongoing management of the system.

2.3 Assess workload and environment

The decision regarding which workloads to migrate needs to be tied to your overall objectives. A Linux migration usually does not have to be an all or nothing approach. Many organizations choose to start small to gain experience with Linux before migrating larger or mission-critical services to Linux. The following sections present some general considerations that should be factored into this decision process.

2.3.1 Workload types

The following sections show some sample workloads, roughly arranged from simple to complex, from the system administrators perspective. These can vary in order of complexity depending on a given organization's environment; however, these are some general guidelines, based on IBM Global Services experience with migrations.

File and print infrastructure

This type of workload can be moved to Linux, as Linux is easily configured to provide file and print service to a wide range of clients, from UNIX-based workstations to Windows. Linux distributions ship with all of these tools available in their default installations.

14

Solaris to Linux Migration: A Guide for System Administrators

Security and directory infrastructure

Linux offers full LDAP, Kerberos, Network Information Service (NIS), firewalling, and other tools in this area that meet the appropriate industry standards. In fact, some of the most modern versions of security services were available from the major Linux distributors earlier than on other non-Linux based systems. LDAP schemas might differ from Solaris to Linux, and work to configure the proper integration between Kerberos levels might also be needed. If you are using NIS+ on Solaris, plan on transitioning to LDAP or NIS.

E-mail
Linux offers full sendmail, IMAP, and POP3 support, but at least one major Linux distributor (Novell SUSE Linux Enterprise Server) now offers the new postfix MTA as the default package (while still offering sendmail as an option). The IMAP and POP daemons (and clients) will each have a learning curve for configuration and use. End-user clients should not see a difference between the Solaris and Linux mail service.

Database serving
This type of migration will not only involve an operating system change, but will also involve moving a major application and its data. Both IBM DB2 and Oracle have Linux offerings, in addition to the open-source MySQL and PostgreSQL packages. The IBM Redbook Up and Running with DB2 for Linux, SG24-6899, might be helpful for information about this topic.

Web and application serving

A full suite of Web services utilities are shipped with most Linux distributions, with all of these containing the Apache HTTP server package. Perl, PHP, and Java (versions from IBM as well as Sun) are all available natively. You will need to research the needs of the application/service to be offered and how to make those things (such as the middleware and all of its prerequisites) available on the Linux platform.

Messaging
Messaging systems incorporate some elements of high availability systems and introduce scalability and networking factors as well. Research will be needed to properly specify your Linux equivalent system requirements. IBM offers an entire suite of messaging middleware (our WebSphere and WebSphere MQ lines) for Linux.

Enterprise resource planning (ERP)

ERP systems tend to be cross-functional and enterprise-wide. Therefore, there are many interactions with other systems and processes for which to plan and account. The system administrator will need to carefully plan for successful integration in addition to the software and OS changes.

Chapter 2. Planning for migration

15

Business Intelligence
These systems incorporate several types of software (such as data mining and decision support applications) and use data from several different sources. In addition to the software research involved, the system administrator will need to plan for system communication and workload management issues.

2.3.2 ISV applications

If you are running ISV applications on systems that you are targeting for migration, you need to determine if the vendors provide compatible versions that support the distribution of Linux that you plan to use. Many ISV applications have other third-party dependencies. Be sure to map out all ISV dependencies, including middleware. Most leading middleware products are available on Linux, and there are many open source alternatives.

2.3.3 Custom application porting

Applications developed in-house might required porting and testing on Linux. The effort involved can vary greatly depending on how portable the application code is. Open source and commercial tools are available to help with tasks such as assessing the portability of your applications. IBM Global Services, as part of their migration services offerings, uses tools developed in cooperation with IBM Research to help with code assessment and conversion.

2.3.4 Databases
Consider whether you want to move the database and the application, or just one or the other. Database administration is a different but related skill to system administration. These skills are as equally transferable to Linux due to the similarity to UNIX-based systems and the fact that the most popular database offerings on the market are also available for Linux now.

2.3.5 Integration with existing network services

A part of your planning will be based on whether you are migrating a client or a server platform. While Linux can work equally well in either role, the differences include what type of distribution you install and what hardware configuration you use. While this book will serve you well in either capacity for the purposes of system administration, the IBM Redbook Linux Client Migration Cookbook A Practical Planning and Implementation Guide for Migrating to Desktop Linux, SG24-6380, might be a helpful reference as well.

16

Solaris to Linux Migration: A Guide for System Administrators

Integration with other Solaris systems

Typically, the Linux migration will occur slowly or as a part of a pilot project, and thus the Linux system will be introduced into an environment where the interaction with Solaris systems is a requirement. You can reference this book in the research stage in order to work out the planning for interoperability in the administration tools and automation.

Integration with other Windows systems

Linux offers several tools for managing integration with Microsoft Windows systems. There are many resources available for further research to plan for integration points such as file sharing, printing, and so on.

2.3.6 Other important infrastructure requirements

You need to closely analyze how your existing functional requirements have defined the design of your current infrastructure. How will you meet these requirements in a Linux-based environment? Factor in your requirements for security, availability, operations, maintenance, and so on. What infrastructure is needed to support your specific workloads? Many of the same infrastructure products available for UNIX are also supported on Linux.

2.3.7 Special hardware

Are there special devices that will be supported? Printers, storage, or other hardware necessary to your operation must be investigated for Linux support.

2.4 Assess skill requirements

One of the key factors that is helping to accelerate migrations from UNIX to Linux is the availability of skills and the ease of transferring UNIX skills into a Linux environment. UNIX and Linux are similar in many respects, enabling system administrators to easily switch to a Linux environment. Also, many of the same tools that they use on UNIX are available on Linux. Still, a skills assessment should be a part of your migration planning so that you can identify areas where skills gaps may exist and plan for how you will cover any skills gaps. Several avenues for Linux training are available, including IBM, Linux distributors, and others.

Hardware-specific skills
Although almost all of the skills related to managing a Linux system are the same regardless of platform, some of the activities related to the interaction of Linux with the underlying hardware and maintenance of the hardware itself are different

Chapter 2. Planning for migration

17

based upon the hardware type. This is especially true for larger systems, including IBM Eserver POWER and zSeries platforms. Administrator education might be required for these systems.

A training option for client users

Many Linux distribution vendors are now providing live or bootable CD-ROM-based distributions. One notable live CD distribution is provided by Knoppix:
http://www.knoppix.com

The following description from Knoppix further explains what a live CD provides: KNOPPIX is a bootable CD with a collection off GNU/Linux software, automatic hardware detection, and support for many graphics cards, sound cards, SCSI and USB devices and other peripherals. KNOPPIX can be used as a Linux demo, educational CD, rescue system, or adapted and used as a platform for commercial software product demos. It is not necessary to install anything on a hard disk. Due to on-the-fly decompression, the CD can have up to 2 GB of executable software installed on it. A live CD can be used to provide users with the ability to run a bootable Linux system on their desktop. They can use it to test and evaluate the UI, applications, and other facets of the Linux client, before an actual client migration occurs. And this can be done without harming the host operating system installation on the local hard disk. Another benefit of using a live CD as part of a migration plan is for detection of hardware and device problems. The live CD helps validate proper hardware detection and device support and identifies any issues prior to actual client migration.

2.5 Build a plan

A key success factor is a well-built plan. Understanding your objectives and having metrics with stakeholder involvement and agreement help provide a good base from which to build a plan. Be sure to build in all key requirements, such as timing, resource needs, and business commitments such as service level agreements (SLAs). Also, be sure to include related aspects of the migration, such as new workloads, infrastructure, and consolidation. As with any change in an organization, cultural change might need to be managed. Anticipating change and involving affected teams early are good ways to manage culture change issues. For example, support staff for hardware might be comfortable with UNIX-related hardware support, where to go for help, and so on. The experts in the prior environment might be less open to change if they feel threatened by new ways of doing things where they are not the expert.

18

Solaris to Linux Migration: A Guide for System Administrators

Build in requirements
Consider the following areas: Resources What hardware will be required? Software? Identify what housing issues are required (are the electrical and cooling inputs equal?). What staff will be required to help with the crossover? Education Does the staff have adequate Linux education? Are there special skills required for the hardware or Linux/hardware combination? Service level agreements While the installation, configuration, and testing of the change is taking place, what are the support mechanisms (both yours and those of any vendors)? What are your commitments to current stakeholders while you are performing the migration? Related aspects to project Be sure to set out what other things are happening in addition to the basic system changeover. Identify skills-related requirements

Include documentation time

When installing and configuring the new system, documenting (even if it is just a short set of notes) will help in reproducing the process and in troubleshooting problems should they appear.

Add a testing period

After the new system is installed, initialized, and configured, there should be time in the schedule for testing the system. Does it interoperate correctly? Can it handle the expected load? Your plan should have some metrics for judging the success of the test period.

Identify and time crossover

It is important to understand what the expected impact will be to the expected usage of the system during the actual migration period. Many administrators schedule crossovers to new systems during low-load periods to minimize disruption of services. In other environments, such as those with no light load times, there will be a period where both new and old systems both remain online until the load can be moved entirely to the new system. This can also be used as the testing period for the migration in these environments.

Chapter 2. Planning for migration

19

20

Solaris to Linux Migration: A Guide for System Administrators

Part 2

Part

System administration differences guide

In this part of the book, we provide chapters describing system administration differences, where each chapter focuses on a specific funtional area. We assume that the reader is an experienced Solaris system administrator who is undertaking (or considering) a migration to Linux. Therefore, each chapter assumes reader knowledge of how the task topic is performed on a Solaris system and compares that with the procedures for doing the similar task on Linux. In each chapter, we attempt to match up Solaris GUI tools, commands, and configuration files with Linux-based counterparts, where they are similar. We also describe differences in or lack of comparable functionality. The chapters include pointers to the appropriate Linux guides and how-to documentation.

Copyright IBM Corp. 2006. All rights reserved.

21

These chapters are not a substitute for the Linux documentation, but rather a cross-reference work for an administrator trying to find out what tools are available on Linux to perform a given job. In this part, we focus on the aspects of system administration in Linux that are hardware-neutral. Part 3, IBM eServer platforms on page 235 covers areas of administration that are unique to a given IBM Eserver platform (POWER technology-based, xSeries, and zSeries systems). Throughout this section, we make references to two of the leading Linux enterprise server products, Red Hat Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SLES). Note: We assume that the reader is an experienced Solaris system administrator who is undertaking (or considering) a migration to Linux. The content in this part of the book should not be used as a substitute for detailed Linux system administration documentation, but rather as a cross-reference work for an administrator trying to find out what tools are available on Linux to perform a given job. The differences guide portion of this book covers the following functional areas: Chapter 3, Operating system installation on page 23 Chapter 4, Disks and file systems on page 51 Chapter 5, Software management on page 75 Chapter 6, Device management on page 85 Chapter 7, Network services on page 107 Chapter 8, Boot and system initialization on page 125 Chapter 9, Managing system resources on page 139 Chapter 10, Printing services on page 151 Chapter 11, Users and groups on page 159 Chapter 12, Monitoring and performance on page 173 Chapter 13, Backup and restore on page 181 Chapter 14, Security and hardening on page 187 Chapter 15, Linux high availability overview on page 205 Chapter 16, Shell scripting on page 211 Chapter 17, Troubleshooting on page 215

22

Solaris to Linux Migration: A Guide for System Administrators

Chapter 3.

Operating system installation

This chapter describes the difference in the how Sun Solaris and Linux do their installations. Both have the option to use either a text-based or GUI installation format and allow installing from a different media sources. This chapter contains the following topics: Basic system installation Advanced installation and automation Network-based install: Solaris and Linux heterogeneous environments

Copyright IBM Corp. 2006. All rights reserved.

23

3.1 Basic system installation

In this section, we discuss the basic installation procedure with a local CD-ROM or DVD-ROM drive. Both Solaris and Linux offer other types of installation methods, which we cover in 3.1.2, Other installation methods on page 25. Local CD/DVD installations are very much the same for Solaris and Linux. The installation program prompts for information when it is required. A menu of options is presented in some parts so that you can select the appropriate choice or choices. As with a Solaris installation, Linux will perform the following operations. Note: The following list is not in any specific order and is a not complete list of the options that will be presented to you. It is a representation of the basic information that will be presented. When using another installation method, these prompts might not be displayed if they are preconfigured in the installation procedure, or you might have extra options. Probe the hardware for devices and load the appropriate device drivers. Assign a host name. Set up a network setup type: DHCP or static IP. If static IP, you will be asked to enter the IP address, subnet mask, default gateway, and DNS servers to use. Select a region type to use for languages. Select a time zone. Select the initial or upgrade installation type. Prompt for a hard disk to use for root partition. Prompt for a automatic or manual disk layout for the file systems Prompt for the software bundle or packages to install. Refer to 3.1.3, System bundle options on page 26 for a high-level view of the software selections available. In Linux, there might be other options that will allow you to set up sound card devices, graphical console resolution settings, and so on. You can proceed with setting these if you want, but most of the default options will be sufficient enough to start with. If not, they can be reconfigured at any time after the installation is done.

24

Solaris to Linux Migration: A Guide for System Administrators

After you selected these options, the installation continues by itself to complete setting up the partitions, building the file systems, and then installing the software packages. As the installation process continues, and depending on what software packages have been selected, it might prompt for other CDs or DVDs to be inserted when required.

3.1.1 Graphical or text installation

What is not mentioned in much detail in the previous section is the console device being used. As with Solaris on some SPARC platforms, you have the choice of using a graphical console to do the installation. And even if you are on a graphical console, you also have the choice of selecting a text-based installation. In Solaris, when installing from a CD set, the Installation CD will always start a graphical installation and the CD 1 will start a text-based installation. On the DVD, you will be prompted for which type of installation to use. In Linux for IA32, the installation boot CD will try to start the graphical installation, and it will fail back to text mode installation if necessary. You can also specify when you boot off the CD or DVD what type of console to use for the installation. If you have a graphical adapter, but for some reason want a text mode installation, you can specify so before booting the kernel and installation program. In addition, you can choose a different screen resolution if the default one is not working on your monitor. Note: Depending on the target platform or system, you might not have a graphical console at all. See Part 3, IBM eServer platforms on page 235.

3.1.2 Other installation methods

What we have discussed in the previous section is the installation of the software packages from a CD or DVD drive. Solaris and Linux both offer other ways to source the packages for installation. In Solaris, you have the following additional choices: Custom Jumpstart Web Start Flash Live Upgrade Factory Jumpstart

Chapter 3. Operating system installation

25

In Linux, you have the following additional choices: Another local hard drive partition (can be USB) NFS server FTP server HTTP server SMB or TFTP servers (only for SUSE) A customized Kickstart or AutoYaST installation from any of the available sources of installation For specific information about how to accomplish each of these methods, refer to the detailed installation documentation provided by the vendors.

3.1.3 System bundle options

If you use an installation method that has not been preconfigured to select the software packages for you, the installation program prompts you for what you want to install. In Table 3-1 on page 27, we outline the different selections available for Solaris and Linux. The most common choices for server platforms are Entire Distribution for Solaris and Server Install for Linux. What comes with each specific collection depends on how the distributor of your package determines what needs to be included. However, most of the common packages for each type of collection are included. But if they are not, you can easily install them later. Refer to Chapter 5, Software management on page 75 for information about how to install packages. Tip: If you are not yet comfortable with software management in Linux and you want to save some time, you can choose the Full Install software bundle to install everything on the Linux server as disk space permits.

26

Solaris to Linux Migration: A Guide for System Administrators

Table 3-1 Logical groupings of software clusters provided by installation Solaris Core: Required operating system files End-user system support: Core plus window environment Developer system support: End user plus the development environment Entire Distribution: Developer system plus enhanced features Entire Distribution Plus OEM: Entire distribution plus third-party hardware drivers (on SPARC only) Red Hat Workstation install (stand-alone client) Developer install (workstation install, plus developer tools, compiler) Server install (server services and software) Full Install (all software/services) Minimal Install (minimum packages for a Linux system) Custom (choose packages manually) SUSE Workstation install (stand-alone client) Developer install (includes developer tools, compiler) Server install (server services and software) Full Install (all software/services) Custom (choose packages manually)

3.1.4 Linux LVM consideration

In Solaris, the installer only works with a raw device in a UFS file system. It neither creates a SVM volume nor does it try to create any type of RAID structure. You are expected to do this step after the OS has been installed completely. In Linux, during the installation, you have the option of using Logical Volume Manager (LVM) and RAID when you are defining the hard disk for the root drive. In addition, you can select which hard drives to use. Important: When you use LVM in Linux, be aware that /boot should not be an LVM structure, but a raw disk partition. The /root and other file systems can be part of an LVM setup. The reason behind this is that the boot loader does not recognize an LVM structure; therefore, it cannot read anything out of an LVM partition. Refer to 4.10, Solaris Volume Manager to Linux LVM on page 65. Note: There might be other disk layout limitations depending on the target platform. Check the platform-specific chapters for more information about these restrictions. See Part 3, IBM eServer platforms on page 235.

Chapter 3. Operating system installation

27

3.1.5 Linux firewall

Linux comes with a firewall based on the iptables kernel modules as part of the installation packages, and it is enabled by default if you do not specify otherwise. For detailed information, consult 14.13, Firewalls on page 202. We cover the basics here because this question is handled as a part of a Linux installation.

Red Hat Enterprise Linux

During the interactive installation, you have the opportunity to select what services to permit, such as remote login (SSH), Web server (HTTP, HTTPS), file transfer (FTP), mail server (SMTP), and your own protocol/port numbers. Or, you might choose to disable the firewall completely. Additionally, Red Hat also has SELinux activated by default. SELinux stands for security-enhanced Linux. It allows more granular control of users, programs, and processes. For more information about SELinux, refer to Red Hat SELinux Guide or the NSA SELinux information page at the following locations:
http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/selinux-guide/ http://www.nsa.gov/selinux

If you need to reconfigure or disable the SELinux or the firewall, you can use the system-config-securitylevel command. This opens a UI tool that can be used to set up security options. Another way to disable SELinux without using the UI tool is to edit the /etc/sysconfig/selinux file and change the line from SELINUX=enforcing to SELINUX=disabled. This permanently disables SELinux at the next reboot. For the firewall, you can use the chkconfig command on the iptables service.

SUSE Linux Enterprise Server

The SUSE firewall is installed and enabled by default after the installation. Its configuration file is called /etc/sysconfig/SuSEfirewall2. There are three startup script files that set up the firewall software: SuSEfirewall2_init SuSEfirewall2_setup SuSEfirewall2_final All three files reside in /etc/init.d and have a symlink to the appropriate run-level subdirectory in which it is started. You can use the chkconfig command or yast2 firewall GUI to check the status of the firewall or to disable it.

28

Solaris to Linux Migration: A Guide for System Administrators

3.2 Advanced installation and automation

Automating the installation can become necessary in a large enterprise environment. For managing multiple systems, automated installation over a network is often the best solution. Where Solaris has custom Jumpstart, Red Hat has Kickstart, and SUSE has AutoYaST. There are no equivalent tools in Linux for the Solaris Web Start Flash and Live Upgrade.

3.2.1 Installing using a serial interface as input/output device

In Solaris, the installation program will use the current console as an I/O device, and if you are using a serial interface, it will use that same serial interface. In Linux for IA32 (see Part 3, IBM eServer platforms on page 235 for other IBM platforms), the default console device used for installation is the graphics card. If you need to install the system using a serial interface, you need to specifically tell the installation program to do so by adding the nofb console=ttyS0 kernel boot options. For example, if you are installing from a CD/DVD, you need to launch the installation from the graphical terminal boot prompt using linux nofb console=ttyS0. You then continue with the installation through the serial interface. If you want to automatically start the installation over the serial interface, one of the options is to remaster the installation CD/DVD in order to modify the isolinux.cfg boot configuration file from the media. For RHEL, edit the isolinux/isolinux.cfg file from the CD/DVD and modify the first line from default linux to default linux nofb console=ttyS0, and erase the timeout 600 option. For SUSE, edit the boot/loader/isolinux.cfg file from the CD/DVD and modify the first line from default harddisk to default linux nofb console=ttyS0, and erase the timeout 200 option. We do not describe here the CD/DVD remastering procedure because it is out of the scope of this book. If you are using the serial interface from a remote location, you probably do not want to change the CD or DVD, so the best way to remotely install with a serial console is to use a network installation and a serial console. For this, you need to set up the network boot and install servers as usual and add the nofb console=ttyS0 options to the kernel line served to clients.

Chapter 3. Operating system installation

29

3.2.2 Configuring a serial console device

The serial interface can be configured to act as the serial console in a similar way as on Sun SPARC systems. However, there might be a limitation based on the platform that you are using. Most IA32-based systems do not have the BIOS capability to have the serial interface activated from the point of power-on self-test (POST) to loading the boot loader. You will not be able to access the system up to this point. However, some enterprise servers have a separate management card that can be accessed through a serial or network and can be used to manage that server even if it is not powered on (refer to Part 3, IBM eServer platforms on page 235). This is equivalent with the Sun LOM feature. The GRUB1 IA32-based boot loader can be configured to be controlled from a serial line, and this is useful if you need to interactively manipulate the booting kernel, the kernel boot parameters, or both. If you are remotely controlling your server through a serial console, it is a good practice to set GRUB to listen to that serial console. To configure the GRUB boot loader to use the serial console, you need to add the parameters shown in the Example 3-1 at the beginning of its /boot/grub/menu.lst configuration file and comment out the splashimage or gfxmenu options.
Example 3-1 Entries required in /boot/grub/menu.lst file serial --unit=0 --speed=9600 --word=8 --parity=no --stop=1 terminal --dumb --timeout=10 serial console # splashimage OPTIONS (in RHEL) # gfxmenu OPTIONS (in SLES)

In this example, you might need to adjust the serial parameters to match your actual hardware setup. If you have a VT100 compatible terminal, you can remove the --dumb option from the terminal line. For more information about those parameters, consult the GRUB documentation. Important: If you do not comment out or erase the splashmenu or gfxmenu options from the GRUBs /boot/grub/menu.lst configuration file, it initializes the graphical mode and does not work on your serial as expected. For setting up the serial console for the rest of the Linux system, you have two options: For a temporary solution, at the GRUB boot prompt, edit the boot command line to have console=ttyS0,9600 for the first serial interface at speed of 9600 bps.
1

The GRand Unified Bootloader; see http://www.gnu.org/software/grub

30

Solaris to Linux Migration: A Guide for System Administrators

For permanent solution, edit the /boot/grub/menu.lst file and add the previous option to the kernel command line. If you want the console messages to be shown on both the graphical console and on the serial line, you can add the console=ttyS0 console=tty0option. And if you need a login prompt on that serial device, you need to add to the /etc/inittab a line similar to the one shown in Example 3-2.
Example 3-2 Enabling a login prompt on the serial line S0:2345:respawn:/bin/agetty -L 9600 ttyS0 vt100

Direct root login can be enabled on this serial interface by adding a ttyS0 to the end of the /etc/securetty file. For more information about configuring console logins over serial ports, refer to 6.3, Terminals and modems on page 101. Some enterprise-level IA32 platforms come with the serial console activated from the POST. Refer to Part 3, IBM eServer platforms on page 235 for more information.

3.2.3 Installing using remote display

In Solaris, you can install using the local graphic display or the serial interface. In Linux, you can install using the local graphic display, serial interface, and a few other remote modes. In RHEL, you have the following options: Add a display=IP_OF_REMOTE_X:0 argument to the kernel boot line for using a remote X Server for displaying the installation GUI. Add a vnc vncpassword=MIN_6_CHAR_PASS argument to the kernel boot line for using a VNC client from another machine to connect to the installation GUI. See Additional RHEL installation boot options on page 280 for more RHEL installation options. In SLES, you have the following options: Add a usessh=1 sshpassword=MIN_1_CHAR_PASS argument to the kernel boot line for starting a ssh installation. If you connect with ssh -X, you can have graphical GUI. Add a vnc=1 vncpassword=MIN_5_CHAR_PASS argument to the kernel boot line for using a vnclient from another machine to connect to the installation GUI.

Chapter 3. Operating system installation

31

See Additional SLES install boot options on page 282 for more SLES installation options.

3.2.4 Solaris Jumpstart

The Jumpstart services enables you to set up a configuration file through the rules file and build customization for each type, class, or group of servers, depending on how you classify your servers in your environment. Jumpstart has two basic components: A boot server An install server These components can either reside in the same server or on different servers. It all depends on your network. Normally, the boot server component of Jumpstart has to be in the same subnet as the client server being installed, but the install server can be anywhere in the network. In an enterprise environment, this might be a bit of an issue. Therefore, new functionalities were built into the Jumpstart procedure that allows the client server to boot from a server outside of its subnet. Jumpstart can now be configured to rely on the DHCP server to provide some boot parameters to the client server. If you have not had the chance to use this functionality, refer to the Supporting Solaris Network Installation with the DHCP Service (Task Map) section in the Solaris Installation Guide and the System Administration Guide: IP Services. On the client server, you can type the boot net - install dhcp command to tell the boot loader to enable the network and configure its IP settings from the DHCP, and that the DHCP will provide additional parameters that it will need to start the installation procedure.

3.2.5 Red Hat Kickstart

Kickstart is an installation configuration file used by the anaconda installation program. There is a wealth of information about anaconda on the Web. Red Hat System Administration Guide has a very detailed how-to as well. To access the Red Hat documentation, refer to:
http://www.redhat.com/docs/manuals/enterprise/

32

Solaris to Linux Migration: A Guide for System Administrators

The following list describes the basic actions Kickstart: It uses a configuration file for a specific server or a group of servers. See Example 3-3 on page 34 and Example 3-4 on page 35 for sample Kickstart files. This configuration file can be supplied in a bootable CD-ROM, bootable diskette, on a network through NFS, FTP, and HTTP, or from a USB flash drive. If you want to install from the network, you can either boot from the network if you have a PXE-enabled system and a network boot server configured, or you can boot from local media such CD, DVD, or USB flash and install the rest of the software from the network. You can use Kickstart for initial installs and upgrades. You can specify the Kickstart file location at the install GRUB prompt with the ks=URL option, or you can use DHCP to supply a Kickstart URL for your client with the ks option. The following sample command line invokes the installation using Kickstart from an NFS server. This line is entered at the GRUB prompt:
linux ks=nfs:/IP_ADDRESS_OF_NFS_SERVER/NFS_PATH/sample-ks.cfg

IP_ADDRESS_OF_NFS_SERVER is the IP address of the server that has the /NFS_PATH directory exported through NFS. That subdirectory contains the sample-ks.cfg file. There are additional steps required to make the network installation work (for example, copying the contents of the CDs or the ISO CD image files to source hard drive). For these instructions, see Red Hat Enterprise Linux 4 Installation Guide for x86, Itaniumtm, AMD64, and Intel@ Extended Memory 64 Technology (Intel@ EM64T) in the chapters Installing Red Hat Enterprise Linux and Performing a Network Installation. The Kickstart configuration file has many options. If used properly, you can do a fully automated, hands-free installation. If you happen to miss an item that the installation needs, it stops the installation at that portion and prompts you to enter what it needs. Items such as disk partitioning, LVM and RAID, software selection, and others can all be defined in the configuration file. You can generate a Kickstart file by using the Kickstart Configurator command system-config-kickstart. This is the graphical-based front end. Or, you can have it automatically generate a Kickstart file for you based on an existing server by issuing the following command:
# system-config-kickstart --generate <output-file-name>

Chapter 3. Operating system installation

33

When you use the Kickstart Configurator, you set up the following sections: Basic configuration information: Language support, keyboard, mouse, time zone, assign a root password, and so on. Installation method: New installation or upgrade, media source (CD, NFS, FTP, HTTP, or hard drive). Boot loader options: Install a new boot loader or not, set up the GRUB password, install MBR, and assign kernel parameters. Partition information: Set up disk partitioning information, use RAID, LVM, or both. Network configuration: Set up network device, use DHCP, static IP, or BOOTP. Authentication: Set up what method to use for user authentication, shadow file, NIS, LDAP, Kerberos 5, and so on. Firewall configuration: Enable or disable the firewall, set up trusted devices and services. Display configuration: Configure X Window System, set up the video card and monitor, set up the display resolution. Package selection: Select the groups of packages to install (can only select from groups and not individual packages). Preinstallation and post-installation script: Set up customized scripts to run before or after the installation process. There are differences between doing a manual Kickstart selection and a system-generated Kickstart file. See the two sample files in Example 3-3 and Example 3-4 on page 35.
Example 3-3 Sample manual Kickstart configuration file #Generated by Kickstart Configurator #platform=x86, AMD64, or Intel EM64T #System language lang en_US #Language modules to install langsupport en_US #System keyboard keyboard us #System mouse mouse #Sytem timezone timezone America/Edmonton #Root password rootpw --iscrypted $1$IRIk0Ucb$gYVbwzLZ62TsYX8IYEN.60

34

Solaris to Linux Migration: A Guide for System Administrators

#Reboot after installation reboot #Use text mode install text #Use interactive kickstart installation method interactive #Install OS instead of upgrade install #Use NFS installation Media nfs --server=9.3.5.11 --dir=/vm/rhel #System bootloader configuration bootloader --location=mbr #Clear the Master Boot Record zerombr yes #Partition clearing information clearpart --all --initlabel #Disk partitioning information part / --fstype ext3 --size 3000 part swap --size 512 #System authorization infomation auth --useshadow --enablemd5 #Network information network --bootproto=dhcp --device=eth0 #Firewall configuration firewall --disabled #XWindows configuration information xconfig --depth=8 --resolution=1024x768 --defaultdesktop=GNOME #Package install information %packages --resolvedeps #@ base-x #@ kde-desktop #@ editors #@ graphical-internet #@ server-cfg #@ admin-tools #@ system-tools #@ printing #@ compat-arch-support

Example 3-4 Sample system-generated Kickstart file from an existing server #Generated by Kickstart Configurator #platform=x86, AMD64, or Intel EM64T #System language lang en_US #Language modules to install langsupport en_US en --default=en_US

Chapter 3. Operating system installation

35

#System keyboard keyboard us #System mouse mouse #Sytem timezone timezone America/North_Dakota/Center #Root password rootpw --iscrypted $1$74WxlsHg$WHkNT3cVQjiNQA2u1.HrG. #Install OS instead of upgrade install #Use CDROM installation media cdrom #Clear the Master Boot Record zerombr yes #Partition clearing information clearpart --linux #Package install information %packages resolvedeps 4Suite Canna Canna-devel Canna-libs ElectricFence FreeWnn FreeWnn-devel FreeWnn-libs GConf2 GConf2-devel . . .(long list of packages cutout here) . . xorg-x11-xfs xpdf xrestop xsane xsane-gimp xscreensaver xsri xterm yelp yp-tools ypbind ypserv zip zisofs-tools zlib zlib-devel

36

Solaris to Linux Migration: A Guide for System Administrators

zsh zsh-html

A manually generated file enables you to customize the information. As for a system-generated one, it picks up only a few of the components. The following components are not gathered by the system-generated Kickstart file. The installation is defaulted to CD-ROM. You might want to change this if doing a network installation. Disk partitioning layout Boot loader setting Authentication method setting Network settings Firewall settings X Window System settings, if required Depending on what you want to achieve, you can combine the two methods to generate a more complete Kickstart file. If your target server will have the same basic software and package configuration as the current server, you can have the system generate a configuration file and have Kickstart open that file and you can customize all the missing components mentioned earlier. Refer to the Red Hat System Administration Guide for more information about the Kickstart installation.

3.2.6 SUSE AutoYaST

AutoYaST uses the same concepts as the Red Hat Kickstart. The following list describes some of its capabilities: The AutoYaST configuration file, called a control file, automatically controls either certain portions or the whole portion of the installation process. Installation source can be from CD/DVD, HDD, NFS, SMB, FTP, HTTP, and TFTP. AutoYaST can use a network PXE boot and DHCP to provide network installation information. AutoYaST is for initial installs and not for upgrades. You can specify at the install GRUB prompt the AutoYaST file location with the autoyast=URL option. AutoYaST can import Red Hat Kickstart files.

Chapter 3. Operating system installation

37

A sample command line to invoke the installation using AutoYaST from an NFS server follows. Enter this line at the GRUB prompt:
install=nfs://IP_ADDRESS_OF_NFS_SERVER/NFS_PATH/ autoyast=nfs://IP_ADDRESS_OF_NFS_SERVER/NFS_PATH/sample-autoyast.xml

Enter these two lines as one line. The IP_ADDRESS_OF_NFS_SERVER is the IP address of the server that has exported through NFS the /NFS_PATH directory. That subdirectory contains also the sample-autoyast.xml file. There are additional steps required to for the network installation, such as copying the contents of the CDs to the hard drive. You can use the yast2 instserver program to copy the install media to the server and configure the relevant services such as SLP, NFS, HTTP, and FTP. Refer to the SUSE Linux AutoYaST Automatic Linux Installation documentation for more information. The AutoYaST control file enables very granular control of the installation process. If used properly, you can do a fully automated, hands-free installation. If you happen to miss an item that the installation needs, it stops the installation at that portion and prompts you to enter what it needs. Items, such as disk partitioning, LVM and RAID, software selection, and others, can all be defined in the control file. You can generate an AutoYaST control file by using the YaST command yast2 autoyast (graphical-based) or yast autoyast (text-based). Unlike Red Hat Kickstart, AutoYaST does not provide a way to automatically generate a control file based on a currently running server. The control file enables you to have much more granular control of system options, that is, up to the /etc/sysconfig settings, as well as the settings for the installation itself. The following list of options are available to you: Software setup: Online update (if enabled): What time of day If download patches only or whole packages

Software packages: Which group of packages to install. After you have selected your choice from the following list, you also have the option to do a detailed selection of packages: Minimum system Minimum graphical system (without KDE) Full installation Default system

Hardware setup: Partitioning for hard drives, use RAID, LVM, or both.

38

Solaris to Linux Migration: A Guide for System Administrators

 Printer configuration: Use direct connects, CUPS, LPD-style, and so on. Sound card configuration. Graphics card and monitor configuration: Enable X Window System, 3D support, color depth, display resolution, and so on. System setup: Boot loader configuration: Set up GRUB, location, and sections. General options: Language support, time zone, hardware clock, keyboard, mouse, and so on. Report and logging options: Enable or disable logging of messages, warnings, and errors. Restore module: Restore files from an archive device or location as part of the installation process. Runlevel editor: Configure the default run level and what services to enable or disable for each run level. /etc/sysconfig editor: Preconfigure kernel values. Network device setup: Set up the type of network device and whether to use DHCP or static IP, set up host name and name server, routing information, and others. Network service setup: Configure DHCP server, DNS server, host names, HTTP server, LDAP server/client, NFS server/client, NIS server/client, Samba server/client, and others. Security and users setup: Configure CAs, certificates, firewalls, VPN, security settings, and create and edit users. Miscellaneous setup: Configure or preload customized application configuration files, set up custom preinstallation and post-installation scripts. When doing the partitioning, note that the default size it uses when you allocate a size for a partition is in kilobytes (KB). You can also enter mb, MB, gb, or GB to the end of the partition size. Example 3-5 shows a sample AutoYaST file.
Example 3-5 Sample AutoYaST file <?xml version="1.0"?> <!DOCTYPE profile SYSTEM "/usr/share/autoinstall/dtd/profile.dtd"> <profile xmlns="http://www.suse.com/1.0/yast2ns" xmlns:config="http://www.suse.c om/1.0/configns"> <configure> <networking>

Chapter 3. Operating system installation

39

<dns> <dhcp_hostname config:type="boolean">false</dhcp_hostname> <dhcp_resolv config:type="boolean">false</dhcp_resolv> <domain>labtest.com</domain> <hostname>sles9serv</hostname> <nameservers config:type="list"> <nameserver>9.3.5.2</nameserver> <nameserver>9.3.5.254</nameserver> </nameservers> <searchlist config:type="list"> <search>labtest.com</search> </searchlist> </dns> <interfaces config:type="list"> <interface> <bootproto>static</bootproto> <broadcast>9.3.5.255</broadcast> <device>eth0</device> <ipaddr>9.3.5.7</ipaddr> <netmask>255.255.255.0</netmask> <network>9.3.5.0</network> <startmode>onboot</startmode> </interface> </interfaces> <modules config:type="list"> <module_entry> <device>static-0</device> <module></module> <options></options> </module_entry> </modules> <routing> <ip_forward config:type="boolean">false</ip_forward> <routes config:type="list"> <route> <destination>default</destination> <device>-</device> <gateway>9.3.5.1</gateway> <netmask>-</netmask> </route> </routes> </routing> </networking> <printer> <cups_installation config:type="symbol">server</cups_installation> <default>local_lp0</default> <printcap config:type="list"> <printcap_entry> <ff config:type="boolean">false</ff>

40

Solaris to Linux Migration: A Guide for System Administrators

<info>Local parallel printer</info> <location>Local direct connection to LPT1</location> <manufacturer>HP</manufacturer> <model>HP LaserJet 5</model> <name>local_lp0</name> <nick>HP LaserJet 5 Foomatic/hpijs (recommended)</nick> <options config:type="list"/> <ppd_options config:type="list"> <ppd_option> <key>Copies</key> <value>1</value> </ppd_option> <ppd_option> <key>Duplex</key> <value>None</value> </ppd_option> <ppd_option> <key>Economode</key> <value>Off</value> </ppd_option> <ppd_option> <key>InputSlot</key> <value>Default</value> </ppd_option> <ppd_option> <key>MPTray</key> <value>First</value> </ppd_option> <ppd_option> <key>PageSize</key> <value>A4</value> </ppd_option> <ppd_option> <key>PrintoutMode</key> <value>Normal</value> </ppd_option> <ppd_option> <key>Quality</key> <value>FromPrintoutMode</value> </ppd_option> <ppd_option> <key>REt</key> <value>Medium</value> </ppd_option> <ppd_option> <key>TonerDensity</key> <value>5</value> </ppd_option> </ppd_options>

Chapter 3. Operating system installation

41

<raw config:type="boolean">false</raw> <uri>parallel:/dev/lp0</uri> </printcap_entry> </printcap> <server_hostname></server_hostname> <spooler>cups</spooler> </printer> <security> <console_shutdown>reboot</console_shutdown> <cracklib_dict_path>/usr/lib/cracklib_dict</cracklib_dict_path> <cwd_in_root_path>yes</cwd_in_root_path> <cwd_in_user_path>yes</cwd_in_user_path> <displaymanager_remote_access>no</displaymanager_remote_access> <enable_sysrq>no</enable_sysrq> <fail_delay>3</fail_delay> <faillog_enab>yes</faillog_enab> <gid_max>60000</gid_max> <gid_min>1000</gid_min> <kdm_shutdown>all</kdm_shutdown> <lastlog_enab>yes</lastlog_enab> <obscure_checks_enab>yes</obscure_checks_enab> <pass_max_days>99999</pass_max_days> <pass_max_len>7</pass_max_len> <pass_min_days>0</pass_min_days> <pass_min_len>5</pass_min_len> <pass_warn_age>7</pass_warn_age> <passwd_encryption>des</passwd_encryption> <passwd_use_cracklib>yes</passwd_use_cracklib> <permission_security>secure</permission_security> <run_updatedb_as>nobody</run_updatedb_as> <system_gid_max>499</system_gid_max> <system_gid_min>100</system_gid_min> <system_uid_max>499</system_uid_max> <system_uid_min>100</system_uid_min> <uid_max>60000</uid_max> <uid_min>500</uid_min> <useradd_cmd>/usr/sbin/useradd.local</useradd_cmd> <userdel_postcmd>/usr/sbin/userdel-post.local</userdel_postcmd> <userdel_precmd>/usr/sbin/userdel-pre.local</userdel_precmd> </security> <x11> <color_depth config:type="integer">16</color_depth> <configure_x11 config:type="boolean">true</configure_x11> <display_manager>kdm</display_manager> <enable_3d config:type="boolean">false</enable_3d> <monitor> <display> <frequency config:type="integer">60</frequency> <height config:type="integer">768</height>

42

Solaris to Linux Migration: A Guide for System Administrators

<width config:type="integer">1024</width> </display> <monitor_device>1024X768@60HZ</monitor_device> <monitor_vendor> LCD</monitor_vendor> </monitor> <resolution>1024x768</resolution> <window_manager>kde</window_manager> </x11> </configure> <install> <bootloader> <activate config:type="boolean">true</activate> <global config:type="list"> <global_entry> <key>color</key> <value>white/blue black/light-gray</value> </global_entry> <global_entry> <key>default</key> <value config:type="integer">0</value> </global_entry> <global_entry> <key>timeout</key> <value config:type="integer">8</value> </global_entry> <global_entry> <key>gfxmenu</key> <value>/boot/message</value> </global_entry> </global> <loader_device>/dev/sda</loader_device> <loader_type>grub</loader_type> <location>mbr</location> <repl_mbr config:type="boolean">false</repl_mbr> <sections config:type="list"> <section config:type="list"> <section_entry> <key>title</key> <value>Linux</value> </section_entry> <section_entry> <key>kernel</key> <value>/boot/vmlinuz root= vga=0x314 selinux=0 splash=silent resume= showopts elevator=cfq</value> </section_entry> <section_entry> <key>initrd</key> <value>/boot/initrd</value>

Chapter 3. Operating system installation

43

</section_entry> </section> <section config:type="list"> <section_entry> <key>title</key> <value>Hard Disk</value> </section_entry> <section_entry> <key>root</key> <value>(hd0)</value> </section_entry> <section_entry> <key>chainloader</key> <value>+1</value> </section_entry> </section> <section config:type="list"> <section_entry> <key>title</key> <value>Failsafe</value> </section_entry> <section_entry> <key>kernel</key> <value>/boot/vmlinuz root= showopts ide=nodma apm=off acpi=off vga=n ormal noresume selinux=0 barrier=off nosmp noapic maxcpus=0 </section_entry> <section_entry> <key>initrd</key> <value>/boot/initrd</value> </section_entry> </section> </sections> </bootloader> <general> <clock> <hwclock>localtime</hwclock> <timezone>US/Central</timezone> </clock> <keyboard> <keymap>english-us</keymap> </keyboard> <language>en_US</language> <mode> <confirm config:type="boolean">true</confirm> <forceboot config:type="boolean">true</forceboot> </mode> <mouse> <id>probe</id> 3</value>

44

Solaris to Linux Migration: A Guide for System Administrators

</mouse> </general> <partitioning config:type="list"> <drive> <device>/dev/hda</device> <initialize config:type="boolean">false</initialize> <partitions config:type="list"> <partition> <crypt>twofish256</crypt> <filesystem config:type="symbol">reiser</filesystem> <format config:type="boolean">true</format> <loop_fs config:type="boolean">false</loop_fs> <mount>/</mount> <partition_id config:type="integer">131</partition_id> <size>4000</size> </partition> <partition> <crypt>twofish256</crypt> <format config:type="boolean">false</format> <loop_fs config:type="boolean">false</loop_fs> <mount></mount> <partition_id config:type="integer">130</partition_id> <size>512</size> </partition> </partitions> <use>free</use> </drive> <drive> <device>/dev/hdb</device> <initialize config:type="boolean">false</initialize> <partitions config:type="list"> <partition> <crypt>twofish256</crypt> <format config:type="boolean">false</format> <loop_fs config:type="boolean">false</loop_fs> <mount>/var</mount> <partition_id config:type="integer">142</partition_id> <size>2048</size> </partition> <partition> <crypt>twofish256</crypt> <format config:type="boolean">false</format> <loop_fs config:type="boolean">false</loop_fs> <mount>/usr</mount> <partition_id config:type="integer">142</partition_id> <size>2048</size> </partition> <partition> <crypt>twofish256</crypt>

Chapter 3. Operating system installation

45

<format config:type="boolean">false</format> <loop_fs config:type="boolean">false</loop_fs> <mount>/home</mount> <partition_id config:type="integer">142</partition_id> <size>3000</size> </partition> </partitions> <use>free</use> </drive> </partitioning> <software> <addons config:type="list"> <addon>Base-System</addon> <addon>Basis-Devel</addon> <addon>Basis-Sound</addon> <addon>File-Server</addon> <addon>Gnome</addon> <addon>HA</addon> <addon>Kde-Desktop</addon> <addon>LAMP</addon> <addon>LSB</addon> <addon>Linux-Tools</addon> <addon>Mail_News-Server</addon> <addon>Print-Server</addon> <addon>SuSE-Documentation</addon> <addon>Various-Tools</addon> <addon>WBEM</addon> <addon>X11</addon> <addon>YaST2</addon> <addon>analyze</addon> <addon>auth</addon> <addon>dhcp_DNS-Server</addon> </addons> <base>Full-Installation</base> </software> </install> </profile>

For more information about how to use AutoYaST, refer to the Special Installation Procedure section in the Novel SUSE Administration Guide and AutoYaST Automatic Linux Installation and Configuration with YaST2. To access the SUSE documentation, refer to:
http://www.novell.com/documentation/sles9/

46

Solaris to Linux Migration: A Guide for System Administrators

3.2.7 The PXE protocol

PXE stands for Pre-boot Execution Environment. Is it a protocol designed by Intel that allows IA32 computers to boot using the network interface. PXE is available for newer network adapters that have execution codes stored in their ROM device. When the computer boots up, the BIOS loads the PXE ROM into the memory and runs it. A menu is then presented on the display and enables the user to boot the loaded operating system through the network. PXE relies on a DHCP or BOOTP server to provide an IP address and other network information. A TFTP server is also needed to supply the boot loader to the client. For basic functionality, you do not need a PXE server. Linux supports PXE for network booting using the syslinux package. The syslinux package is bundled in both RHEL and SLES. For more information about this topic, read 8.5, Network booting on page 135 and the following section.

3.3 Network-based install: Solaris and Linux heterogeneous environments

Depending on your migration plan, you might have an environment where you will need to configure a Solaris system to serve a Linux network installation, or a Linux system to serve a Solaris network installation.

3.3.1 Solaris system serving Linux network installation

To configure a Solaris system to serve a Linux network installation, you have to enable and configure the DHCP, TFTP, and NFS services on the Solaris server for booting and installing a Linux network client. Instead of NFS, you can also use FTP or HTTP protocols for a package distribution. You can set up the Solaris server to serve standard Linux network installs and also custom Kickstart or AutoYaST installs. There are no ready-to-use tools for this task, so you need to configure all these services by hand. But if you follow the Linux vendor instructions for setting up a Linux server for network installations, you will be able to replicate that scenario for a Solaris server. We provide a little more technical how-to information in this section, because this topic is not covered by any vendor documentation.

Solaris system serving RHEL network installation

Here is a short outline of the needed steps to configure a Solaris system to serve RHEL for IA32 network installations: 1. Get the pxelinux.0 file from PXELINUX tools at http://syslinux.zytor.com or from the RHEL syslinux package (/usr/lib/syslinux/pxelinux.0 in RHEL) and

Chapter 3. Operating system installation

47

place it in the TFTP share directory of the Solaris system, which by default is /tftpboot. This file is an IA32 PXE Linux loader that is needed for the network boot process. 2. Copy the vmlinuz and initrd.img files from the /isolinux directory found in the CD1 of RHEL to the same /tftpboot folder on the Solaris server. 3. Create the pxelinux.cfg configuration directory in the same place you put the pxelinux.0 file, and create a default file for the PXELINUX configuration. (See Example 3-6.)
Example 3-6 Sample default PXE configuration file default linux label linux kernel vmlinuz append initrd=initrd.img ramdisk_size=65536 ks=KS_URL

For more information about syntax, check the PXELINUX documentation, and look at the isolinux.cfg file from the same /isolinux directory on the CD1. 4. Edit /etc/inetd.conf file to enable the tftpd daemon if not already active, and send a HUP signal to inetd to reload the configuration if you modified it. 5. Configure and start the DHCP daemon in the normal way, and add the following standard DHCP options:
BootFile=pxelinux.0 BootSrvA=IP_address_of_your_Solaris_server

6. Copy the RHEL CD images into a local directory and share that directory through NFS. 7. Optionally, create a Kickstart file for the clients and place it in the NFS shared directory and reference it in the default configuration file of the pxelinux, for example:
ks=nfs:IP_OF_NFS_SERVER:/SHARE/ks.cfg

Now, you are ready to boot and install your Linux RHEL clients from this Solaris server. This example is for installing a similar configuration on all clients, but you can make a custom configuration for each separate client by adding a specific MAC-based or IP-based PXE configuration file for each client and then customize it by specifying the right Kickstart file for the installation configuration. For platforms other than IA32, the process is similar, but you will need to use another boot loader instead of pxelinux.0.

48

Solaris to Linux Migration: A Guide for System Administrators

Solaris system serving SLES network installation

Here is a short outline of the needed steps to configure a Solaris system to serve SLES for IA32 network installations: 1. Get the pxelinux.0 file from PXELINUX tools at http://syslinux.zytor.com or from the SLES syslinux package (/usr/share/syslinux/pxelinux.0 in SLES) and place it in the TFTP share directory of the Solaris system, which by default is /tftpboot. This file is an IA32 PXE Linux loader that is needed for the network boot process. 2. Copy the linux and initrd files from the /boot/loader directory found in the CD1 of SLES to the same /tftpboot folder on the Solaris server. 3. Create the pxelinux.cfg configuration directory in the same place you put the pxelinux.0 file, and create a default file for the PXELINUX configuration (see Example 3-7).
Example 3-7 Sample default PXE configuration file default linux label linux kernel linux append initrd=initrd ramdisk_size=65536 install=URL autoyast=URL

For more information about its syntax, check the PXELINUX documentation, and look at the isolinux.cfg file from the same /boot/loader directory on the CD1. 4. Edit /etc/inetd.conf to enable the tftpd daemon if not already active, and send a HUP signal to inetd to reload the configuration if you modified it. 5. Configure and start the DHCP daemon in the normal way, and add the following standard DHCP options:
BootFile=pxelinux.0 BootSrvA=IP_address_of_your_Solaris_server

6. Copy the CDs or DVD contents of the SLES distribution in a local directory, share that directory through NFS, and point the installation option to the right location, for example:
install=nfs://IP_OF_NFS_SERVER/NFS_SHARE/

7. Optionally, create a AutoYaST file for the clients and place it in a NFS shared directory and reference it in the default configuration file of the pxelinux, for example:
autoyast=nfs://IP_OF_NFS_SERVER/AUTOYAST_SHARE/autoyast.cfg

Now, you are ready to boot and install your Linux SLES clients from this Solaris server.

Chapter 3. Operating system installation

49

This example is for installing the same configuration on all clients, but you can make a custom configuration for each separate client by adding a specific MAC-based or IP-based PXE configuration file for each client and then customize it by specifying the right AutoYaST file for the installation configuration.

3.3.2 Linux system serving Solaris network installation

To configure a Linux system to serve a Solaris network installation, you have to decide which type of network initialization you will use: the default RARP+Bootparams, or the more streamlined DHCP. If you use the RARP+Bootparams services, you need to install, configure, and activate these services, because they are not normally used in Linux by default. If you choose the DHCP service, you need to configure and start this service and configure the Solaris client to ask for the DHCP service instead of RARP+Bootparams. Then, you need to configure the TFTP and NFS services for serving the kernel and the installation files.There are no ready to use tools for this task, so you need to configure all these services by hand. Our recommendation is to use the DHCP protocol instead of RARP+Bootparams, because all needed services are in the main distribution of RHEL and SLES.

50

Solaris to Linux Migration: A Guide for System Administrators

Chapter 4.

Disks and file systems

This chapter provides the information required to understand how to define, configure, and set up disks and file systems for use on Linux. This chapter describes the following topics: Disks and disk partitioning Disk-based file system management Virtual file systems Network File System (NFS) Swap file systems File system journaling File system organization AutoFSCK AutoFS Solaris Volume Manager to Linux LVM VERITAS VxVM and VxFS

Copyright IBM Corp. 2006. All rights reserved.

51

4.1 Disks and disk partitioning

In Solaris, partitions and slices are used interchangeably to mean the same thing: raw disk slices. Linux uses the term partitions, but some documentation also mentions slices when discussing the same thing.

4.1.1 Disks
There is no distinction made in this section as to the type of attachment (that is, direct, SAN, or other), as long as the operating system can recognize that it is attached to the server.

Task table
Table 4-1 illustrates device creation commands.
Table 4-1 Device creation commands Device creation task Create device files for new device Solaris devfsadm Linux udev

Disk recognition on Solaris

Before a disk can be sliced for use, the disk has to be recognized by the operating system. For new SCSI devices in the system, the sd.conf file in the /kernel/drv directory has to be set up to recognize a specific target and LUN number. In addition, a reconfiguration reboot has to be completed before the device will be recognized. If the sd.conf file already has the entry before the last reboot, the devfsadm command is run to generate the appropriate device files.

Disk recognition on Linux

Linux also requires that first the disk needs to be recognized by the operating system. This is done automatically by the kernel at the boot time if there are no hardware conflicts between disks (check for SCSI ID conflicts, or IDE master/slave conflicts). However, if hot plug is used to add new disks, we need to tell the kernel that a new disk is available for use. For more information about hot plug, refer to 6.1.4, Hot-plug devices on page 93. In RHEL and SLES, the device files are created every time a system boots by the udevd daemon, so a reconfiguration reboot is not necessary.

52

Solaris to Linux Migration: A Guide for System Administrators

Multipath
Both Red Hat and SUSE began supporting multipathing in their latest releases of RHEL and SLES. Refer to the vendor documentation and support channels for the latest updates about how to set up and use multipathing. SUSE has an online support page describing the setup tasks for SLES9, available at:
http://portal.suse.com/sdb/en/2005/04/sles_multipathing.html

See the following link for information about multipath support in RHEL4 Update 2:
https://rhn.redhat.com/errata/RHEA-2005-280.html

4.1.2 Disk partitions

In Solaris, slices and partitions are used interchangeably to mean the same thing. For Linux, most references use the word partition. Disk partitioning is a step that needs to be done before any new disk can be used. Each slice or partition will have to be created to the appropriate size as per your requirement.

Task table
Table 4-2 illustrates disk partitioning commands.
Table 4-2 Disk partitioning commands Disk partitoning task Create partitions on the disk Display partition table Maximum usable partitions Solaris format or fmthard prtvtoc 7 Linux fdisk or parted fdisk or parted 15 for SCSI 63 for IDE

Solaris
Use the format or fmthard commands to create the appropriate slice sizes in the disk. To display the partition table, use the prtvtoc command.

Linux
You can use the fdisk or parted commands to create the appropriate partition sizes in the disk. To display the partition table, use the fdisk -l command. Refer to the man pages for the fdisk and parted commands for more information.

Other considerations
In Solaris, you can create and use a maximum of seven slices per disk excluding the WholeDisk slice number 2.

Chapter 4. Disks and file systems

53

In Linux, we have primary partitions and logical partitions contained in (primary) extended partitions. Primary partitions and extended partitions can be created only in the Master Boot Record (MBR) of the disk and there is room for only four such partitions in MBR. Logical partitions are partitions created in extended partitions on the disk. The extended partition cannot be used directly, but only by means of the logical partitions created in it. To summarize, we can have up to four primary partitions, or one extended partition and up to three primary partitions defined in the MBR. Using either the fdisk or parted command, you only need to create primary and logical partitions, and they then make the extended partitions automatically. Another thing to note is that, in Linux, each partition has a partition type, which describes its intended usage. This type can be Linux, Linux LVM, Linux RAID, and so on. When creating a Linux partition, you decide what partition type to use. This depends on for what you intend to use the partition. If it is for data file system data storage, you will have to select ext2, ext3, Reiserfs, LVM, or one of the other options available. Refer to the man pages for the fdisk and parted commands for more information. The interactive fdisk command displays the available partitions types when you use the l subcommand. See Example 4-1 and Example 4-2 on page 55.
Example 4-1 Sample prtvtoc output # * * * * * * * * * * * * * * * * prtvtoc /dev/rdsk/c0t1d0s2 /dev/rdsk/c0t1d0s2 partition map Dimensions: 512 bytes/sector 424 sectors/track 24 tracks/cylinder 10176 sectors/cylinder 14089 cylinders 14087 accessible cylinders Flags: 1: unmountable 10: read-only First Sector Last Flags Sector Count Sector Mount Directory 00 1058304 20972736 22031039 / 01 0 1058304 1058303 01 0 143349312 143349311 00 22031040 121308096 143339135 /space 00 143339136 10176 143349311

Partition 0 1 2 3 7

Tag 2 3 5 8 0

54

Solaris to Linux Migration: A Guide for System Administrators

Example 4-2 Sample fdisk -l output # fdisk -l Disk /dev/sda: 9100 MB, 9100369920 bytes 255 heads, 63 sectors/track, 1106 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Device Boot /dev/sda1 /dev/sda2 * Start 1 65 End 64 1106 Blocks 514048+ 8369865 Id 82 83 System Linux swap Linux

4.2 Disk-based file system management

This section discusses the different methods to manage a disk-based file system.

Task table
Table 4-3 illustrates file system management commands.
Table 4-3 File system management commands File system management task To create a file system on disk Solaris newfs or mkfs Linux mkfs, mkfs.ext3, mk2fs -j, parted, or mkfs.reiserfs (SUSE only) fsck, fsck.ext3, e2fsck

Check file system for consistency

fsck

Overview of Solaris
Solaris natively only supports the UFS file system for hard disks. Use the newfs or the mkfs command to create the file system and the fsck command to check and fix the file system for consistency. The other types of disk-based file systems are HSFS (for CD-ROM), PCFS (for diskettes), and UDF (for DVD). Bad block management uses the command addbadsec to map out the bad sector.

Linux
In Linux, there are several disk file systems from which to choose. Depending on its intended use, one file system might be better than another. If you do not know which to choose, a safe guideline is to use the default file system of that distribution. On Red Hat, the default disk file system is ext3, and on SUSE, it is reiserfs, and both file systems are journaled.

Chapter 4. Disks and file systems

55

For creating a file system we have the following commands: mkfs.ext3 or mke2fs -j for creating ext3 file systems mkfs.reiserfs or mkreiserfs for creating reiserfs file systems (SUSE only) You can also still use the mkfs command, but you need to specify the file system type option with -t fstype. For checking the file system consistency, we have the following commands: fsck.ext3 or e2fsck for ext3 file systems fsck.reiserfs or reiserfsck for reiserfs file systems You can also still use the fsck -t fstype command. If you do not use the -t fstype option, the default file system will be used, ext3 for Red Hat and reiserfs for SUSE. Linux can also access UFS partitions, if needed for migration purposes. Table 4-4 shows the different disk file systems in Solaris and Linux.
Table 4-4 File system types File system type UNIX file system Solaris ufs Red Hat Source available but not compiled SUSE ufs (Refer to the mount commands man page for support statements.) reiserfs ext3 iso9660 udf msdos vfat

Reiser file system Ext3 file system CD-ROM file system DVD file system MS-DOS file system Windows 95 and later file system

N/A N/A hsfs udf pcsf pcfs

Source available but not compiled ext3 iso9660 udf msdos vfat

4.3 Virtual file systems

These types of file systems (except CacheFS) do not have a 1:1 corresponding disk slice used to store the data, but instead, use the system RAM

56

Solaris to Linux Migration: A Guide for System Administrators

to store information. Anything that gets stored in here will not be persistent across a reboot.

Overview of Solaris
There are several different types and uses for virtual file systems in Solaris: /tmp Based on the TMPFS file system. This file system is used for storing temporary files generated by the operating system and applications. When this space is fills up to a certain portion, the physical swap is used as the backing store in which to move these files. This space is also used as a buffer space for file system reads and writes in order to increase access time. A place to store temporary system files that are currently in use, but that are not required across a reboot. This space also provides access to special kernel information and facilities. Used for caching slow devices on local harddrive, such as simple or double-speed CD-ROMs or NFS over slow networks. A way for you to create a virtual file system by using a different path to access the files within that file system. /proc This file system contains all the active processes in the system by process number.

/var/run

CacheFS

Loopback file system

Linux
There are many options for virtual file system types in Linux. We only discuss a few of them that have a Solaris equivalent: /proc In Linux, this contains all the information about each process, information about the running kernel, devices, networking stack, and so on. See /proc on page 89 for more information. An equivalent in Linux does not exist. But for similar functionality, you can investigate the mount command with the --bind, --rbind, and --move options.

Loopback file system

Chapter 4. Disks and file systems

57

/tmp

Unlike Solaris, in Linux /tmp is a disk file system and is persistent across reboots. There is no /var/run equivalent. If you need a fast file system that will not be required to be persistent across a reboot, you can create a tmpfs or a ramfs. An equivalent to the Solaris CacheFS is not available in Linux.

CacheFS

4.4 Network File System (NFS)

Network File System (NFS) is available for both Solaris and Linux. You can use it in heterogeneous networks with both Solaris and Linux operating systems, with each of them being NFS server or client. For more information about NFS, refer to 7.8.7, NFS on page 121.

4.5 Swap file systems

Swap space is handled similarly to Solaris, in that a swap space is created at the time the disk is partitioned to handle paging from physical memory.

Task table
Table 4-4 on page 56 illustrates swap commands.
Table 4-5 Swap commands Swap task Create a swap space Enable a swap space Enable all swap spaces Disable a swap space Disable all swap spaces Delete a swap space Display the swap space usage Display the swap space status Set swap priority level Solaris swap -a swap -a N/A N/A N/A swap -d swap -s swap -l N/A Linux mkswap swapon swapon -a swapoff swapoff -a N/A swapon -s N/A swapon -p

58

Solaris to Linux Migration: A Guide for System Administrators

Overview of Solaris
swap is the only swap configuration command on Solaris. You can add, remove, and list the available swaps. Swap can be defined in dedicated disk slices, which is the preferred way, or as files within file systems, which is usually done for emergency and temporary purposes.

Linux
In Linux, we can also have swap defined in dedicated disk partitions, which is the preferred way, or as files within file systems, which is usually done for emergency and temporary purposes. In Linux, there are several commands for managing swap space: mkswap to format the swap space in order to be used by the system swapon to activate a swap space after it was formatted with mkswap swapoff to deactivate a swap space that is in use In addition, you can use the free command to see how much swap is used by the system.

4.6 File system journaling

File system journaling is a way to store transactions before they are written to the file system itself. After it is stored, a transaction can be applied to the file system at a later time. When a hard reboot is encountered, the operating system will only have to check and replay the journal log and will not be required to check the whole file system, thereby reducing the time it will take to perform the checks at boot up.

Overview of Solaris task

The SVM default is logging enabled for any UFS file system unless specifically disabled by itself due to available free space or specifically disabled by the system administrator. SVM logging is all internal to the mounted file system and does not make use of any external device to keep its log entries.

Linux task
The Linux default file systems for both RHEL and SLES have logging capabilities as well. You do not need any extra steps to activate the logging with the ext3 and reiserfs file systems. Both file systems use internal journaling for storing the

Chapter 4. Disks and file systems

59

transaction logs by default. It is possible to use an external device for journaling if needed. For journal mode information for the reiserfs and ext3 file systems, see the following IBM publications: Tuning Red Hat Enterprise Linux on IBM Eserver xSeries Servers, REDP-3861 Tuning SUSE LINUX Enterprise Server on IBM Eserver xSeries Servers, REDP-3862 The Linux ext3 default for journal mode is ordered. Meaning, all data is being forced to be written to the file system before the metadata has been committed to the journal. There are two other modes that ext3 supports, journal and writeback: Journal mode Writeback mode Commits the metadata into the journal before any the file system is written. This is the safest mode. Data ordering is not preserved. The data can be written to the file system after the metadata has been committed to the journal. This is considered to provide better performance than the previous two modes, and still provides high system integrity, but there is a tendency for older data in files to appear after the server has crashed and journal recovered.

Novell SUSE has two additional modes, user_xattr and acl: user_xattr acl Enables extended user attributes. Refer to the attr(5) man page for more information. Enables POSIX access control lists. Refer to the acl(5) man page for more information.

SUSE Linux also provides ReiserFS. ReiserFS is designed to provide improved features for disk space utilization, access performance, and crash recovery. For more information about ReiserFS, see SUSE Linux Enterprise Server 9 Administration and Installation Guide.

4.7 File system organization

The Linux file system organization is standardized in the Filesystem Hierarchy Standard (FHS, http://www.pathname.com/fhs/) and is laid out as follows: /bin Essential command binaries

60

Solaris to Linux Migration: A Guide for System Administrators

/boot /dev /etc /etc/opt /etc/X11 /home /lib /media /mnt /opt /proc /root /sbin /srv /tmp /usr /usr/bin /usr/include /usr/lib /usr/local /usr/sbin /usr/share /usr/X11R6 /var /var/account /var/cache /var/crash /var/lib /var/local /var/lock /var/log /var/opt

Static files of the boot loader Device files Host-specific system configuration Configuration files for add-on packages Configuration for X User home directories Essential shared libraries and kernel modules Mount point for removable media Mount point for mounting a file system temporarily Add-on application software packages Kernel and process information Roots home directory Essential system binaries Data for services provided by this system Temporary files Secondary hierarchy Most user commands Header files included by C programs Libraries Local hierarchy (empty after main installation) Non-vital system binaries Architecture-independent data X files Variable data Process accounting logs (optional) Application cache data System crash dumps (optional) Variable state information Variable data for /usr/local Lock files Log files and directories Variable data for /opt

Chapter 4. Disks and file systems

61

/var/mail /var/run /var/spool /var/tmp /var/yp

User mailbox files (optional) Data relevant to running processes Application spool data Temporary files preserved between system reboots Network Information Service (NIS) database files (optional)

4.8 AutoFSCK
In Linux, the ext2 and ext3 file system types have an option to force the checking of the consistency even if the file system was cleanly unmounted and is not marked as dirty. This is useful when you have an ext3 journaled file system that will never be dirty, but you want to check it from time to time just to make sure that there are no inconsistencies on it. You have two variables, one that counts the number of mounts, and other that is time based. Both of them can be manipulated with the tune2fs program. To disable this feature, you need to set both of these variables to 0. There is no equivalent functionality for this task in Solaris.

4.9 AutoFS
Also known as automount or automountd, this is a way to automatically mount a Network File System (NFS) when a specific directory path is being accessed by the user. When there is no activity on these subdirectories after a certain period of time, the automountd daemon unmounts the file system.

Task table
Table 4-6 illustrates the AutoFS information.
Table 4-6 AutoFS files AutoFS task Auto mounter daemon Master configuration file Other common configuration file Solaris automountd /etc/auto_master /etc/auto_home Linux automount /etc/auto.master /etc/auto.misc

62

Solaris to Linux Migration: A Guide for System Administrators

Solaris
The automount is configured in the /etc/auto_master (and its other corresponding) file and is activated by the startup script /etc/rc2.d/S74autofs, which runs during the server boot to run level 2 or 3. After automountd is running as a daemon, all NFS mounting and unmounting is automatically handled.

Linux
In Linux, the automount is configured in the /etc/auto.master file, as opposed to Solaris /etc/auto_master file, and is activated by the startup script /etc/init.d/autofs. After automountd is running as a daemon, all NFS mounting and unmounting is automatically handled.

Sample contents of configuration files

The following examples provide the sample contents of configuration files.
Example 4-3 Solaris /etc/auto_master # Master map for automounter # +auto_master /net -hosts #/home auto_home /xfn -xfn

-nosuid,nobrowse -nobrowse

Example 4-4 Solaris /etc/auto_home # Home directory map for automounter # +auto_home

Example 4-5 Linux /etc/auto.master # # $Id: auto.master,v 1.3 2003/09/29 08:22:35 raven Exp $ # # Sample auto.master file # This is an automounter map and it has the following format # key [ -mount-options-separated-by-comma ] location # For details of the format look at autofs(5). #/misc /etc/auto.misc --timeout=60 #/misc /etc/auto.misc #/net /etc/auto.net

Chapter 4. Disks and file systems

63

Example 4-6 Linux /etc/auto.misc # # # # # # cd # the following #linux #boot #floppy #floppy #e2floppy #jaz #removable

$Id: auto.misc,v 1.2 2003/09/29 08:22:35 raven Exp $ This is an automounter map and it has the following format key [ -mount-options-separated-by-comma ] location Details may be found in the autofs(5) manpage -fstype=iso9660,ro,nosuid,nodev :/dev/cdrom entries are samples to pique your imagination -ro,soft,intr ftp.example.org:/pub/linux -fstype=ext2 :/dev/hda1 -fstype=auto :/dev/fd0 -fstype=ext2 :/dev/fd0 -fstype=ext2 :/dev/fd0 -fstype=ext2 :/dev/sdc1 -fstype=ext2 :/dev/hdd

Example 4-7 Linux /etc/auto.net #!/bin/sh # $Id: auto.net,v 1.5 2003/09/29 08:22:35 raven Exp $ # Look at what a host is exporting to determine what we can mount. # This is very simple, but it appears to work surprisingly well key="$1" # add "nosymlink" here if you want to suppress symlinking local filesystems # add "nonstrict" to make it OK for some filesystems to not mount opts="-fstype=nfs,hard,intr,nodev,nosuid" # Showmount comes in a number of names and varieties. "showmount" is # typically an older version which accepts the '--no-headers' flag # but ignores it. "kshowmount" is the newer version installed with knfsd, # which both accepts and acts on the '--no-headers' flag. #SHOWMOUNT="kshowmount --no-headers -e $key" #SHOWMOUNT="showmount -e $key | tail +2" # Newer distributions get this right SHOWMOUNT="/usr/sbin/showmount --no-headers -e $key" $SHOWMOUNT | sort +0 | \ awk -v key="$key" -v opts="$opts" -- ' BEGIN { ORS=""; first=1 }

64

Solaris to Linux Migration: A Guide for System Administrators

{ if (first) { print opts; first=0 }; print " \\\n\t" $1, key ":" $1 } END ' { if (!first) print "\n"; else exit 1 }

4.10 Solaris Volume Manager to Linux LVM

The Sun Solstice DiskSuite (SDS) has been changed to Solaris Volume Manager (SVM) since the introduction of Solaris 9. The functionality and concept is still very much the same, but SVM introduces the new feature called soft partitions. All of the following references to SVM refer to both SDS and SVM. Where appropriate, a more detailed distinction between SDS and SVM is noted.

Task table
Table 4-7 illustrates volume management commands.
Table 4-7 Volume management commands Volume management task Initialize a disk for VM Solaris Create partition area/size metainit metainit volumename raidtype devices... N/A Linux Create partition area/size pvcreate mkraid, mdadm lvcreate LVname vgcreate VGname devicename mkraid, mdadm lvchange -a y LVname vgchange -a y VGname raidstart lvchange -a n LVname vgchange -a n VGname raidstop, mdadm lvremove LVname vgremove VGname lvextend LVname vgextend VGname newdevname lvreduce LVname vgreduce VGname devicename raidreconf

Create a volume or volume group (VG) Enable the volume or volume group Disable the volume or volume group Delete the volume or volume group Add a device to the volume or volume group Delete a device from the volume or volume group

N/A

metaclear metattach or metainit metadetach

Chapter 4. Disks and file systems

65

Volume management task Create a soft partition or logical volume (no RAID) Create a soft partition or logical volume (RAID 0) Create a soft partition or logical volume on a specific device

Solaris metainit -p metainit with RAID 0 on devices first, then metainit -p to create the soft partition volume Same as above, but the second metainit -p has the device name at the end of the command line metaclear metattach volume (size or device name) growfs

Linux lvcreate -Lsize -nLVname VGname lvcreate -iNumOfStripes -IStripeSize -nLVname VGname mdadm, mkraid Same as above, but add the device name to the end of the command line lvremove /dev/VGname/LVname raidreconf lvextend -Lsize /dev/VGname/LVname raidreconf Red Hat ext2/ext3: resize2fs SUSE ext2: resize2fs Red Hat reiser: resize_reiserfs Note: View the man pages for unmount requirement, and check if online version is available.

Delete a soft partition or logical volume Extend a volume or logical volume Extend a file system after volume has been grown

Reduce a volume or logical volume

metadetach Volname devicename

ext2: resize2fs, lvreduce reiser: resize_reiserfs, lvreduce

Note: View the man pages for unmount requirements.

The RAID and LVM2 commands installed on your system will be in either the /sbin or /usr/sbin directory. The packages that provide logical volume management are mdadm*, raidtools*, and lvm2*. Table 4-8 on page 67 presents a quick review of the equivalent commands.

66

Solaris to Linux Migration: A Guide for System Administrators

Table 4-8 Equivalent volume management commands Volume management task Set up or display metadb Display metadevice status Initialize raw devices to metadevices Attach metadevices Detach metadevices Clear and remove metadevices Replace a metadevice Rename a volume Check metadevice ID configuration Manage hot spares Set submirrors offline Set submirrors online Change volume parameters Back up volume group metadata Recover soft partition configuration information Set up root file system for mirroring Administer disk sets for sharing Resynchronize volume during reboot Expand a file system size Has entries that starts up the kernel metadevice modules Sets the default number of available volumes metarecover metaroot metaset metasync growfs /etc/system /kernel/drv/md.conf Solaris metadb metastat metainit metattach metadetach metaclear metareplace metarename metadevadm metahs metaoffline metaonline metaparam Linux vgdisplay, lvdisplay, lsraid vgdisplay, lvdisplay cat /proc/mdstat pvcreate vgchange, lvchange vgchange, lvchange vgremove, lvremove raidreconf raidreconf mdadm mdadm, raidhotadd, raidhotremove mdadm mdadm vgchange, lvchange vgcfgbackup vgcfgrestore N/A N/A N/A vgextend, lvextend, resize2fs, resize_reiserfs, raidreconf N/A /etc/sysconfig/lvm /etc/raidtab

Chapter 4. Disks and file systems

67

Solaris
SVM uses the state database to store a copy of the configuration. It requires a minimum of half + 1 copies, called a quorum, to be in good state in order to boot up the operating system. Prior to soft partitions, Solaris created a new file system the same size as the metadevice. So, if an SVM volume is a concatenation of two disks or a RAID 5 structure of four disks, the file system will be created to the maximum space allowed by it. With soft partitions, you can partition the whole SVM volume into smaller partitions as you see fit, and the file system that created on these soft partitions will be the size that you create. The creation and management of the SVM volumes uses some of the subcommands such as metainit, metattach, metastat, and metarename. Refer to the task table in Table 4-8 on page 67.

Linux
In Linux, the equivalent of SVM is LVM2, which is present in both Red Hat and SUSE. The basic concepts behind LVM for Linux are the same as in SVM. The sequence of operations are simplified: 1. Partition the disk. 2. Initialize partitions. 3. The disk can be used to create a new volume or volume group (VG), or added into an existing volume or volume group. 4. Logical volumes (LV) can be created or extended on this new disk. 5. The file system can be extended to this new disk. Note: The task table in Table 4-8 on page 67 lists the commands required to do the previous sequence. Refer to the man pages for each command for detailed information about other options available. For a good guide about LVM configuration and its use on Linux, see:
http://www.tldp.org/HOWTO/html_single/LVM-HOWTO/

Other Linux LVM tasks

Aside from the previously discussed normal volume management task, the Linux LVM has the additional following commands: lvmdiskscan lvscan pvdisplay Scan for devices visible to LVM2 Scan disks for logical volumes Display physical volume attributes

68

Solaris to Linux Migration: A Guide for System Administrators

pvs pvscan vgck vgscan vgimport vgexport

Report physical volume information Scan all supported logical volume block devices for physical disk Perform consistency check on the volume group Scan all disks for volume group data and rebuild caches Import a volume group to this server Export a volume group out of this server

LVM, by default, does not log errors and debug messages to a file. You can change this by editing the /etc/lvm/lvm.conf file under the log section.

Linux LVM sample configuration

Example 4-8 shows df output for a sample disk configuration. The root disk is being used as a simple raw device. The mount point /vm shows that it is using an LVM structure.
Example 4-8 Sample output of df # df -h Filesystem Size Used Avail Use% Mounted on /dev/sda1 9.7G 6.0G 3.2G 66% / none 2.0G 0 2.0G 0% /dev/shm /dev/sda6 1.4G 38M 1.3G 3% /home /dev/sda5 2.9G 37M 2.7G 2% /opt /dev/mapper/DataVG-LogVol00 100G 12G 84G 13% /vm

Example 4-9 shows vgdisplay output, without any parameters passed to it, for the volume group (VG). If you have multiple VGs, this command will include each one of them in the list output. To display only a specific VG, provide the VG name as a parameter to the command.
Example 4-9 Sample default output of vgdisplay with no options # vgdisplay --- Volume group --VG Name System ID Format Metadata Areas Metadata Sequence No VG Access VG Status MAX LV

DataVG lvm2 4 2 read/write resizable 0

Chapter 4. Disks and file systems

69

Cur LV Open LV Max PV Cur PV Act PV VG Size PE Size Total PE Alloc PE / Size Free PE / Size VG UUID

1 1 0 4 4 101.62 GB 32.00 MB 3252 3248 / 101.50 GB 4 / 128.00 MB egqMUy-dNyj-DSKG-zupE-V22E-cPHn-2cLLrn

Example 4-10 shows the output of the vgdisplay command in abbreviated format (using the -s option).
Example 4-10 Sample output of vgdisplay -s # vgdisplay -s "DataVG" 101.62 GB [101.50 GB used / 128.00 MB free]

Example 4-11 shows what a verbose display looks like. It also shows the information about the logical volumes (LVs) defined within the VG and the corresponding physical volumes (PVs) that are used for this VG.
Example 4-11 Sample output of vgdisplay -v # vgdisplay -v Finding all volume groups Finding volume group "DataVG" --- Volume group --VG Name DataVG System ID Format lvm2 Metadata Areas 4 Metadata Sequence No 2 VG Access read/write VG Status resizable MAX LV 0 Cur LV 1 Open LV 1 Max PV 0 Cur PV 4 Act PV 4 VG Size 101.62 GB PE Size 32.00 MB Total PE 3252 Alloc PE / Size 3248 / 101.50 GB

70

Solaris to Linux Migration: A Guide for System Administrators

Free PE / Size VG UUID

4 / 128.00 MB egqMUy-dNyj-DSKG-zupE-V22E-cPHn-2cLLrn

--- Logical volume --LV Name /dev/DataVG/LogVol00 VG Name DataVG LV UUID pmIJYM-OwVm-ZpqN-Z4dR-9TQl-Gtg0-xCU5kS LV Write Access read/write LV Status available # open 1 LV Size 101.50 GB Current LE 3248 Segments 4 Allocation inherit Read ahead sectors 0 Block device 253:0 --- Physical volumes PV Name PV UUID PV Status Total PE / Free PE PV Name PV UUID PV Status Total PE / Free PE PV Name PV UUID PV Status Total PE / Free PE PV Name PV UUID PV Status Total PE / Free PE --/dev/sdb1 6j6Ld4-tyMR-Swe5-oCpM-N38w-yTYZ-HTubDo allocatable 1084 / 0 /dev/sdc1 NU5VjA-N7LU-SagN-u91f-xAmI-ysbY-5AbvK7 allocatable 1084 / 0 /dev/sdd1 CtXfvq-yrmD-SVHb-l7x3-fKUh-pm5m-GysOF7 allocatable 542 / 0 /dev/sde1 2WJB4T-YTQX-ibjC-rOe5-HcBC-8QoE-ywFCiJ allocatable 542 / 4

Example 4-12 shows the information provided by the pvdisplay command for a specific physical disk partition. It gives a little bit more information than the previous output.
Example 4-12 Sample output of pvdisplay # pvdisplay /dev/sdc1 --- Physical volume --PV Name /dev/sdc1 VG Name DataVG

Chapter 4. Disks and file systems

71

PV Size Allocatable PE Size (KByte) Total PE Free PE Allocated PE PV UUID

33.88 GB / not usable 0 yes (but full) 32768 1084 0 1084 NU5VjA-N7LU-SagN-u91f-xAmI-ysbY-5AbvK7

Example 4-13 shows the information provided by the lvdisplay command for a logical volume.
Example 4-13 Sample default output of lvdisplay # lvdisplay /dev/DataVG/LogVol00 --- Logical volume --LV Name /dev/DataVG/LogVol00 VG Name DataVG LV UUID pmIJYM-OwVm-ZpqN-Z4dR-9TQl-Gtg0-xCU5kS LV Write Access read/write LV Status available # open 1 LV Size 101.50 GB Current LE 3248 Segments 4 Allocation inherit Read ahead sectors 0 Block device 253:0

Example 4-14 shows a lot more information for an LV, including which physical disk device partitions the LV resides in and how it is laid out. The sample LV shows that it uses a linear structure across the four PVs. Note that the logical extents numbering from the first PV to the last PV are in sequence. This is equivalent to the concatenation model for SVM. This shows the mapping of logical extents to physical volumes and physical extents.
Example 4-14 Sample output of lvdisplay --maps # lvdisplay --maps /dev/DataVG/LogVol00 --- Logical volume --LV Name /dev/DataVG/LogVol00 VG Name DataVG LV UUID pmIJYM-OwVm-ZpqN-Z4dR-9TQl-Gtg0-xCU5kS LV Write Access read/write LV Status available # open 1 LV Size 101.50 GB Current LE 3248

72

Solaris to Linux Migration: A Guide for System Administrators

Segments Allocation Read ahead sectors Block device

4 inherit 0 253:0

--- Segments --Logical extent 0 to 1083: Type linear Physical volume /dev/sdb1 Physical extents 0 to 1083 Logical extent 1084 to 2167: Type linear Physical volume /dev/sdc1 Physical extents 0 to 1083 Logical extent 2168 to 2709: Type linear Physical volume /dev/sdd1 Physical extents 0 to 541 Logical extent 2710 to 3247: Type linear Physical volume /dev/sde1 Physical extents 0 to 537

4.11 VERITAS VxVM and VxFS

There are VERITAS VxVM and VxFS products for Linux systems. You can also use the Linux LVM2, which is part of the software bundle for both Red Hat and SUSE. See 4.10, Solaris Volume Manager to Linux LVM on page 65. For more information about the VERITAS products (now Symantec), visit:
http://www.veritas.com/

Chapter 4. Disks and file systems

73

74

Solaris to Linux Migration: A Guide for System Administrators

Chapter 5.

Software management
This chapter describes how to manage software packages and patches in Linux. Software management, in this context, includes the operating system and any software that runs on the operating system. This chapter discusses the following topics: Packages Patching Dependencies Package distribution methods Automated software management Activating fixes after updating Compiling patches in Linux

Copyright IBM Corp. 2006. All rights reserved.

75

5.1 Packages
A package is a collection or group of program files and subdirectories that make up a software product. A package can be part of the operating system or an add-on software for specific functionality.

Task table
Table 5-1 illustrates package management commands.
Table 5-1 Package management commands Package management task Install packages Display installed packages Remove software package Upgrade/install a package Verify correct installation List the contents of an installed package Check which file to which a package belongs Check package information Solaris pkgadd pkginfo or pkgparam pkgrm pkgadd pkgchk Look in /var/sadm/install/contents Look in /var/sadm/install/contents Look in /var/sadm/pkg/PKGNAME/pkginfo Linux rpm -i rpm -qa rpm -e rpm -U rpm -V rpm -ql rpm -qf rpm -qi

5.1.1 Package management in Solaris

The Solaris package format is based on the System V Interface Definition for Application Binary Interface (ABI) and has tools for managing these packages. These management tools include the ability to add and remove packages, check for consistency, and display package information. If system administrators use the pkgadd and pkgrm tools to manage packages, the tools update the software product database accordingly.

5.1.2 Package management in Linux

Red Hat and Novell SUSE use the Red Hat Package Manager (RPM), which includes features for package management such as dependency and signature checking and other advanced options. For more information about RPM, go to the official site:
http://www.rpm.org

76

Solaris to Linux Migration: A Guide for System Administrators

Refer to the following Web sites for distribution-related information: Red Hat: Chapter 15, Package Management with RPM, in Red Hat Enterprise Linux 4: Reference Guide, available at:
http://www.redhat.com/docs/manuals/enterprise/

Novell SUSE: The RPM --- the Package Manager section in the Installation chapter of SUSE LINUX Enterprise Server 9 Administration Guide for OES, available at:
http://www.novell.com/documentation/sles9/

Example 5-1, Example 5-2, and Example 5-3 on page 78 provide sample output from using the rpm command.
Example 5-1 Sample output: List all packages in server # rpm -aq filesystem-2.3.0-1 bzip2-libs-1.0.2-13 gdbm-1.8.0-24 libsepol-1.1.1-2 bash-3.0-19.2 popt-1.9.1-9_nonptl vim-minimal-6.3.046-0.40E.4 gawk-3.1.3-10.1 openssl-0.9.7a-43.1 shadow-utils-4.0.3-41.1 cracklib-dicts-2.7-29 .... mozilla-nspr-1.7.7-1.4.2 intltool-0.31.2-1 libcroco-0.6.0-4 vim-enhanced-6.3.046-0.40E.4 gail-1.8.0-2 metacity-2.8.6-2.1

Example 5-2 Sample output: List files in a particular package # rpm -ql GConf2-2.8.1-1 /etc/gconf/2 /etc/gconf/2/path /etc/gconf/gconf.xml.defaults /etc/gconf/gconf.xml.mandatory /usr/bin/gconf-merge-tree /usr/bin/gconftool-2 /usr/lib/GConf /usr/lib/GConf/2 /usr/lib/GConf/2/libgconfbackend-oldxml.so

Chapter 5. Software management

77

/usr/lib/GConf/2/libgconfbackend-xml.so .... /usr/share/locale/yi/LC_MESSAGES/GConf2.mo /usr/share/locale/zh_CN/LC_MESSAGES/GConf2.mo /usr/share/locale/zh_TW/LC_MESSAGES/GConf2.mo /usr/share/man/man1/gconftool-2.1.gz /usr/share/sgml /usr/share/sgml/gconf /usr/share/sgml/gconf/gconf-1.0.dtd

Example 5-3 Sample output: Package information # rpm -qi GConf2-2.8.1-1 Name : GConf2 Relocations: (not relocatable) Version : 2.8.1 Vendor: Red Hat, Inc. Release : 1 Build Date: Tue 12 Oct 2004 12:16:24 PM CDT Install Date: Thu 13 Oct 2005 03:15:55 PM CDT Build Host: tweety.build.redhat.com Group : System Environment/Base Source RPM: GConf2-2.8.1-1.src.rpm Size : 4050741 License: LGPL Signature : DSA/SHA1, Wed 05 Jan 2005 03:47:04 PM CST, Key ID 219180cddb42a60e Packager : Red Hat, Inc. <http://bugzilla.redhat.com/bugzilla> URL : http://www.gnome.org Summary : A process-transparent configuration system Description : GConf is a process-transparent configuration database API used to store user preferences. It has pluggable backends and features to support workgroup administration.

5.2 Patching
A patch is a subset of a package that adds new functionality to a package or fixes a problem with the installed package.

78

Solaris to Linux Migration: A Guide for System Administrators

Task table
Table 5-2 illustrates patch management commands.
Table 5-2 Patch management commands Patch management task Install a patch Remove a patch Display installed patches Solaris patchadd patchrm showrev -p Linux rpm -F (upgrades only if already present) N/A N/A

5.2.1 Patching in Solaris

Patches in Solaris are handled using the following commands: patchadd, patchrm, and showrev. The last two digits of the patch file name show the revision level.

5.2.2 Patching in Linux

In comparison to Solaris, Linux manages software updates and fixes in a different way. Instead of creating patches, vendors create full packages of fixed software to simplify package maintenance. If a system administrator needs to install a new package, the process only requires installing the latest version. This method is in contrast to the process of requiring the installation of the base version package first, followed by all the associated patches. For RPM-based distributions a system administrator can use the rpm -U command to upgrade a package in Linux. In addition, administrators can use the rpm -F command to update only if the package is already present. Note: The concept of a patch differs in Linux as compared to Solaris. A patch in Linux refers to a text file that contains code that must be applied to a preexisting source code or script file. Patching source code files also requires the user to recompile the binaries in order to benefit from the update. See 5.7, Compiling patches in Linux on page 82 for more information.

5.3 Dependencies
In certain situations, a package or patch will have a dependency on one or more packages or patches before it can be installed.

Chapter 5. Software management

79

5.3.1 Dependency management in Solaris

A Solaris system administrator has several options available for resolving package dependencies: Manual (that is, an individual package or patch). A patch cluster, which already contains most of what is needed for the core operating system. An automated patch manager (see 5.5, Automated software management on page 81). Each of these options has certain considerations that need to be taken into account according to the needs of the system administrator. For example, to target a specific issue, only downloading the required patches and its prerequisites is required. However, consider a cluster patch when there is a need to ensure that the key components of the operating system are all up to date.

5.3.2 Dependency management in Linux

In an RPM-based distribution of Linux, you can use two methods for resolving package dependencies: 1. Manual: Using the rpm -i command to install a package provides information about what dependent packages are missing, if any. 2. Automatic: Enterprise Linux distribution vendors typically provide tools for automatic dependency management. Some examples include: Red Hat: up2date Novell SUSE: you Refer to 5.5.2, Automated software management in Linux on page 81 for more information.

5.4 Package distribution methods

Solaris has the SunSolve Web site for retrieving distributed packages and patches. Red Hat and Novell also provide facilities for package distribution: Red Hat: The Red Hat Network provides package updates and files:
https://rhn.redhat.com

Novell SUSE: The SUSE Linux portal provides package updates and fixes (requires registration):
https://portal.suse.com

80

Solaris to Linux Migration: A Guide for System Administrators

5.5 Automated software management

Automated software management is an easy way to update a system without having to individually check through each patch or package for dependencies and requirements.

5.5.1 Automated software management in Solaris

There are several options available from the SunSolve Web site for automated software management. The software in this context refers to patches only and not packages. Those options include: Patch Manager for Solaris 8 and 9 PatchPro Interactive PatchPro Expert These tools generate a complete list of patches based on your current system, which includes patches for software components that are installed on the system but are not being used. Some of these tools enable you to pick from the generated list so that you can target only specific software components.

5.5.2 Automated software management in Linux

In Linux, each distributor has their own way of performing automated software management. With both systems, you can configure a default policy for software upgrades and have a daemon perform the updates in the background based on the schedule defined in the policy.

Red Hat
The default tool for automated dependency and update management is up2date. This tool connects to the Red Hat Network and checks for available updates for your system. After up2date gathers a list of updates, it presents you with the option to select which packages to download and update. After the selections step, the tool completes the remaining installation and update steps. For more information, refer to the Red Hat Network 4.0 Reference Guide, available at:
http://www.redhat.com/docs/manuals/RHNetwork/

Novell SUSE
The default tool for automated dependency and update management is you (YaST online update), a YaST2 module that has a similar functionality to the Red Hat up2date tool. For more information, see the YaST Online Update section in

Chapter 5. Software management

81

the Installation chapter of SUSE LINUX Enterprise Server 9 Administration Guide for OES, available at:
http://www.novell.com/documentation/sles9/

5.6 Activating fixes after updating

Certain packages and patches require a reboot to activate the update or fix, especially for kernel and library patches.

5.6.1 Patch activation in Solaris

Occasionally, a system will have to regenerate its devices during the reboot if the fix is for driver components. In most kernel or library patches, Sun recommends installing in single user mode. This allows the system to be immediately rebooted after applying the update, while minimizing the potential for a loss of user data.

5.6.2 Patch activation in Linux

In Linux, there is no such thing as reconfiguration reboot or device regeneration. Sometimes it is possible to update kernel modules without rebooting the system by removing the affected module from kernel, updating the kernel module, and reinserting the new kernel module. For information about this, seeInstalling device driver modules on page 92. In addition, you can update the system with a new kernel while it is running and wait to boot the new kernel during a maintenance window. The system continues to run with the old kernel, without interruption, until the system is rebooted.

5.7 Compiling patches in Linux

For a Solaris system administrator, there is no need to compile the kernel, driver, or any of its core tools because there is no source code available in all versions of Solaris before version 10. Compilation is usually done when GNU freeware is required and has to be installed. Even in this situation, there are usually downloadable precompiled binaries. In Linux, it is usually possible to compile programs because the majority of the applications come with source code. All the tools necessary for compiling, assembling, and linking are usually present in the operating system. If not, then those tools can be installed. Some driver or third-party applications might require compilation to allow them to work with the systems kernel. Most source code and third-party applications

82

Solaris to Linux Migration: A Guide for System Administrators

come with instructions about how to compile and install. If you still need more help, you can contact the provider of the source or third-party application to guide you through this process. The following tutorials are available in the open source section of the IBM developerWorks Web site:
http://www.ibm.com/developerworks/opensource

Compiling sources and managing packages, available at:

http://www.ibm.com/developerworks/edu/l-dw-linux-lpir25-i.html

Configuring and compiling the kernel, available at:

http://www.ibm.com/developerworks/edu/l-dw-linux-lpir26-i.html

Chapter 5. Software management

83

84

Solaris to Linux Migration: A Guide for System Administrators

Chapter 6.

Device management
This chapter provides a description of common tasks for managing devices in Linux for the Solaris system administrator. This chapter includes the following topics: Device access and configuration Removable media devices Terminals and modems Distribution-based device management tools

Copyright IBM Corp. 2006. All rights reserved.

85

6.1 Device access and configuration

This section outlines the specific tasks needed to access and configure devices in Linux. We describe the following topics: Device naming and access Displaying device configuration information Adding a device Hot-plug devices

Task table
Table 6-1 illustrates device management commands.
Table 6-1 Solaris versus Linux device management commands Device management task Run multiple tasks in a GUI environment Solaris admintool Red Hat system-config-* (keyboard, mouse, printer, network, soundcard, and so on) /dev/MAKEDEV kudzu mknod kudzu /dev/MAKEDEV ls /sys/devices cat /proc/devices insmod lsmod rmmod SUSE yast2

Configure a device Define a device Remove a device List devices Install a device driver kernel module List installed device driver kernel modules Remove a device driver kernel module

devfsadm devfsadm devfsadm -C or -c sysdef or prtconf modload modinfo modunload

/dev/MAKEDEV hwinfo mknod hwinfo /dev/MAKEDEV ls /sys/devices cat /proc/devices insmod lsmod rmmod

6.1.1 Device naming and access

Linux is similar to Solaris in the way it represents devices on the system by using logical and physical device files.

86

Solaris to Linux Migration: A Guide for System Administrators

Table 6-2 Solaris versus Linux device file location Solaris /dev /devices devfsadm Linux /dev /sys/devices udev Description Contains logical device files Contains physical device files Command that creates and manages the physical device files and the associated logical device files

Solaris and Linux both use the logical device names in the /dev file system for many of the disk-related system administration tasks. However, the file and directory naming convention differs between the two operating systems.

Solaris logical disk devices

Solaris administrators refer to disk devices by selecting the subdirectory that it is linked to (either /dev/rdsk or /dev/dsk), followed by a string of information that indicates the specific controller, disk, and slice:
/dev/dsk/c[1]t[2]d[3]s[4]

Where: 1: Logical controller number 2: Physical bus target number 3: Driver number 4: Slice or partition number

Linux logical disk devices

The Linux kernel represents disk devices as a pair of numbers <major>:<minor>. Some major numbers are reserved for particular device drivers, others are dynamically assigned to a device driver when Linux boots. A major number can also be shared by multiple device drivers. The device driver uses the minor number to distinguish individual physical or logical devices. Device drivers assign device names to their devices, according to a device driver-specific naming scheme. For example: The first SCSI disk is named /dev/sda, the second disk is named /dev/sdb, the third is /dev/sdc, and so on. After /dev/sdz is reached, the cycle starts over with /dev/sdaa, followed by /dev/sdab, and so on. The first floppy drive is named /dev/fd0. The first SCSI CD-ROM is named /dev/scd0, which is also known as /dev/sr0. The master disk on the IDE primary controller is named /dev/hda and the slave disks is named /dev/hdb, while the master and slave disks of the secondary controller are /dev/hdc and /dev/hdd.

Chapter 6. Device management

87

The output in Figure 6-1 is an example of what a Linux system administrator might see when performing an ls -l command on /dev/hda1.

Figure 6-1 Example of IDE device in Linux

The previous listing represents the first partition of the first drive on the first IDE controller. Linux defines device files with either type c for character devices, or type b for block devices. All disks, such as the previous one, are represented as block devices in Linux. For an example of a character device, refer to 6.3.3, Serial port management on page 103. This next output is an example of what a Linux system administrator might see when performing an ls -l command on /dev/sda1, as shown in Figure 6-2.

Figure 6-2 Example of SCSI device in Linux

For more information about device names and numbers, the Linux Assigned Names And Numbers Authority (LANANA) maintains the Linux Device List. This is the official listing of device numbers and /dev directory nodes for Linux, available at:
http://www.lanana.org/

88

Solaris to Linux Migration: A Guide for System Administrators

Accessing devices
In Solaris, there is the concept of having a raw device logical interface name (rdsk) and a block device logical interface name (dsk) for the same device. The name a system administrator uses depends on the task. In Linux, there is only one type of logical device name scheme used. Table 6-3 lists a few Solaris versus Linux commonly used commands for accessing devices.
Table 6-3 Commonly used logical device access commands in Solaris and Linux Solaris example df /dev/dsk/c1t0d0s0 fsck -p /dev/rdsk/c1t3d0s4 mount /dev/dsk/c1t1d0s0 /mnt/1 newfs /dev/rdsk/c0t0d1s1 prtvtoc /dev/dsk/c1t1d0s0 Linux example df /dev/sda1 fsck -y /dev/sda2 mount /dev/hda2 /data mkfs /dev/sdc1 fdisk -l /dev/hdc

Refer to the man pages for these commands for more information.

6.1.2 Displaying device configuration information

In Solaris, there are three commands for displaying system and device configuration information: prtconf sysdef dmesg There are several methods to obtain similar device-related information in Linux.

/proc
The Linux /proc virtual file system differs from the Solaris implementation in that besides having information about running processes, it also contains runtime system information about areas such as devices mounted, system memory, and hardware configuration. Table 6-4 on page 90 outlines some of the files and directories that contain device-related information.

Chapter 6. Device management

89

Table 6-4 Device-related information in the /proc file system Location /proc/bus/ /proc/devices /proc/ide /proc/scsi Description Directory containing bus-specific information File that lists block and character device drivers configured into the current running kernel Directory containing information about all IDE devices known to the kernel Directory containing information about all SCSI devices known to the kernel

Red Hat and Novell SUSE also provide additional information about the /proc file system: In Red Hat, refer to Chapter 5, The /proc File System, in Red Hat Enterprise Linux 4: Reference Guide, available at:
http://www.redhat.com/docs/manuals/enterprise/

In Novell SUSE, refer to the The /proc File System, section of the Administration chapter of SUSE LINUX Enterprise Server 9 Administration Guide for OES, available at:
http://www.novell.com/documentation/sles9/

Important: Although some device driver information can be found in /proc, using it for this purpose is deprecated in recent Linux distributions, and the sysfs virtual file system (mounted under /sys) should be used instead.

/sys
The sysfs virtual file system provides a hierarchical view in the user space of kernel data structures, their attributes, and the links between them. The hotplug command notifies user space when a device is added or removed, keeping the file system dynamic. Features include: sysfs is linked to the kobject infrastructure. Objects in the kernel are represented in a directory tree. Each kernel object is represented as a subdirectory in the tree. sysfs forwards file I/O operations to methods defined for the attributes, providing a means to read and write kernel attributes. sysfs enables a user to determine which device has been assigned to which device file by the kernel. Previously, this information was not easily obtainable. sysfs can be mounted just like other virtual file systems. Only cat(1) and echo(1) are needed, but tree(1) is also good to have for viewing the overall structure.

90

Solaris to Linux Migration: A Guide for System Administrators

Table 6-5 outlines some of the directories that contain device-related information.
Table 6-5 Device-related information in the /sys file system Location /sys/block/ /sys/bus/ /sys/class/ Description Has subdirectory for each block device discovered in the system. Has subdirectories for each supported physical bus type. Every device class that is registered with the kernel is represented here. An example of a class is sound or printer. A hierarchical structure that has every physical device discovered by the bus types seen by the system.

/sys/device/

For an in-depth look at sysfs, refer to The sysfs file system, by Patrick Mochel, presented in volume 1 of the 2005 Ottawa Linux Symposium proceedings, available at:
http://www.linuxsymposium.org/2005/

The lspci and lsusb commands

The lspci and lsusb commands display information about all PCI and USB buses in the system and all devices connected to them. For more information and options about these commands, refer to the man page.

The dmesg command

This command essentially performs the same function as it does in Solaris. For more information about this command, refer to the man page. Red Hat and Novell SUSE also provide their own files and commands for obtaining device information: Red Hat: /etc/sysconfig/hwconf: This file contains information about all the hardware detected on the system, which includes the drivers used, vendor ID, and device ID information. For more information, refer to Chapter 4, The sysconfig Directory, in Red Hat Enterprise Linux 4: Reference Guide, available at:
http://www.redhat.com/docs/manuals/enterprise/

Chapter 6. Device management

91

 hwbrowser: This command invokes an X Window System graphical user interface that enables you to browse the hardware configurations. For more information, refer to Section 4 in Chapter 40, Gathering System Information, in Red Hat Enterprise Linux 4: System Administration Guide, available at:
http://www.redhat.com/docs/manuals/enterprise/

Novell SUSE: hwinfo: This command probes for the hardware present in the system and can generate a system overview log. Refer to the man page and the readme file located under /usr/share/doc/packages/hwinfo for more information about this command.

6.1.3 Adding a device

The process of adding a device (excluding hot-pluggable) involves shutting down the system, physically connecting the device, and powering the system back up. If the device contains its own power source, make sure that the power is turned on before powering up the system.

Device driver modules

The Linux operating system uses kernel modules for device drivers in a manner similar to that used in Solaris. Most device driver modules are located under /kernel/drivers in the release directory within the /lib/modules/ file system. The specific release level is determined by issuing the uname -r command. For example, a Red Hat Enterprise Linux system might return the following output, which means /lib/modules/2.6.9-11.ELsmp/kernel/drivers contains the loadable device driver modules for the system:
# uname -r 2.6.9-11.ELsmp

Automatic device configuration

Red Hat and Novell SUSE both support automatic device configuration at boot time. This means that the system will recognize most devices and make them available after powering up the system.

Installing device driver modules

Occasionally, a system administrator might need to update or install new device driver modules. In these circumstances, follow the instructions provided by the device manufacturer or module creator. Refer to Table 6-1 on page 86 for module-related tasks in Solaris and their associations in Linux. Table 6-6 on page 93 lists additional module-related commands. Refer to the commands man page for information about its use.

92

Solaris to Linux Migration: A Guide for System Administrators

Table 6-6 Additional module-related commands Module-related task Add and remove modules based on the system configuration Display information about a kernel module Command modprobe modinfo

6.1.4 Hot-plug devices

Hot-plugging is the adding and removing of devices while the system is running. Linux includes hot-plug support for the Small Computer System Interface (SCSI), Peripheral Component Interconnect (PCI), and Universal Serial Bus (USB), devices just as Sun Solaris, as well as additional devices such as IEEE 1394 (Firewire) support. While Solaris requires the use of the cfgadm command to manage hot-plug devices, a Linux system automatically recognizes and configures a supported hot-pluggable device when connected. Conversely, the system automatically deconfigures a hot-pluggable device when removed. Linux provides the hotplug and udev commands for managing hot-plug devices, and the man pages contain instructions about using both of these commands. Red Hat and Novell SUSE also provide additional information about Linux hot-plugging: Red Hat: Refer to 7.1.1, Hotplug Devices, in the Red Hat Enterprise Linux 3: System Administration Guide, available at:
http://www.redhat.com/docs/manuals/enterprise/

Novell SUSE: Refer to the The Hotplug System section in SUSE LINUX Enterprise Server 9 Administration Guide for OES, available at:
http://www.novell.com/documentation/sles9/

Background information
For general information about Linux hot-plugging, refer to the project Web site at:
http://linux-hotplug.sourceforge.net/

Also refer to the following Ottawa Linux Symposium (OLS) presentations by Greg Kroah-Hartman: OLS 2001: Hotpluggable devices and the Linux Kernel OLS 2003: udev A Userspace Implementation of devfs The presentations are available at:
http://www.kroah.com/linux/

Chapter 6. Device management

93

6.2 Removable media devices

This section covers the following topics: Supported types Managing and accessing removable media Formatting removable media CDs and DVDs Tape drives

6.2.1 Supported types

Linux fully supports many types of removable media, ATAPI, SCSI, and USB connected devices, such as: DVD-ROM/R/RW CD-ROM/R/RW Iomega Zip and Jaz drives Diskettes Tapes The locations of the device mount points differ from the Solaris operating system, and Table 6-7 lists information for accessing some common removable devices.
Table 6-7 How to access common removable media devices What to access Files on the first floppy disk Files on the second floppy disk Files on the first CD or DVD Files on the second CD or DVD Where to access /media/floppy /media/floppy1 /media/cdrom /media/cdrom1

For questions about Linux support for a specific device, contact the Linux distributor or the device manufacturer.

6.2.2 Managing and accessing removable media

Methods implemented in the operating system for automatic mounting of removable devices vary between Solaris and Linux, and also between different Linux distributions. In addition, depending on which graphical session manager

94

Solaris to Linux Migration: A Guide for System Administrators

you optionally choose to run on top of the operating system, the types of device management and access tools available in those session managers vary as well.

Solaris
Solaris uses the volume manager (vold) to actively manage removable media devices. The volume manager simplifies the process for using removable media by automatically performing many of the tasks required to mount and access removable media (on earlier versions of Solaris).

Red Hat
Red Hat Enterprise Linux semi-automatically manages removable media devices. Most removable devices are auto-detected for Nautilus File Manager users, who will see a desktop link to the device upon startup. Shell console users will have to manually mount and unmount removable devices. For more information about accessing removable devices, refer to Chapter 13, Diskettes and CD, in Red Hat Enterprise Linux 4: Red Hat Enterprise Linux Step By Step Guide, available at:
http://www.redhat.com/docs/manuals/enterprise/

Novell SUSE
Novell SUSE Linux Enterprise System operating system automatically manages removable media devices using the submount file system (subfs). The submount program automatically mounts removable media devices when the mount point is accessed, and subsequently unmounts them when they are no longer in use. Issuing the cd /media/<device> command initiates the automatic mounting process. Note: Media cannot be ejected as long as it is being accessed by a program. For more information about submount, consult the man pages or the project Web site at:
http://submount.sourceforge.net/

Access
System administrators access files on mounted removable media devices using the same commands as they do when accessing other file systems. Commands such as ls, cp, rm, and mv behave the same way on removable media, with CD/DVD and tape media being the exception (see 6.2.4, CDs and DVDs on page 96 and 6.2.5, Tape drives on page 98).

Chapter 6. Device management

95

Remote systems
The process of allowing remote systems to access removable media file systems is no different than it is for standard file systems. You can export the removable media file system mount points using Network File System (NFS), Samba, or any other supported networked file system protocols. In addition, accessing a remote removable media file system is the same process as it is for standard remote file systems.

The eject command

The eject command operates in the same manner in Linux as it does in Solaris. For details about its usage options, refer to the man page.

The mzip command

The mzip command is part of the mtools package that enables you to issue Zip and Jaz disk-specific commands on Linux. Refer to the man pages about mzip and mtools for more information.

6.2.3 Formatting removable media

Linux does not have an equivalent command to Solariss rmformat. Therefore, superuser access must be attained in order to format removable media from the command line. The standard file system commands, such as fdisk, fsck, and mkfs, are applicable to most removable media devices (see 6.2.4, CDs and DVDs on page 96 and 6.2.5, Tape drives on page 98 for exceptions).

The fdformat command

The fdformat command does a low-level format of the floppy diskette. Consult the man page for further information.

The gfloppy command

This is a GUI desktop interface for formatting floppy disks. Consult the man page for further information.

6.2.4 CDs and DVDs

This section provides an overview of commands and tools for accessing, writing, and erasing CD and DVD optical media.

Accessing
When mounted, both CD and DVD media can be accessed using standard file system commands, with the exception of commands that involve the writing and erasing of data, that is, rm, cp, and mv.

96

Solaris to Linux Migration: A Guide for System Administrators

Writing
Both Red Hat and SUSE provide graphical tools for creating audio and data CDs through their desktop applications: Red Hat The Nautilus File Manager has an integrated tool called the CD/DVD Creator. Novell SUSE The K3b application is provided for CD burning. For more information about this application, go to:
http://k3b.sourceforge.net/

The YaST tool provides a CD creator module. For more information about this application, run the yast2 cd-creator command as root. If you do not want to use one of the previous tools, important command line tools for managing this process include: cdrecord: This command is similar to cdrw on Solaris and is for writing CD file systems. Linux and cdrecord support the ISO 9660 format with Rock Ridge or Joliet extensions. For more information about other supported formats and instructions for using cdrecord, refer to the man page or the project Web site at:
http://freshmeat.net/projects/cdrecord/

mkisofs: This command is the same for both Solaris and Linux. Refer to the man page for more information about its use. dvd+rw-tools: Red Hat and Novell SUSE ship with this package, which provides the following commands for manipulating +RW/+R and -R[W] DVD media: dvd+rw-booktype dvd+rw-format dvd+rw-mediainfo dvd-ram-control growisofs For more information about this package, refer to the growisofs man page and the following system documentation: Red Hat: /usr/share/doc/dvd+rw-tools-<version> Novell SUSE: /usr/share/doc/packages/dvd+rw-tools

Chapter 6. Device management

97

dvdrecord: Red Hat provides this additional command, which is a stub for the cdrecord command. The man page documentation recommends using the growisofs command from the dvd+rw-tools package instead of this command. The previously mentioned CD/DVD Creator (Red Hat) and K3b (Novell SUSE) applications also provide DVD burning capabilities.

Erasing
The cdrecord command is capable of erasing CD-RW media. The dvd+rw-tools package provides the ability to format and erase DVD media.

Audio and video

Table 6-8 describes the commands and tools used to extract and play audio data from CD and DVD media.
Table 6-8 Solaris versus Linux audio extraction and playback tools Task Extraction shell command Extraction GUI command Playback shell command Playback GUI command Solaris cdrw xmcd cda xmcd Red Hat cdparanoia sound-juicer cdda2wav xmms Novell SUSE cdparanoia k3b cdda2wav xmms

Note: Although Sun recommends xmcd, Solaris does not include the tool. Red Hat and Novell SUSE do not provide tools for video extraction or playback. However, the following Web site provides software and information resources related to using DVDs with Linux:
http://dvd.sourceforge.net/

6.2.5 Tape drives

Linux supports an assortment of tape drives. For information about whether or not a particular drive is supported under Linux, refer to the manufacturers Web site.

Supported formats
The following list outlines the tape formats supported in Linux: Travan

98

Solaris to Linux Migration: A Guide for System Administrators

Digital Audio Tape (DAT) Digital Linear Tape (DLT) Super Digital Linear Tape (Super DLT) 8 mm Advanced Intelligent Tape (AIT)

Tape device name format

Solaris administrators refer to tape devices by selecting the subdirectory that it is linked to /dev/rmt, followed by a string of information that indicates the specific drive number, density, and device options:
/dev/rmt/[1][2][3][4]

Where: 1: Drive number 2: Optional density flag: [l]ow, [m]edium, [h]igh, [u]ltra, or [c]ompressed 3: Berkeley compatibility flag: [b] 4: No rewind option: [n] Linux has a different scheme for naming physical tape devices depending on the tape device type. Due to this variance on names, the /dev/tape device name is used for simplicity and links to the first tape device. Following sections highlight a few examples of supported tape device types and their naming scheme. For a compete list, see the device list maintained by the Linux Assigned Names And Numbers Authority (LANANA), available at:
http://www.lanana.org

SCSI tape device

A generic SCSI tape device is indicated by an st in the device name. Most SCSI tape devices use this name scheme. The mode setting is optional and defined by the manufacturer. Figure 6-3 on page 100 shows a SCSI tape example.

Chapter 6. Device management

99

Figure 6-3 SCSI tape example

IDE tape device

An IDE tape device is indicated by an ht in the device name. Figure 6-4 shows a IDE tape example.

Figure 6-4 IDE tape example

QIC-117 tape device

A QIC-117 tape device is indicated by a qft in the device name. The compression part of the name can be left out to indicate no compression. In addition, this same name section can contain the word raw to indicate no file marks. Figure 6-5 on page 101 shows a QIC-117 tape example.

100

Solaris to Linux Migration: A Guide for System Administrators

Figure 6-5 QIC-117 tape example

Displaying tape drive status

There are several ways to display the status of a Linux tape device. Issuing a man -k tape command lists the tape-related commands available for Red Hat and Novell SUSE. The following commands are a commonly used subset of both distributions tape-related commands.

The mt command
Both Solaris and Linux have the mt command for controlling magnetic tapes. The Linux version contains all the options of the Solaris version and more. Refer to the man page for more information about this command.

The mtx command

The mtx command controls SCSI media changers such as tape changers and tape libraries. Refer to the man page for more information about this command.

The tapeinfo command

The tapeinfo command reads information from SCSI tape drives that is not usually available by using the tape drivers supplied by vendors. Refer to the man page for more information about this command.

6.3 Terminals and modems

This section covers the following terminal- and modem-related topics: Terminal setup and initialization Modem setup tools Serial port management Port monitoring

Chapter 6. Device management

101

6.3.1 Terminal setup and initialization

The Linux operating system uses virtual terminals to establish console access through the keyboard and monitor attached to the system. There are six virtual terminals configured, tty1 through tty6, and each terminal allows user logins. Users can switch between virtual terminals by typing Alt+F1 through Alt+F6. If the system is using X Window System, users must use Ctrl+Alt+F1 through Ctrl+Alt+F6. The /etc/inittab file contains the configuration information for these tty terminals, and the LANANA device list contains information about all tty devices, available at:
http://www.lanana.org

Often, a Solaris system administrator needs to configure a terminal to run across a serial port. Sun recommends using Solaris Management Consoles (smc) Serial Ports Tool for setting up terminals in Solaris. The Linux operating system represents serial port devices under the /dev file system as ttyS#, where # represents the serial port number. Unfortunately, neither Red Hat nor Novell SUSE provide a tool for configuring serial port terminals; however, the process to accomplish this setup is simple. The following sections provide the steps required to enable this feature in Linux.

Red Hat
Perform the following steps: 1. Open the /etc/inittab file. 2. Locate the following section:
# Run gettys in standard runlevels 1:2345:respawn:/sbin/mingetty tty1 2:2345:respawn:/sbin/mingetty tty2 3:2345:respawn:/sbin/mingetty tty3 4:2345:respawn:/sbin/mingetty tty4 5:2345:respawn:/sbin/mingetty tty5 6:2345:respawn:/sbin/mingetty tty6

3. Add the following line below this section:

S0:12345:respawn:/sbin/agetty -L <baud> <serial device> <TERM>

Where: <baud> = baud rate of connection, for example, 38400 <serial device> = serial port device to which to attach terminal, for example, ttyS0 <TERM> = terminal type to run over connection, for example, xterm

102

Solaris to Linux Migration: A Guide for System Administrators

4. Run the telinit q command, which tells the init program to check, or query, the /etc/inittab file and make changes as needed.

Novell SUSE
Perform the following steps: 1. Open the /etc/inittab file. 2. Locate the following section:
# #S0:12345:respawn:/sbin/agetty -L 9600 ttyS0 vt102

3. Remove the # and set the baud rate, serial device, and terminal type as needed. 4. Run the telinit q command, which tells the init program to check, or query, the /etc/inittab file and make changes as needed.

6.3.2 Modem setup tools

Red Hat and Novell SUSE provide configuration tools and instructions for setting up modems, as described in Table 6-9.
Table 6-9 Modem configuration Linux distribution Red Hat Command to start tool system-config-network Location of instructions Refer to Chapter 18, Network Configuration, in Red Hat Enterprise Linux 4: System Administration Guide, at: http://www.redhat.com/docs/manuals/enterprise/ Refer to the Network Integration section in the Services chapter of SUSE LINUX Enterprise Server 9 Administration Guide for OES, at: http://www.novell.com/documentation/sles9/

Novell SUSE

yast modem

6.3.3 Serial port management

Linux provides the setserial command for setting and retrieving information about serial ports. The setserial command can configure this common list of parameters: port: Configures the I/O port number irq: Configures the IRQ number uart: Configures the universal asynchronous receiver/transmitter (UART) type

Chapter 6. Device management

103

baud_base: Configures the speed of the serial port Refer to the man page for further information about, as well as options for, the setserial command.

Port initialization
To initialize a specific serial port, such as ttyS0, run the following command:
setserial /dev/ttyS0 autoconfig

Novell SUSE provides an additional method for configuring serial ports. Initialization can be accomplished using the setserial script, located in /etc/init.d. Red Hat Enterprise Linux also provides a setserial command for managing initialization and settings. Refer the man page for setserial for more information.

Serial console port configuration

Assuming a login is configured on ttyS0 (see 6.3.1, Terminal setup and initialization on page 102), use the following boot option to configure a serial console port in Linux:
console=ttyS0

This option can be issued at the boot prompt or appended into the bootloader configuration file. For information about booting, refer to 8.1, Booting a system on page 126 and 8.3, Boot configuration files on page 132.

6.3.4 Port monitoring

Solaris provides ttymon for monitoring and managing ports. Table 6-10 lists and describes the various port monitoring programs available in Linux. Consult the man pages for information about their options.
Table 6-10 Port monitoring programs available for Linux Port monitoring program agetty Description The default alternative getty program for Linux, which includes many nonstandard features as compared to getty The smart modem getty for use with modems A minimal getty for use only on Linux virtual consoles

mgetty mingetty

Novell SUSE
Novell SUSE also provides the original getty program for monitoring ports.

104

Solaris to Linux Migration: A Guide for System Administrators

Services administration
Solaris provides the pmadm command to enable system administrators to list the services of ports associated with a port monitor, add services, and enable or disable a service. Linux does not have a similar command; however, you can accomplish the same functionality provided by pmadm. Table 6-11 lists the tty service-related tasks in Solaris and how to accomplish them in Linux.
Table 6-11 Solaris versus Linux tty port service commands Port service task Add a TTY port service View status on TTY port service Solaris pmadm -a pmadm -l Linux Add a service line in /etc/inittab. See man page. cat /etc/inittab ps -ef | grep tty ps -ewft cat /proc/tty/drivers Run telinit q after adding the service to the /etc/inittab file. Run telinit q after removing the service from the /etc/inittab file.

Enable a TTY port service Disable a TTY port service

pmadm -e pmadm -d

6.4 Distribution-based device management tools

Red Hat and Novell SUSE provide their own interactive tools for managing devices: Red Hat kudzu: This tool detects and configures hardware on a system. For information about using this tool, refer to its man page. Novell SUSE yast2: This system configuration utility supplies modules for configuring common hardware, such as disk controllers and displays. For information about using this tool, refer to its man page.

Chapter 6. Device management

105

106

Solaris to Linux Migration: A Guide for System Administrators

Chapter 7.

Network services
This chapter discusses the differences in networking and networking services between Solaris and Linux. We do not describe what TCP/IP is and how to plan for your networking services, because this is the same for both operating systems. This chapter includes the following topics: IPv4 IPv6 Mixed IPv4 and IPv6 networks Static and dynamic routing IPSec and IKE Network multipath Network trunking IP network services TCP wrappers IP stateful firewalling

Copyright IBM Corp. 2006. All rights reserved.

107

7.1 IPv4
IP version 4 is the default networking stack for all recent operating systems, and this is no exception for Solaris and Linux. However, there are a number of small differences between the two operating systems, and in this chapter, we highlight those differences.

IPv4 in Solaris
The installation program of Solaris will usually set the right networking parameters for you, so that after installation, the system is ready to operate in the network. To set or modify the TCP/IP parameters of a Solaris box after installation, you need to edit several files: /etc/hostname.interface /etc/netmasks /etc/nodename /etc/defaultdomain /etc/defaultrouter /etc/inet/hosts /etc/nsswitch.conf /etc/resolv.conf /etc/bootparams /etc/ethers File containing IP address or host name for each network interface you need to use. Contains subnetting information. Contains the host name of the system. Contains the fully qualified domain name of the system. Contains an entry for each router that the server will use. Database contains the local IP address to name mappings. The /etc/hosts is a symlink to this file. Contains configuration of network name services. Contains the DNS client configuration of the system. Contains information for serving network booting services. Contains information about the MAC to IP address mapping used by the rarpd daemon.

There are other files used by IPv4 networking (/etc/inet/services or /etc/inet/protocols), but their default values are fine for most applications and usually do not need any modification. To use dhcpagent, the DHCP client for configuring a network interface, you need to create an empty file /etc/dhcp.interface (such as /etc/dhcp.hme0). You can alter the default dhcpagent behavior by editing the /etc/default/dhcpagent configuration file. In addition, you can use the sys-unconfig command and reconfigure your Solaris box after a reboot.

108

Solaris to Linux Migration: A Guide for System Administrators

IPv4 in Linux
The installation program of Linux also sets up networking parameters for you, so that after the installation process, the system is ready to operate in the network. To set or modify the TCP/IP parameters after the initial installation, you have the option of editing the configuration files or using a GUI. The GUI program for accomplishing this in RHEL is system-config-network and in SLES is yast2. The most important configuration files used by Linux networking services are: In RHEL: /etc/sysconfig/network-scripts/ifcfg-interface-name This file contains all the information needed for configuring a specific interface, including IP, subnet mask, broadcast, custom MAC address, and gateway, or if the interface is to be configured by DHCP or BOOTP protocol. In addition, you can configure whether the interface should be activated at boot time or later. If you choose to activate it later, you can do so with the ifup command. Other general networking options can be specified in the /etc/sysconfig/network file. For more information about the possible options to configure in these files, refer to the /usr/share/doc/initscripts-version/sysconfig.txt file and RHEL System Administration Guide. In SLES: /etc/sysconfig/network/ifcfg-interface-id-mac-address This file contains all the information needed for configuring that specific interface, including IP, subnet mask, broadcast, custom MAC address, and gateway, or if the interface is to be configured by DHCP or BOOTP protocol. In addition, you can configure whether the interface should be activated at boot time or later. If you choose to activate it later, you can do so with the ifup command. Other general networking options can be specified in the /etc/sysconfig/network/config file. For more information about the possible options to configure in these files, refer to the SLES Administration Guide.

Common network configuration files in Solaris and Linux

There are several networking configuration files that are common to both Linux to the Solaris: /etc/hosts /etc/nsswitch.conf /etc/resolv.conf /etc/bootparams This file specifies IP address to name mappings. Contains configuration of network name services. Contains the DNS client configuration of the system. Contains information for serving network booting services only to Solaris systems. Linux clients do not need this service.

Chapter 7. Network services

109

/etc/services /etc/protocols

Contains TCP/UDP port numbers to service names mapping. Contains IP protocol numbers to protocol names mapping.

Other networking differences

The interface naming convention is different. In Solaris, we the interface typenum (for example, le0, le1, hme0, hme1, ce0, ce1), while in Linux we have only ethnum (for example, eth0, eth1), where num is the interface number. Linux network interfaces do not need plumbing; they are active from the moment their driver is loaded until the moment we unload the driver from the kernel. For configuring IP networking parameters in Solaris, use the ndd command. In Linux, you directly change the parameters in the /proc/sys/net directory with the sysctl program, or for a permanent change, modify the /etc/sysctl.conf configuration file of the sysctl program. Linux provides for network profiles, which means that we can create several network profiles and change them as necessary. However, they are most useful for mobile workstations and computers, rather than servers, so we do not cover them here.

Task table
Table 7-1 illustrates network commands and configuration files in Solaris and their equivalent in Linux.
Table 7-1 Network commands and configuration files Task Configure TCP/IP Solaris Edit the following files: /etc/hostname.* /etc/inet/* /etc/defaultrouter /etc/defaultdomain /etc/nodename /etc/netmasks ifconfig netstat -i ifconfig Linux RHEL: /etc/sysconfig/network and /etc/sysconfig/networking/* SLES: /etc/sysconfig/network/*

Display interface settings Display interface status and statistics Configure interface

ip ifconfig netstat -i ifconfig ip ifconfig

110

Solaris to Linux Migration: A Guide for System Administrators

Task Check various network statistics Change resolver Configure name services Display kernel network parameters Configure kernel network parameters Check for network link Rename a network interface Check routing table Testing for connectivity Check IP path Capturing network packets

Solaris netstat vi /etc/resolv.conf vi /etc/nsswitch.conf ndd /dev/ip \? ndd /dev/tcp \? ndd ndd N/A netstat -r ping traceroute snoop

Linux netstat vi /etc/resolv.conf vi /etc/nsswitch.conf sysctl -a | grep ^net sysctl -w variable=value ethtool mii-tool ip netstat -r route ping traceroute tcpdump tethereal ethereal

Example 7-1 shows sample output from ifconfig on Solaris.

Example 7-1 Sample Solaris ifconfig output # ifconfig -a lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index 1 inet 127.0.0.1 netmask ff000000 eri0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2 inet 9.3.5.145 netmask ffffff00 broadcast 9.3.5.255 ether 0:3:ba:29:bf:1

Example 7-2 on page 112 shows sample output from ifconfig on Linux. Note the additional information provided. If the -a option is provided, it also lists all the inactive interfaces. Remember that Linux does not have the plumb/unplumb feature to enable and disable a network device.

Chapter 7. Network services

111

Example 7-2 Sample Linux ifconfig output # ifconfig eth0 Link encap:Ethernet HWaddr 00:02:55:7B:02:B0 inet addr:9.3.5.11 Bcast:9.3.5.255 Mask:255.255.255.0 inet6 addr: fe80::202:55ff:fe7b:2b0/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:847931 errors:0 dropped:0 overruns:0 frame:0 TX packets:211207 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:74462901 (71.0 MiB) TX bytes:280066144 (267.0 MiB) Base address:0x2500 Memory:fbfe0000-fc000000 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:4944 errors:0 dropped:0 overruns:0 frame:0 TX packets:4944 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:3062806 (2.9 MiB) TX bytes:3062806 (2.9 MiB)

# # ifconfig -a eth0 Link encap:Ethernet HWaddr 00:02:55:7B:02:B0 inet addr:9.3.5.11 Bcast:9.3.5.255 Mask:255.255.255.0 inet6 addr: fe80::202:55ff:fe7b:2b0/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:3569066 errors:0 dropped:0 overruns:0 frame:0 TX packets:4008490 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:426383647 (406.6 MiB) TX bytes:1746883088 (1.6 GiB) Base address:0x2500 Memory:fbfe0000-fc000000 eth1 Link encap:Ethernet HWaddr 00:02:55:7B:02:B1 BROADCAST MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 b) TX bytes:0 (0.0 b) Base address:0x2540 Memory:fbfc0000-fbfe0000 Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:5340 errors:0 dropped:0 overruns:0 frame:0 TX packets:5340 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:3194582 (3.0 MiB) TX bytes:3194582 (3.0 MiB)

lo

112

Solaris to Linux Migration: A Guide for System Administrators

sit0

Link encap:IPv6-in-IPv4 NOARP MTU:1480 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)

7.2 IPv6
We discuss IPv6 differences in this section.

IPv6 in Solaris
The installation program of Solaris asks if you want to configure IPv6 on that server and set the right networking parameters for you, so that after install, the system is ready to operate in an IPv6 network. To manually configure IPv6 on a Solaris system, you need to touch the /etc/hostname6.interface file for each IPv6 interface and reboot the system. Optionally, you can add IP to name mappings in the /etc/inet/ipnodes file. You can use most of the tools from IPv4 with special switches for IPv6.

IPv6 in Linux
The IPv6 stack is present by default in both RHEL and SLES, and you will have a link-local address on all your network interfaces. To modify the IPv6 stack configuration, you need to edit the following files: In RHEL: /etc/sysconfig/network-scripts/ifcfg-interface-name The file contains all the information needed for configuring IPv6 on that specific interface. You can also configure whether the interface should be activated at boot time or later. If you choose to activate it later, you can do so with the ifup command. Other general networking options can be specified in the /etc/sysconfig/network file. For more information about the possible options to configure in these files, refer to the /usr/share/doc/initscripts-version/sysconfig.txt file.

Chapter 7. Network services

113

In SLES: /etc/sysconfig/network/ifcfg-interface-id-mac-address This file contains all the information needed for configuring IPv6 on that specific interface. You can also configure whether the interface should be activated at boot time or later. If you choose to activate it later, you can do so with the ifup command. You can specify other general networking options in the /etc/sysconfig/network/config file.

7.3 Mixed IPv4 and IPv6 networks

Because IPv6 is not yet a widely deployed protocol, we might have IPv6 islands separated by IPv4 routers. In order to link those IPv6 islands, we can use tunnels. Both Solaris and Linux support IPv6 tunneling in IPv4.

Solaris tunneling
In Solaris, you can add tunnel interfaces using the ifconfig command or adding options to the /etc/hostname6.interface file.

Linux tunneling
In Linux, you can create tunnels by adding configuration files for them in the /etc/sysconfig/network directory. For more information, see the documentation for your Linux distribution.

7.4 Static and dynamic routing

We discuss the use of static and dynamic routes in two major cases: when the server is acting as a network router providing routing services to clients, and when the server is acting like a end-user system and it needs a gateway to communicate with the world.

Routing in Solaris
To configure a Solaris box to act as a router, you need at least two network interfaces on that box. Then, you create a /etc/hostname.interface file for each network interface installed, specifying the host name or the IP address of that interface. Then, you must enter those host names and IP addresses in the /etc/inet/hosts file for IPv4 and /etc/inet/ipnodes for IPv6. If this Solaris router is part of any subnetted network, we add the subnets in the /etc/inet/netmask file. The startup scripts then determine if they need to start a routing daemon for learning and advertising routes (daemons such as RIP or RDISC) or use static routing.

114

Solaris to Linux Migration: A Guide for System Administrators

On an end-user Solaris box, we can control the use of a static default route by editing the /etc/defaultrouter file. If the file contains a default gateway, the system uses that default gateway and does not launch a dynamic routing protocol such as RIP or RDISC. If the /etc/defaultrouter file is empty, the system tries to use a dynamic gateway first, using /usr/sbin/in.rdisc if it exists, and if this method fails, it starts the RIP routing protocol in passive, non-advertising mode. A machine with only one network interface can be forced to be a router by creating an empty /etc/gateways file. A machine with more than one network interfaces can be forced to be a multi-homed host instead of a router by creating an empty /etc/notrouter file. For IPv6 routing, you need to use the ndpd and ripngd daemons.

Routing in Linux
On a Linux box, regardless of how many network interfaces you have, the routing functionality can be enabled by the sysctl program. To enable routing, enter the following command in the /etc/sysctl.conf file and then run sysctl -p:
net.ipv4.ip_forward = 1

This configures the kernel to permit IP packet forwarding from one network interface to another one. Also check if the IP firewall service is configured properly for your environment. By default, static routing is used for both network router and end-user configurations. If a DHCP client is used for configuring the network parameters, the default gateway DHCP option is also considered a static route because it is not dynamically modified during the lease period. If dynamic routing is required, we have several daemons available in the quagga package: ripd ripngd ospfd ospf6d bgpd radvd For RIP routing protocol For IPv6 RIP routing protocol For OSPF routing protocol For IPv6 OSPF routing protocol For BPG routing protocol For router advertisement daemon for IPv6

All these routing protocol daemons have their own configuration file in the /etc/ directory.

Chapter 7. Network services

115

For more details about the configuration of dynamic routing protocols, consult the documentation of the quagga package and the distribution manuals or visit the Web site at:
http://www.quagga.net/

7.5 IPSec and IKE

IPSec and IKE protocols are supported by both Solaris and Linux for VPN connectivity. We cover the topics about configuring and administering IPSec and IKE features in detail in Chapter 14, Security and hardening on page 187.

7.6 Network multipath

In Solaris, there is a feature called IP Network Multipath that enables you to create one or more groups of physical interfaces and use them as one virtual interface. It provides for: Failure detection Repair detection Outbound load spreading In Linux, we can implement a similar solution with some subsets of the interface bonding feature. See 7.7, Network trunking on page 116 for more information about this feature.

7.7 Network trunking

The interface trunking feature in Solaris is almost equivalent to the interface bonding feature in Linux. In order for this feature to work, you need to properly configure the switch at the other end of the network cables. Cisco switches call this feature EtherChannel. Tip: Linux bonding has more features than the Solaris trunking, and in some configuration modes, it can be used without switch support for EtherChannel.

116

Solaris to Linux Migration: A Guide for System Administrators

Solaris trunking
In Solaris, you need to have a special quad network adapter for creating network trunks. Then, you can use several network interfaces from this adapter to create a trunk interface, which is like a big pipe composed of all the individual interfaces. You assign the IP address only to the trunk interface.

Linux bonding
The Linux bonding driver provides a method for incorporating multiple Ethernet network interfaces into a single bonded interface. You can create configuration files for them in the /etc/sysconfig/network directory. The bonded interfaces can provide either hot standby or load balancing services. In addition, link integrity monitoring can be performed. The ifenslave user-level control program is included in both RHEL and SLES distributions. For more information about configuring a bonded interface, refer to the documentation located in the /Documentation/networking/bonding.txt file of the kernel source tree or visit the official project Web site at:
http://sourceforge.net/projects/bonding

Tip: If you only need more bandwidth, Linux supports 1 Gb and 10 Gb network interfaces.

Linux bridging
In Linux, you can create interface bridges with any Ethernet network interface by using commands from the bridge-utils package. For more information, see the Linux distribution and bridge-utils package documentation, the /Documentation/networking/bridging.txt file of the kernel source tree, and visit the official project Web site at:
http://bridge.sourceforge.net

For an example of using the bridge-utils package, refer to Physical and virtual network bridging on page 296.

7.8 IP network services

This section contains information about a set of commonly used network services.

7.8.1 inetd-based versus xinetd-based network services

In this section, we describe the differences between inetd for Solaris and xinetd for Linux.

Chapter 7. Network services

117

Solaris inetd server

In Solaris, standard network services, such as Telnet, FTP, Rlogin, TFTP, and others, are started on demand by the inetd network daemon through the /etc/inetd.conf configuration file. The /etc/services configuration file contains the portname to portnumber mappings for network applications that can be used with the inetd daemon.You can activate usage logging and TCP wrappers by specifying the following parameters in the /etc/default/inetd configuration file:
ENABLE_CONNECTION_LOGGING=YES ENABLE_TCPWRAPPERS=YES

All services to be run by the inetd daemon are defined in the /etc/inetd.conf configuration file.

Linux xinetd server

In Linux, the xinetd network daemon is the extended Internet services daemon. The xinetd daemon general behavior is controlled by its /etc/xinetd.conf configuration file. However, unlike the previous inetd daemon, xinetd has all the network services defined in separate files, one file per service, in a separate directory. The default directory for services configuration is /etc/xinetd.d/, but it can be changed to something else in the /etc/xinetd.conf global configuration file. Here are some basic services and their associated configuration files: Telnet TFTP /etc/xinetd.d/telnet /etc/xinetd.d/tftpd Remote shell /etc/xinetd.d/rsh

The syntax of the xinetd configuration files is also different than inetd. Example 7-3 shows an example of a configuration file for Telnet services.
Example 7-3 Example xinetd service configuration service telnet { socket_type = stream protocol = tcp wait = no user = root server = /usr/sbin/in.telnetd disable =yes }

Note that the default state for an xinetd service is disabled. With xinetd, it is possible to configure specific logging and network access control for each separate service. The network access control feature can be implemented in two

118

Solaris to Linux Migration: A Guide for System Administrators

ways: with the built-in support of xinetd, or with the standard tcpd program from the tcp-wrappers package. If any of the configuration files are modified while xinetd is running, we need to send a SIGHUP signal to xinetd in order to reread its configuration; otherwise, it will still run with the old configuration. For more information about xinetd, check the man page of xinetd, xinetd.conf, and xinetd.log. Tip: It is not uncommon that services in one system are inetd or xinetd based, and in other systems are stand-alone services. This depends on your distribution and release level.

7.8.2 DHCP
The Dynamic Host Configuration Protocol (DHCP) server from Solaris and Linux have comparable features: Support for previous BOOTP clients Support for local and remote clients by using DHCP relay services Large network support Network booting information for DHCP clients Support for vendor options DDNS updates However, the server itself is very different in terms of management scripts and configuration files.

Solaris DHCP server

In Solaris, there are several utilities for controlling the in.dhcpd server daemon such as dhcpconfig, dhtadm, pntadm, and the GUI dhcpmgr. The configuration is stored in the /etc/inet/dhcpsvc.conf file and some other tables, such as dhcptab, macros, and so on.

Linux DHCP server

In Linux, the controlling script of the DHCPD server is /etc/init.d/dhcpd, which supports the start, stop, restart, and status arguments among others. The daemon configuration file is /etc/sysconfig/dhcpd, and the database containing all the information is in /etc/dhcpd.conf. Leases are kept in /var/lib/dhcpd.leases. There is no GUI or tool for configuring DHCP in RHEL; the only way to configure

Chapter 7. Network services

119

it is to edit the files as needed and restart the daemon with /etc/init.d/dhcpd restart. In SLES, there is a DHCP server configuration module in the YaST2 program. Important: In order to use the DHCP server with IBM Eserver i5 and Eserver p5 OpenPower firmware, you need to run a different DHCP server than the supplied with SLES9. By default, the SLES9 DHCP server uses Berkeley Software Design, Inc. (BSD) type sockets. This DHCP server always sends a broadcast response that the OpenPower firmware does not understand. The Linux Packet Filter (LPF) type sockets send a unicast response, to which the firmware will reply. Because of this caveat, SLES9 ships with two DHCPD binaries, dhcpd and dhcpd.lpf. In order to use the LPF DHCP server, you need to change the DHCPD_BINARY variable in /etc/sysconfig/dhcpd to:
DHCPD_BINARY="/usr/sbin/dhcpd.lpf"

For more information about DHCP, check the man pages for dhcpd and dhcpd.conf and the documentation that came with your distribution.

7.8.3 DNS
Both operating systems use BIND as the Domain Name System (DNS) server. Modern Linux distributions are based on BIND 9, so if you are familiar with BIND 8 or later on Solaris, you will be able to configure BIND 9 on Linux. The main configuration file is the same: /etc/named.conf in both Solaris and Linux.

7.8.4 NTP
Linux supports both a client and server Network Time Server (NTP) daemon service. While the configuration files for this service are in /etc/inetd/* on Solaris, Linux configures this using the /etc/ntp.conf file.

7.8.5 LDAP
Both Solaris and Linux support Lightweight Directory Access Protocol (LDAP) as a client and server. For information about this topic, see 11.3, Directory services on page 163.

120

Solaris to Linux Migration: A Guide for System Administrators

7.8.6 NIS and NIS+

Both Solaris and Linux support Network Information Service (NIS) and NIS+ protocols. For information about this topic, see 11.4, NIS and NIS+ on page 164.

7.8.7 NFS
The Network File System (NFS) is used on both Solaris and Linux for sharing file systems across a network. Both operating systems can act as servers and as clients, even at the same time.

NFS server setup in Solaris

In Solaris, you add the subdirectory paths for NFS sharing in the /etc/dfs/dfstab file, and start the NFS server with /etc/init.d/nfs.server start. If you already have an NFS server running and you update the content of the dfstab file, you can run the shareall command to update the NFS shares without restarting the service.

NFS client setup in Solaris

To mount a shared NFS resource, you can either do it manually with the mount command or automatically at boot time by placing an entry in the /etc/vfstab file.

NFS server setup in Linux

In Linux, you add the subdirectory paths for NFS sharing in the /etc/exports file, and start the NFS server with /etc/init.d/nfs start on RHEL and /etc/init.d/nfsserver start on SLES. The Linux equivalence of the share and shareall commands is the exportfs command, which performs the same thing as the Solaris counterparts. The syntax of /etc/dfs/dfstab in Solaris is also different from the syntax of /etc/exports in Linux. Example 7-4 shows a sample /etc/exports file.
Example 7-4 NFS /etc/exports sample /pub (ro,insecure,all_squash) /home/joe joe_pc(rw,all_squash,anonuid=150,anongid=100) /usr *.local.domain(ro) @trusted(rw) /projects proj*.local.domain(rw) / master(rw) trusty(rw,no_root_squash)

Consult the man page for exports for information about the options in the /etc/exports file.

Chapter 7. Network services

121

Both Red Hat (system-config-nfs) and SUSE (yast2) offer GUI-based NFS server configuration.

NFS client setup in Linux

To mount a shared NFS resource, you can either do it manually with the mount command or automatically at the boot time by placing an entry in the /etc/fstab file. Table 7-2 lists the main differences between the Solaris and Linux NFS tasks and configuration files.
Table 7-2 Solaris NFS and Linux NFS differences Tasks and configuration files Manually start the NFS server Solaris /etc/init.d/nfs.server start mount -F nfs server://resource /mnt/point /etc/default/nfs /etc/default/nfslogd shareall share /etc/dfs/dfstab Linux /etc/init.d/nfs start (in RHEL) /etc/init.d/nfsserver start (in SLES) mount server://resource /mnt/point /etc/sysconfig/nfs /etc/sysconfig/nfs exportfs -a exportfs -o /etc/exports

Mount a resource on a NFS client NFS server general config file NFS logging options config file Share all exported file systems Share a new file system Config file of shared file systems

7.8.8 Web proxy and cache servers

Solaris offers a follow-on product (Web Proxy Server) for this capability. Linux ships by default the squid package for proxy and Web cache services.

7.8.9 Mail services

Solaris uses sendmail as the mail transfer agent. RHEL uses sendmail as mail transfer agent as well, so you are already familiar with this MTA. postfix is available as an alternative. SLES uses postfix as the mail transfer agent, but sendmail is available for this distribution as well. Both Linux distributions have POP3 and IMAP4 servers available by default.

122

Solaris to Linux Migration: A Guide for System Administrators

7.9 TCP wrappers

Wrappers (tcpd) functionality is offered by both Solaris and Linux in the same manner. On Linux, the configuration of /etc/xinetd.conf (instead of /etc/inetd.conf) is required. For more information, refer to 7.8.1, inetd-based versus xinetd-based network services on page 117 and 14.3, Securing and removing services on page 189.

7.10 IP stateful firewalling

Solaris has the SunScreen product for IP firewalling, and Linux has the iptables-based firewall. Read 14.13, Firewalls on page 202 for more information about firewalls.

Chapter 7. Network services

123

124

Solaris to Linux Migration: A Guide for System Administrators

Chapter 8.

Boot and system initialization

This chapter describes the differences in system booting and initialization between Solaris and Linux. We discuss the following subjects: Booting a system Run levels Boot configuration files Shutting down Network booting

Copyright IBM Corp. 2006. All rights reserved.

125

8.1 Booting a system

This section provides information about specific ways to boot a system, different booting sources, and ends with an overview of the booting process.

8.1.1 Booting types

A server can be booted in several ways, depending on the tasks the system administrator needs to perform. Where the Sun SPARC platform has the OpenPROM firmware to control and initiate the system booting, Linux has several platform-specific boot loaders for performing the task of booting a system. The current default Linux boot loader for the IA32 platform is GRand Unified Bootloader (GRUB). Prior to this, there was Linux Loader (LILO) and LOADLIN. Note: GRUB is standard on Intel-based and AMD-based servers and is used in combination with BIOS on PC-based hardware. For other platform-specific boot loaders, read Part 3, IBM eServer platforms on page 235. Table 8-1 presents the equivalence between the Solaris and Linux boot processes.
Table 8-1 Boot types Boot type Default boot mode Solaris boot Or if auto-boot is set to true, it is automatic. Note: The commands listed in this column assume that you are at the OpenPROM prompt. Single user mode Reconfiguration mode Verbose boot mode boot -s boot -r boot -v Linux Automatic or press Enter at the grub prompt. Note: The commands listed in this column assume that you are at the GRUB prompt. Add -s, single, S, or 1 at the end of boot line. Not needed. in RHEL, remove rhgb quiet from the boot command line. In SLES, press Esc when booting. Edit grub boot line to specify other kernel and options.

Interactive boot mode

boot -a

126

Solaris to Linux Migration: A Guide for System Administrators

Boot type Recovery boot mode Interactive start of services while booting Emergency boot mode

Solaris boot cdrom boot net N/A N/A

Linux Boot from a rescue CD-ROM or from the network. Possible, but vendor dependent. Add emergency or the -b option at the end of boot line.

8.1.2 Booting sources

Solaris and Linux can both boot from different media types. Table 8-2 presents the possible media sources for Solaris and Linux.
Table 8-2 Boot sources Boot source Hard disk Solaris boot Or automatic if auto-boot is set to true. boot cdrom boot net Not supported. Linux Automatic if set in BIOS (on PC-based hardware). Automatic if set in BIOS, selection at boot time (on PC-based hardware). Need DHCP (PXE) and TFTP server configured for network boot. Automatic if BIOS supports booting from these devices (on PC-based hardware).

CD-ROM/DVD Network Floppy and various USB devices

8.1.3 Booting process overview

Here we describe how a system is started up. Table 8-3 shows how Solaris and Linux control the booting process.
Table 8-3 Boot process Boot process Boot control mechanism Boot control location Solaris OpenPROM NVRAM Linux BIOS to MBR to GRUB /boot/grub/menu.lst Note: This is platform dependent. Applicable to IA32 systems.

Chapter 8. Boot and system initialization

127

Overview of SPARC Solaris booting process

In Solaris for SPARC, booting a kernel is the task of the OpenPROM firmware, which: 1. Runs POST. 2. Identifies and initializes the proper device for booting. 3. Loads bootblk. 4. This then loads and executes ufsboot. 5. Then ufsboot loads and executes the core kernel. 6. Initializes the core kernel data structures. 7. Loads other kernel modules based on the /etc/system file. 8. Finally, starts the /sbin/init program that brings up the system based on information in the /etc/inittab file. See the Table 8-4 on page 130 for the kernel location in Solaris.

Overview of Linux for IA32 booting process

On the IA32 platform, the BIOS performs the following tasks: 1. Initializes the system and peripheral devices. 2. Locates and runs the MBR code, which is in the first disk sector from the first BIOS defined disk drive, or can be configured to boot from a CD-ROM, floppy, or network device. The Master Boot Record (MBR) is unique to Intel Architecture (IA) systems and contains a description of the disk (primary partitions, boot partition) and a little program called boot loader that further loads the operating system. In Linux for the IA32 platform, the default boot loader is GRand Unified Bootloader (GRUB), which can be used to boot not only Linux systems, but also other operating systems, such as DR-DOS and Microsoft Windows. You can also use GRUB to boot other kernel versions in the same Linux system or a same kernel with different boot parameters. The configuration file for GRUB is /boot/grub/menu.lst, and it tells GRUB where to find the kernel to boot from and what the parameters are for the kernel. In Linux, the kernel to boot is placed in the /boot partition or directory along with the necessary kernel modules for proper booting. These kernel modules along with some early scripts are packed in a special format called the initrd image. Check the man page for initrd for more information about this file.

128

Solaris to Linux Migration: A Guide for System Administrators

The kernel starts initializing devices and loads additional device drivers from the initrd image. It then starts the /sbin/init program, which brings up the system based on information in the /etc/inittab file. The GRUB boot loader is also interactive, so you can modify what kernel to boot and its parameters before booting the system. This is equivalent to the interactive booting of Solaris with the boot -a command, but allows more control. At the GRUB prompt, you can change the automatic startup and edit the boot option strings. To edit the current boot entry in GRUB, you need to press e in Red Hat or Esc in SUSE. When in the GRUB interactive prompt, you can change the boot parameters before the kernel is loaded. Example 8-1shows some sample boot strings for RHEL.
Example 8-1 Changing boot parameters in GRUB root (hd0,0) kernel /boot/vmlinuz-2.6.9-11.ELsmp ro root=LABEL=/1 rhgb quiet initrd /boot/initrd-2.6.9-11.ELsmp.img

For example, if you want a verbose boot for the above boot string, remove rhgb quiet. For more information about GRUB, refer to the Linux man pages or the project Web site at:
http://www.gnu.org/software/grub/

Whenever you update the Linux kernel, a new boot entry is created for you in the GRUB configuration file and a new initrd image is created to match the new kernel. If you need to create a custom initrd image, refer to the man pages for the mkinitrd command. All kernel modules that can be used for a specific kernel are in the /lib/modules/uname -r/ directory. Table 8-4 on page 130 shows the major differences between booting in Solaris and Linux.

Chapter 8. Boot and system initialization

129

Table 8-4 Booting process Booting information Boot process Solaris Phases: 1. OpenPROM: Displays system information, runs POST, loads bootblk, locates ufsboot. 2. Boot programs: bootblk loads and executes the ufsboot. 3. Kernel initialization: ufsboot loads and executes the core kernel, initializes core kernel data structures, loads other kernel modules based on the /etc/system file, starts /sbin/init program. 4. init: Starts other processes based on the /etc/inittab file. 5. /sbin/rcX runs the corresponding /etc/rcX.d scripts to start up other components. Linux Phases: 1. BIOS: Checks the system and peripheral devices. Locates and runs the Master Boot Record (MBR). 2. MBR loads GRUB (GRand Unified Bootloader). 3. GRUB boots the kernel using information in /boot/grub/menu.lst. 4. Kernel starts initializing devices and loads additional device drivers from initrd image, and then starts init process from /sbin/init. 5. init: Starts other processes based on the /etc/inittab file. 6. /etc/rc.d/rc in RHEL or /etc/init.d/rc in SLES runs the corresponding /etc/rcX.d scripts to start up other components. /boot contains the kernels and initrd images.

Default kernel location

/kernel - Contains common components. /platform/platform-name/kernel Contains the platform-specific kernel components. /platform/hardware-class-name/ker nel - Contains kernel components specific to this hardware class. /usr/kernel - Contains kernel components common to all platforms within a particular CPU set.

130

Solaris to Linux Migration: A Guide for System Administrators

Booting information Default kernel modules location

Solaris Each of the subdirectories mentioned in the previous cell might contain the following subdirectories: drv - Contains loadable device drivers. exec - Contains modules that run programs stored in different file formats. fs - Contains file system modules. misc - Contains miscellaneous system-related modules. sched - Contains operating system schedulers. strmod - Contains loadble System V STREAMS modules. sys - Contains loadable system calls. cpu - Contains CPU type-specific modules. tod - Contains time-of-day hardware modules. sparcv9 - Contains 64-bit versions of specific modules.

Linux /lib/modules/uname -r/

8.2 Run levels

The concept of run levels is the same in both Solaris and Linux, and both have /etc/inittab and rc.d-style initialization scripts. However, there are some differences between the use of run levels in Solaris and Linux, as you can see in Table 8-5 on page 132.

Chapter 8. Boot and system initialization

131

Table 8-5 Run levels Run level S or s Solaris Single user level (boot -s). Only some file systems are mounted, as opposed to run level 1. Shuts down the operating system. Does not attempt to turn off the power. Returns to OpenPROM ok prompt. Administrative single user mode. Can access all available file systems. User logins disabled Multiuser without NFS resources shared. Multiuser with NFS resources shared (default run level for Solaris). The login is text or graphical, depending on the console capabilities. N/A (alternate multiuser). Power-down state. Shuts down the operating system and tries to turn power off if supported by the system. Reboot. N/A. Linux Minimal single user mode on SUSE. In Red Hat, it is the same as init level 1 Power-down state. Shuts down the operating system and tries to turn power off if supported by the system. Equivalent to Solaris run level 5 Single user mode.

2 3

Multiuser without NFS resources shared. Text-only login. Multiuser with NFS resources shared. Text-only login.

4 5

N/A (alternate multiuser). Full multiuser, NFS resources shared, and graphic login with X display manager. Reboot. Same as in Solaris. Ignore the /etc/inittab file and launch the default shell.

6 emergency

In both the Solaris and Linux operating systems, you can check out the current run level with the who -r command, and in Linux, you can additionally use the runlevel command for this task. Note: In Linux, the default run level is 5 (full multiuser with graphic login) unless there is no graphic card available, and then the default run level is 3 (full multiuser with text login).

8.3 Boot configuration files

Differences in how inittab and run control files are used are discussed in this section.

132

Solaris to Linux Migration: A Guide for System Administrators

8.3.1 /etc/inittab file

The /etc/inittab file has the same purpose in both operating systems, that is when the kernel loads the /sbin/init program, it reads this file to see what is the default run level and what scripts need to run. After initialization, the information in this file is used whenever the administrator changes the run level of the system. If you modify the /etc/inittab file and want to instruct init to reload it, you can do so with the init q or init Q commands. The syntax of this file is almost the same in both Solaris and Linux. For more in-depth information about this file, refer to the man page for inittab. Note: The /etc/inittab file is configured in Solaris to reach run level 3 by first running the scripts for run level 2 and then the scripts for the run level 3. But in Linux, there is only one set of scripts for each run level, meaning that for run level 3, the init program will run only the scripts for run level 3. It does not run the lower run level scripts to get to its intended run level. Pay extra attention to where you manually place an rc script.

8.3.2 Run control files (rc files)

Each run level has a set of scripts associated with it called run control files or rc files. Note: We also discuss the control of the system services through rc files in detail in 9.3, Starting and stopping system services on page 141. In Solaris, each run level has a master script /sbin/rcX, where X is the run level, and an associated directory in /etc/rcX.d/ where X is again the run level. When entering a run level, the init program starts the corresponding /sbin/rcX script, which in turn, executes the associated scripts from /etc/rcX.d/. In Linux, there is one single master script: /etc/rc.d/rc in Red Hat and /etc/init.d/rc in SUSE. This script the run level to start as an argument. The scripts directory associated with each run level is /etc/rcX.d/ in Red Hat and /etc/init.d/rcX.d in SUSE, where X is the run level, and usually the scripts in those directories are mostly symbolic links to the /etc/init.d/ main service scripts. For instance, in Red Hat, /etc/rc3.d/S80sendmail is a symbolic link to the ../init.d/sendmail script and /etc/rc1.d/K30sendmail is a symbolic link to the same

Chapter 8. Boot and system initialization

133

../init.d/sendmail script, and depending on the name of this link, the sendmail script will start or stop the service. Also new in Linux is the ability to restart or check the status of most of the services by passing the argument restart or status to the service script in the /etc/init.d directory:
# /etc/init.d/sendmail restart

And:
# /etc/init.d/sendmail status

Running the service script without options lists the supported options for that specific script:
# /etc/init.d/sendmail Usage: /etc/init.d/sendmail {start|stop|restart|condrestart|status} #

Tip: Red Hat provides the service script in /sbin, which enables you to run service scripts without specifying the file path:
# service sendmail status

Novell SUSE links many of the /etc/init.d/ service scripts to /usr/sbin/rc* files, which also removes the file path restriction:
# rcnfsserver Usage: /usr/sbin/rcnfsserver {start|stop|status|try-restart|restart|force-reload|reload}

8.3.3 Disabling rc scripts

At times, there is a need to prevent certain services from running. In Solaris, the suggested method is to rename the starting script by adding a _ character in front of the name, for example:
mv /etc/rcX.d/S30someservice /etc/rcX.d/_S30someservice

In Linux, we can use the same method as in Solaris for a temporary change, but the proper way is to use the chkconfig command or GUI programs such as system-config-services in Red Hat, or the System/Runlevel Editor module from yast2 in SUSE. For more information about chkconfig, refer to its man page or the Red Hat and Novell SUSE documentation.

134

Solaris to Linux Migration: A Guide for System Administrators

8.4 Shutting down

Shutting down a system in Solaris and Linux is almost the same. The shutdown, reboot, halt, poweroff, init, and telinit commands exist in both Solaris and Linux, but there can be minor differences in options, so check the man pages of those commands before using them. One thing you need to remember is that in Solaris run level 5 is for shutting down the system and turning off the power. In Linux, run level 5 is associated with a full multiuser running system with graphical login. Moreover, run level 0 in Solaris just brings down the operating system returning to the OpenPROM ok prompt, while in Linux it shuts down the operating system and also turns off the power if possible.

8.5 Network booting

Sometimes it is useful to boot from the network. Both Solaris and Linux support network booting. This section describes how to set up a machine to act as a network boot server and how to boot a Linux system from the network as a client.

Solaris network boot server setup

In Solaris, you need to configure the services shown in Table 8-6.
Table 8-6 Solaris network boot services Service rarp bootparams dhcp tftp nfs Description For serving IP address information For boot information Can be used instead of rarp and bootparams for network boot configuration For serving the proper kernel from /tftpboot For sharing file systems

This can be done manually by editing the configuration files, or automatically by using the add_install_client script from the installation CD or DVD.

Linux network boot server setup

In Linux, you need to configure the services shown in Table 8-7 on page 136.

Chapter 8. Boot and system initialization

135

Table 8-7 Linux network boot services Service dhcp tftp nfs Description For network boot configuration For serving the kernel and initrd image to boot For sharing file systems

Again, you can execute this task by modifying the configuration files or by using a GUI. The GUI is called system-config-netboot in RHEL, and in SLES, you can use the Network Services module from yast2 to configure dhcp/tftp/nfs services. You might need to manually configure support for network booting. Table 8-8 provides a list of the configuration files for controlling the server processes.
Table 8-8 Linux network boot configuration files File /etc/dhcpd.conf /etc/xinetd.d/tftp /etc/exports Description For DHCP services For TFTP services For NFS services

Then, make sure that the services are restarted and the firewall policy is not blocking any of the needed protocols.

Network boot services differences

Table 8-9 lists the differences between services needed for booting a Solaris system and those needed for booting a Linux system.
Table 8-9 Network boot services Services for network boot RARP Bootparams DHCP TFTP Solaris Needed if not using DHCP. Needed if not using DHCP. Not needed if using RARP and Bootparams. Needed for serving kernels. Linux Not needed; uses DHCP. Not needed; uses DHCP. Needed for client network configuration. Needed for serving kernels and initrd images.

136

Solaris to Linux Migration: A Guide for System Administrators

Services for network boot NFS

Solaris Needed for remote mounting of file systems.

Linux Not needed if the initrd image is enough or using other means for network file access.

Solaris network boot client

On Solaris, it is usually as simple as typing boot net at the OpenPROM ok prompt, but there are some exceptions. When the server was configured with the add_install_client script, the server is actually configured for serving a network install to the client that was specified. If this is not what you want, you need to boot with boot net -s to boot from the network in single user mode. In addition, if the server is configured for DHCP instead of RARP+Bootparams, you need to configure OpenPROM to use the DHCP protocol.

Linux network boot client

For booting an Intel-based Linux client from the network, determine if your network card has support for PXE booting. In addition, check the BIOS or system firmware to make sure that the network boot support is enabled. After the system is enabled to boot from the network, it automatically searches for a DHCP server, and then the DHCP server assigns an IP address and transfers a network file to boot from a TFTP server. This network boot file loads and presents a booting menu for choosing a specific kernel to boot, or it can automatically load and boot a default kernel, depending on the configuration. Tip: It is also possible to boot a Solaris client from a Linux server or a Linux client from a Solaris server, but this setup is possible only with manual configuration. For more information, read 3.3, Network-based install: Solaris and Linux heterogeneous environments on page 47.

Chapter 8. Boot and system initialization

137

138

Solaris to Linux Migration: A Guide for System Administrators

Chapter 9.

Managing system resources

This chapter describes the differences in managing system resources from Solaris to Linux. We describe these differences in the system resources in context with the tasks and commands used to control them. In this chapter, we discuss the following topics: Displaying system information Resource management Starting and stopping system services Scheduling and cron services Quotas Process accounting Remote system management services

Copyright IBM Corp. 2006. All rights reserved.

139

9.1 Displaying system information

Table 9-1shows some commands used to gather and report system information. On Linux, you can also use GUI tools to view this information. On Red Hat, you can use the hwbrowser command, and on Novell SUSE, use the yast command to launch the GUI tools.

Task table
Table 9-1 illustrates system information commands.
Table 9-1 System information commands System information task List system information List processor information List system memory size List disk space usage List file space usage List physical disks Show the systems host name List systems network configuration List active processes List system configuration List system diagnostic information Solaris uname psrinfo prtconf | grep Memory df du format hostname ifconfig prstat prtconf prtdiag Linux uname cat /proc/cpuinfo cat /proc/meminfo | grep MemTotal df du fdisk -l hostname ifconfig ps -e In Red Hat: lshal In Novell SUSE: hwinfo In Red Hat: lspci and /proc/* files In Novell SUSE: hwinfo

9.2 Resource management

In this section, we discuss optional tools for resource management.

Solaris Resource Manager

Solaris Resource Manager is a tool that helps the Solaris system administrator to manage the system resources for better system utilization and resource

140

Solaris to Linux Migration: A Guide for System Administrators

availability. If this is a product that you have used on Solaris systems, you might consider using the Class-based Kernel Resource Management framework in Linux, or IBM Tivoli Workload Scheduler.

Linux CKRM
Linux Class-based Kernel Resource Management (CKRM) is a framework for allocating system resources (such as CPU, memory, I/O, and network) based on user-defined classifications of work. CKRM gives a user-level application the ability to allocate hardware resources to applications as defined by the system administrator. CKRM is currently shipped with SUSE SLES9. For more information about how to use CKRM, visit the CKRM project Web site:
http://ckrm.sourceforge.net/

Tivoli Workload Scheduler

IBM Tivoli Workload Scheduler is an automated workload management solution that helps manage and operate mission-critical applications within a secure, fault-tolerant, and scalable IT infrastructure. It analyzes the status of the production work and drives the processing of the workload according to your business policies. It supports a multiple-end-user environment, enabling you to distribute processing and control across sites and departments within your enterprise. For more information, see:
http://www.ibm.com/software/tivoli/sw-bycategory/

9.3 Starting and stopping system services

System services are defined as processes or daemons that usually run in the background, are not associated with any user login, and do have a user interface.

9.3.1 Solaris system services

In Solaris, the /etc/rc*.d/ directory contains all the Solaris system startup and shutdown scripts, some of which are soft links to the actual file that resides in /etc/init.d/. Startup script names start with an S and shutdown scripts start with a K. To disable a service from running at startup time, a script is simply renamed so that it does not begin with a capital S. To manually start or stop a service, use the /etc/init.d/service {start, stop} commands.

Chapter 9. Managing system resources

141

Network services and inetd

Both Solaris and Linux configure the network service ports through the /etc/services file. Solaris uses the /etc/inetd.conf file to configure the Internet inetd processes.

9.3.2 Linux system services

On Linux, you can use the commands listed in this section, or use the user interfaces to manage system services. On Red Hat, use the system-config-services command to launch the GUI. On Novell SUSE, use the yast command. The /etc/init.d directory is the default location for system service startup scripts used to start and stop the service. The scripts listed there are named the same as the system service. Most startup scripts at least provide start, restart, and stop functionality and usually support a status option, for example:
/etc/init.d/dhcpd /etc/init.d/dhcpd /etc/init.d/dhcpd /etc/init.d/dhcpd start stop restart status

SUSE provides a set of soft links to these scripts for managing system resources. The linked files are /usr/sbin/rc*. These files correspond to the scripts in /etc/init.d. You can use either of these commands to check the status of the cron daemon:
/etc/init.d/cron status rccron status (in path from /usr/sbin)

Red Hat provides the service command. You can use either of these commands to check the status of the cron daemon:
/etc/init.d/crond status service crond status

Another useful version of the service command is:

service --status-all

Example 9-1 provides a sample output of the previous command.

Example 9-1 Sample output from service --status-all (shortened) # service --status-all acpid (pid 2242) is running... amd is stopped .. ..

142

Solaris to Linux Migration: A Guide for System Administrators

xinetd (pid 2319) is running... ypbind is stopped rpc.yppasswdd is stopped ypserv is stopped

Any of the previous services can be started or stopped by issuing the command:
/sbin/service smbd start

Or:
/sbin/service smbd start

Note that the service command is only available for Red Hat. The chkconfig command has the same function as the service command and is available on both Red Hat and SUSE.

The chkconfig command

The chkconfig command is used to add and enable system services to run at boot time. The chkconfig command exists and performs the same basic function on both Red Hat and Novell SUSE, but behaves slightly differently. For more information about the chkconfig command, refer to the man page. Example 9-2 shows a sample output of a chkconfig --list command.
Example 9-2 Sample output of chkconfig --list command (shortened) # chkconfig --list kudzu 0:off dhcp6s 0:off vsftpd 0:off anacron 0:off .. .. diskdump 0:off innd 0:off ntpd 0:off ypxfrd 0:off cyrus-imapd 0:off ypbind 0:off xinetd based services: time: off tftp: on .. .. amanda: off rlogin: off chargen-udp:

1:off 1:off 1:off 1:off

2:off 2:off 2:off 2:on

3:on 3:off 3:off 3:on

4:on 4:off 4:off 4:on

5:on 5:off 5:off 5:on

6:off 6:off 6:off 6:off

1:off 1:off 1:off 1:off 1:off 1:off

2:off 2:off 2:off 2:off 2:off 2:off

3:off 3:off 3:off 3:off 3:off 3:off

4:off 4:off 4:off 4:off 4:off 4:off

5:off 5:off 5:off 5:off 5:off 5:off

6:off 6:off 6:off 6:off 6:off 6:off

off

Chapter 9. Managing system resources

143

Note the different run levels for which the services are being enabled. Keep in mind that, in Linux, unlike Solaris, the run level is not run from the lowest to the desired run level, but instead, it only runs the startup scripts specified at the desired run level.

Network services and inetd

Linux uses the xinetd service and the /etc/xinetd.conf and /etc/xinetd.d directory entries for configuring the Internet processes. For more information about these services and configuration differences, refer to Chapter 7, Network services on page 107.

Task table
Table 9-2 illustrates system service management commands.
Table 9-2 Summary of system service commands Task Add a service that does not already exist in chkconfig Show existing service links Enable service to start at boot time Solaris Copy start script into the appropriate run level directory, prefixed with S ls run level directory Copy start script into the appropriate run level directory, prefixed with S Rename the service start script to a name that does not start with S (typically lowercase s) Red Hat chkconfig --add service name SUSE chkconfig -a or --add service name

chkconfig --list | grep service name chkconfig service name on

chkconfig service name chkconfig -s or --set service name on

Disable service from running at boot time

chkconfig service name off

chkconfig -s or --set service name off

9.4 Scheduling and cron services

This section discusses the command scheduling services and their differences. Scheduling involves being able to run commands at predefined times. The at command enables a command to run at a predefined time. The batch command works the same way, except that it waits until system load levels permit.

144

Solaris to Linux Migration: A Guide for System Administrators

The cron command allows for a command to be run at a predefined time, but offers much more flexibility than the at command. Both Solaris and Linux handle the two scheduling methods the same way. The anacron command is unique to Linux. This command is used just like cron, except that it allows for those scheduled commands that were not run due to the system being powered off or in suspended state to be run after the system is up and running normally again, sort of like a catch-up mode. This can be very useful in the case of desktops and notebooks that are sometimes powered off. The cron and anacron services use tables to schedule the jobs to run. For more information about configuring commands to run, refer to the man pages.

Task table
Table 9-3 illustrates scheduling and cron management commands.
Table 9-3 Scheduling and cron service commands Task Run a command at a predefined time Run a command or set of commands periodically Run a command or set of commands periodically, running when possible if system is off at the scheduled time Solaris at cron and crontab N/A Linux at cron and crontab anacron and anacrontab

9.5 Quotas
This section describes the differences in the quota commands. The quota commands are for managing disk usage and limits. The quota commands on Linux are very similar to those on Solaris; however, there are subtle differences in the command line parameters available. Check the man pages for further details about the differences.

Task table
Table 9-4 on page 146 illustrates quota commands.

Chapter 9. Managing system resources

145

Table 9-4 Quota commands Quota task Display disk usage and limits Edit users quotas Check consistency between a file system and the disk quota file Solaris quota edquota quotacheck Linux quota edquota quotacheck

9.6 Process accounting

This section describes the differences in the process accounting services. Process accounting is the method of keeping records on what commands are run and the user executing the command, as well as system resources used.

9.6.1 Solaris
The commands on Solaris are very similar to those available on Linux with the exception of the ac command. There is not an equivalent command with similar functionality.

9.6.2 Linux
On Linux, the commands differ somewhat from the commands on Solaris. The accton and lastcomm commands are very similar to those found on Solaris. Refer to the man pages for any differences in these commands.

Task table
Table 9-5 illustrates package commands.
Table 9-5 Process accounting commands Process accounting task Turn on process accounting List users connection time List commands run previously List users connection time and idle time Solaris accton none exists lastcomm who Linux accton ac lastcomm who

146

Solaris to Linux Migration: A Guide for System Administrators

9.7 Remote system management services

This section describes the differences in the common remote system management services between Solaris and Linux. You might or might not have used these services on Solaris, but you might want to consider their use when you migrate to Linux. These system management services enable a server or set of servers to be configured and monitored remotely using several different optionally available management frameworks or applications. Web-Based Enterprise Management (WBEM) and the Common Information Model (CIM) are the object-based architecture standards and data model promoted by the industry consortia Distributed Management Task Force (DMTF). The WBEM services described here are implementations of these standards. For more information about these standards, refer to the DMTF Web site:
http://www.dmtf.org

Another commonly used remote system management service is the Simple Network Management Protocol (SNMP). SNMP is an Internet Engineering Task Force (IETF) standard and is perhaps the most widely implemented standard for system management. The data model used by an SNMP service is called a Management Information Base (MIB). Different sections of this MIB are supported by different devices and servers depending on the services available on the system. For more information about the SNMP service, refer to the ITEF Web site:
http://www.ietf.org

9.7.1 Web-Based Enterprise Management (WBEM)

WBEM services are made available for use by a system management application through a CIM object manager or CIM server. Providers are written that gather and manage system resources. These providers register their classes with the CIM server to enable these objects to be used by a management application or even other providers. There are increasing numbers of applications adopting this new management standard. For more information about the components in the WBEM architecture, refer to the DMTF Web site:
http://www.dmtf.org/standards/wbem

Solaris
Some of the Solaris commands for the WBEM services include: init.wbem wbemadmin wbemlogviewer Used to start, stop, or retrieve status from the CIM server. Sun WBEM user manager. WBEM log viewer GUI.

Chapter 9. Managing system resources

147

The CIM server on the Sun Solaris platform is managed with the inet.wbem command. The inet.wbem command is in the /etc/initd directory. The init.wbem command accepts the start, stop, and status parameters.

Red Hat
Starting in RHEL4 Update 2, the OpenPegasus CIM server is available. For more information about the OpenPegasus CIM Server, refer to the OpenPegasus Web site:
http://www.openpegasus.org

Additionally, you can read more about how to set up and use the service by reviewing the documents in the /usr/share/doc/tog-pegasus-* directory. Some of the commands available for the OpenPegasus WBEM services include: cimconfig cimserver cimauth wbemexec cimprovider Provides command line interface to manage CIM server configuration properties. Starts and stops the CIM server. Adds, modifies, removes, or lists CIM user authorizations. Provides command line interface for querying CIM objects. Disables, enables, removes, and lists registered CIM providers or CIM provider modules and module status.

OpenPegasus is started and stopped with the cimserver command in /usr/sbin, or with the tog-pegasus command in /etc/inet.d. These commands can only be executed as a privileged user. The cimserver command starts the service, and using the -s parameter, stops the service. The osinfo command is useful for checking the status of the cimserver process and it also reports the system information. For more information about these and other commands, refer to the man pages. To enable the service to start at run time, run the chkconfig commands:
chkconfig --add tog-pegasus chkconfig tog-pegasus on

Then to start the service, use the tog-pegasus command in /etc/init.d:

/etc/init.d/tog-pegasus start

Or run the cimserver command. There are a very basic set of providers that are included with the OpenPegasus package. To make the service more useful, a more extensive set of

148

Solaris to Linux Migration: A Guide for System Administrators

instrumentation for Linux is available through the Standards Based Linux Instrumentation for Manageability (SBLIM) project. For more information about the SBLIM project, refer to the project Web site:
http://sblim.sourceforge.net

Novell SUSE
On SUSE, the OpenPegasus CIM server is also available, but it is on the SDK CD and must be installed separately. The default CIM server for SUSE is the OpenWBEM CIM server. For more information about the OpenWBEM CIM server, refer to the OpenWBEM Web site:
http://www.openwbem.org

In addition, there is documentation about how to configure and use the service in the /usr/share/doc/packages/openwbem directory. Some of the commands available for the OpenWBEM service include: owcimomd rcowcimomd owenumclasses OpenWBEM CIM server. Starts, stops, and queries status of CIM server. Lists CIM classes.

The OpenWBEM service can be started by using the owcimomd command in /usr/sbin. More information about these and other commands can be learned from the man pages. To start the service and to enable it to start at run time, run the chkconfig command:
chkconfig owcimomd on

The more extensive set of instrumentation for Linux referenced earlier, the SBLIM providers, are included in the Novell SUSE release. You can install these providers to greatly expand the amount of management information available.

9.7.2 Simple Network Management Protocol (SNMP)

SNMP services are implemented through a background server or daemon. This daemon listens for requests, and if the object is supported in the set of MIB extensions maintained by the server, the action is performed and the response returned. The SNMP service also has the ability to send and receive system events called traps. And, with the deployment and wide spread adoption of secure SNMPv3, it is now even safer to use.

Chapter 9. Managing system resources

149

Solaris
The Solaris SNMP service is started at boot time through the startup script /etc/rc3.d/S76snmpdx. To start and stop the service manually, run the script with the start and stop command line parameters:
/etc/rc3.d/S76snmpdx start /etc/rc3.d/S76snmpdx stop

The configuration file, snmpd.conf, is in /etc/snmp subdirectory for Red Hat, and in /etc for Novell SUSE.

Linux
Both Red Hat and Novell SUSE include the Net-SNMP package as their default SNMP service. For more information about the Net-SNMP package, refer to the Net-SNMP project page:
http://www.net-snmp.org/

You can also read more about Net-SNMP in the documents in the /usr/share/doc/packages/net-snmp directory. There are some excellent SNMP tutorials available at:
http://www.simpleweb.org/

Starting the Linux SNMP daemon is done either by running the snmpd command or enabling the service to start at boot time. There are many command line options, but usually, if started from the command line, the service is just simply started with the default command line options by running the snmpd command. For more information about the snmpd command, review the man page. To enable the service to start at run time, run the chkconfig command:
chkconfig snmpd on

To start the SNMP service:

/etc/inet.d/snmpd start

To configure the SNMP service, you need to edit the /etc/snmpd.conf file. The file has basic instructions for configuration, but you might want to refer to the man page for snmpd.conf as well. To make the SNMP service secure, you need to configure SNMPv3. A full tutorial about how to set this up is beyond the scope of this book. For more information, refer to the SNMPv3 readme in the /usr/share/doc/packages/net-snmp directory.

150

Solaris to Linux Migration: A Guide for System Administrators

10

Chapter 10.

Printing services
This chapter discusses commonalities and differences in printer management between Sun Solaris and Linux. In this chapter, we discuss the following topics in detail: Linux CUPS Client and server setup Printer management using the lp commands Printer types and CUPS support

Copyright IBM Corp. 2006. All rights reserved.

151

10.1 Overview
The printing subsystem on Linux is, on the surface, similar to the one on Solaris, supporting all the familiar BSD and System V UNIX-based lp command line utilities. Underneath, however, is the Common UNIX Printing System (CUPS), which is based on the new Internet Printing Protocol (IPP) standard. IPP has been adopted by many printer and print server makers, and was added by Microsoft into its Microsoft Windows 2000 product. While CUPS also supports the widespread UNIX-based LPD protocol, it is worthwhile to understand the Linux standard mechanism.

10.2 Linux CUPS

The Common UNIX Printing System (CUPS) is the standard for the Linux print service. It supports the Internet Printing Protocol (IPP) as its primary transport, but supports the line printer daemon (LPD) (client and server) as well. CUPS offers the following advantages over the earlier UNIX LPD/LPR mechanism: Individual setting of print parameters at user level Better support for PostScript printers Network browsing Web-based user and administrator front ends IPP defines a standard protocol for printing, as well as managing print jobs and printer options such as media size, resolution, and so forth. IPP also supports access control, authentication, and encryption. IPP is layered on top of HTTP. This enables users to view documentation, check status information about a printer or server, and manage their printers, classes, and jobs using their Web browser. CUPS supports the following protocols for network printing: IPP (preferred default), LPD (RFC 1179), SMB (Windows share), and socket. While we recommend using the CUPS guides provided by your Linux distribution first, for complete and detailed CUPS documentation and tutorials for both system administrators and users, refer to:
http://www.cups.org/ http://www.linuxprinting.org/

152

Solaris to Linux Migration: A Guide for System Administrators

10.3 Client and server setup

CUPS provides two methods for adding printers: lpadmin and a Web interface. The lpadmin command enables you to perform most printer administration tasks from the command line and is located in /usr/sbin. The Web interface is located at http://localhost:631/admin and steps you through the printer configuration. The IPP protocol is based on HTTP, and thus CUPS Web management can be done without the requirement of an HTTP server. The CUPS Web interface allows for full networked and remote printer management, after a CUPS administrator (who does not have to be root) is assigned using the lppasswd command. Additionally, your Linux distribution will usually provide a third way, a printer administration GUI (a yast2 module in SUSE, system-config-printer-gui in Red Hat). Both SUSE and Red Hat recommend that you attach and configure printers using their GUIs; you should follow your distributions recommendation until you have experience in using CUPS.

Locally attached printers

Linux CUPS supports printers attached through serial (including USB) or parallel ports. They can be added through all three interfaces. Follow the manufacturers instructions for installation, including installing any necessary device drivers. Use of the parallel port might require special configuration; consult the CUPS guide that came with your distribution. Restriction: For zSeries (S/390): Printers provided by zVM that you can connect locally with the zSeries hardware platform are not supported by CUPS. On these platforms, network printing is required.

Network printing
Linux CUPS supports networked printing through IPP, LPD (RFC 1179), SMB (Windows share), or direct socket. Configuration is possible for all through the command line, Web, or GUI. IPP is the preferred method between two CUPS servers or between the CUPS server and client. CUPS also supports printers attached through Novell NetWare technology and HP JetDirect system.

Chapter 10. Printing services

153

10.4 Printer management using the lp commands

The basic BSD and System V UNIX-based command line utilities are available on Linux as well as Solaris OS: lp lpadmin lpc lpr lpstat lpq lprm lpmove enable disable accept reject cancel Print files Configure the printer services and classes Line printer control Submit print requests Display print service status information Display print queue status Cancel print jobs Move print requests Start the named printers Stop the named printers Accept print jobs for the specified destination Reject print jobs for the specified destination Cancel print request

While the basic option/flag sets of these commands are the same, there are differences in details based on how CUPS manages printers.

Solaris specific
Solaris provides some additional utilities to manage the print services, which are not available on Linux. They are as follows: lpsched lpset lpshut lpsystem lpusers lpforms lpget lpfilter Start the LP print service daemon Set printing configuration in /etc/printers.conf Stop the lp print service Register remote systems with print service Set printing queue priorities Administer forms used with the lp service Get printing configuration Administer filters used with the lp service

154

Solaris to Linux Migration: A Guide for System Administrators

Linux specific
Linux also provides some additional functions and capability to manage the print services. They are as follows: cupsd lpinfo lpoptions lppasswd Start the common UNIX printing system daemon Show available devices or drivers Display or set printer options and defaults Add, change, or delete digest password in CUPS

Important: Most Linux distribution vendors advise against configuring the printing system through the direct editing of the configuration files. The CUPS service dynamically builds the configuration files each time it is started, and changes will be lost. The /etc/printcap file (for example) is created only for the purpose of compatibility with printcap-based applications.

Task table
Table 10-1 organizes the command comparison by administrator task.
Table 10-1 Print service management task commands Task Run multiple tasks in a GUI environment Add a printer Start a print queue Stop a print queue Display print queue status Cancel a print job Add a print queue Change a print queue Remove a print queue Display settings of a print queue Set print queue settings Print files Solaris admintool lpadmin accept, enable, or lpc reject, disable, or lpc lpstat or lpc cancel lpadmin lpadmin lpadmin lpadmin lpadmin lp or lpr Red Hat system-config-printe r-gui lpadmin accept, enable, or lpc reject, disable, or lpc lpc or lpq cancel or lprm lpadmin lpadmin lpadmin lpadmin or lpoptions lpoptions lp or lpr SUSE yast2 lpadmin accept, enable, or lpc reject, disable, or lpc lpc or lpq cancel or lprm lpadmin lpadmin lpadmin lpadmin or lpoptions lpoptions lp or lpr

Chapter 10. Printing services

155

Task Display print service status information Move the print request to another print queue Print services daemon Start and stop script for the print services daemon Put printing config information in the system config database Display printing config information from the system config database Register remote print queues in local system Set printing queue priorities for users Configure forms for the print queue Apply filters to the print queue Display available devices or drivers Manage digest password in CUPS

Solaris lpstat lpmove lpsched /etc/init.d/lp or lpshut lpset

Red Hat lpstat lpmove cupsd /etc/rc.d/cups

SUSE lpstat lpmove cupsd /etc/rc.d/init.d/cups

N/A

N/A

lpget

N/A

N/A

lpsystem lpusers lpforms lpfilter N/A N/A

N/A N/A N/A N/A lpinfo lppasswd

N/A N/A N/A N/A lpinfo lppasswd

10.5 Printer types and CUPS support

There are three classes of printers to consider, when matching a printer to a Linux/CUPS system, and the division is by language.

156

Solaris to Linux Migration: A Guide for System Administrators

PostScript
CUPS supports these printers by default, using either supplied PostScript Printer Description (PPD) files for your brand and model of printer, or those available from your printer manufacturer, or from places such as:
http://www.linuxprinting.org/

Standard printer languages (PCL, ESC/P, or similar)

There are many printers, laser as well inkjet, that use these languages for processing. The PCL language is mostly used in certain lines of HP printers (and their clones). ESC/P is used by some Epson models. Linux can usually support these printers well through a filtering process known as interpreting (usually by using Ghostscript tools, which come with Linux). There are very few native Linux drivers for these printers, with the exception of HP products. CUPS comes with special filters, known as foomatic (in the cups-drivers RPM package) and rastertoprinter (the cups-drivers-stp RPM package) to handle these printers if they do not have their own drivers. Consult the CUPS section of the Help Guide that comes with your Linux distribution.

Proprietary
A few of the least expensive printers use proprietary communication models and come with Windows-only drivers. Lexmark inkjet printers are a common example of this. These are usually poorly-supported by Linux, if they are supported at all. You might find help for some models at:
http://www.linuxprinting.org/

Chapter 10. Printing services

157

158

Solaris to Linux Migration: A Guide for System Administrators

11

Chapter 11.

Users and groups

This chapter discusses basic user and group administration tasks and utilities, including: Commands Directory services NIS and NIS+ User ID and group ID differences

Copyright IBM Corp. 2006. All rights reserved.

159

11.1 Basic user administration

In general, user and group administration is the same as on most UNIX-based systems, including Solaris. Configuration files, such as /etc/group, are in the same locations and are used in the same manner. Most user administration tasks can be performed either from the command line or from your Linux distributions GUI (for example yast2 on SUSE and system-config-users on Red Hat). Red Hat starts regular user IDs at 500; SUSE Linux starts them at 1000.

11.2 Commands
The following command line-based user ID administration tools have very similar basic functionality between Solaris and Linux: useradd usermod userdel For these command line-based tools, functions such as adding comments, setting home directory path, base group ID, and other group IDs, creating a home directory, using skeleton profiles, and the default shell settings are the same between Solaris and Linux.

11.2.1 Adding user accounts

The key differences for the useradd command are as follows: Solaris: -P and -R options -p option -A option Red Hat: -p option -r option Assign a user passwd when adding the user account. Assign a UID that is lower than the UID_MIN set in /etc/login.defs, which can make the user ID a system account. The useradd command creates a user home directory as per system-wide the configuration parameter set in Add RBAC profile and role in the useradd command. Assign a project name to the user. Assign an authorization attribute for executing restricted functions.

-M option

160

Solaris to Linux Migration: A Guide for System Administrators

/etc/login.defs. By giving it this parameter, it overwrites that option to not create it. -n option The useradd command creates a group name to the same user name by default. By giving this parameter, it will not create the group name with the same name as the user name. Not to add the user to the last login log. Assign a user passwd when adding the user account. Assign a UID that is lower than the UID_MIN set in /etc/login.defs, which can make the user ID a system account.

-l option SUSE: -p option -r option

-D or --binddn <binddn> option Add the user using the Distinguished Name binddn to bind to the LDAP directory. -P or --path <path> option Add the user ID in a different path for the passwd and shadow files. --service option --usage option Add the user account using the special service, which can be a local file, LDAP, NIS, or NIS+. Display a short list of valid options.

--show-defaults option Display the default settings for the user parameters: homedir, expiry, inactivity, gid, other groups, and default shell. --save-defaults options Change the default settings for the user parameters. Note: Both Solaris and Red Hat use the -D option to set and display the user parameter defaults. However, SUSE Linux uses the -D option for a different purpose.

11.2.2 Changing user account information

The differences between Solaris, Red Hat and SUSE for the usermod command are the same as the previous list, although some of the previous options are specific to the useradd command. The usermod utility can change most of the

Chapter 11. Users and groups

161

option settings defined in the previous list. The command differences are as follows: Solaris: The usermod command enables you to change most of the parameters that are set by the useradd command. Red Hat: This also enables you to change most of the parameters that are set by the useradd command, with the following available options: -L option -U option SUSE: The usermod does not allow the changes to any of the basic user account parameter except for the following changes: -D <binddn> option -P <path> option Change the Distinguished Name binddn. Change the path for the passwd and shadow files. Change the user accounts password. -r <service> option Change the accounts special service procedure. Either local files, LDAP, NIS, or NIS+. Lock the user account. Unlock the user account.

-p <new password> option

11.2.3 Removing user accounts

There are no differences between Solaris and Red Hat for the userdel command. The -r option works the same way on both. However, SUSE does have differences, including: -D <binddn> or --binddn <binddn> option Add or remove the user account using the Distinguished Name binddn to the LDAP directory. -P <path> or --path <path> option Remove the user account from the specified directory to passwd and shadow file. --service <service> option Remove the user account using the special service. Either local files, LDAP, NIS, or NIS+. -f option Force the removal of files in the users home directory even if the files are not owned by the account.

162

Solaris to Linux Migration: A Guide for System Administrators

11.2.4 Home directory

Linux creates user home directories by default into /home. It has a location for creating a default user profile, a set of preconfigured files (such as .profile or X11 desktop files) used to populate a newly created users home directory called /etc/skel/*.

11.3 Directory services

The Solaris version of LDAP uses the iPlanet Directory Server. Solaris 8 uses the 4.11 version and Solaris 9 uses the 5.1 version. There is also the more recent Sun One Directory Server (5.2). Both Red Hat and SUSE Linux use OpenLDAP (http://www.OpenLDAP.org/), a software package founded with support from IBM. This package provides both a stand-alone and update replication daemon, as well as libraries, utilities, Java classes, and sample clients. OpenLDAP supports LDAPv3 (RFC 3377) with the addition of the following optional features: Schema updates over LDAP through cn=config (2.3 only) Component Matching (RFC 3687) (2.3 only) Language Tag and Range options (RFC 3866) DNS-based service location (RFC 2247 & RFC 3088) Password Modify operation (RFC 3062) Named Referrals/ManageDSAit control (RFC 3296) Simple Paged Result Control (RFC 2696) All Operational Attributes plus attribute list feature (RFC 3673) supportedFeatures discover mechanism (RFC 3674) Content Synchronization operation WhoAmI? operation Proxy Authorization control Matched Values control (RFC 3876) Assertion control Pre/Post Read controls No-Op control Modify/Increment extension Absolute True (&) and False (|) Filter extension

Chapter 11. Users and groups

163

OpenLDAP is also available for Solaris. In addition to the documentation provided by SUSE and Red Hat, for a good how-to for LDAP administration, refer to the Linux Documentation Project (http://www.tldp.org/HOWTO/LDAP-HOWTO/index.html). There are modules (pam_ldap and nss_ldap) available to allow for workstation authentication against LDAP. The /etc/ldap.conf and /etc/nsswitch.conf files have the same content on both operating systems. If you plan to connect Linux LDAP clients to Suns LDAP offerings, the connection should be straightforward. If you are moving from Suns offerings to a Linux LDAP server solution, keep in mind that authentication methods and schemas will probably differthe LDAP technical community have not completely standardized in these areas. The basic configuration command on RHEL4 is system-config-authentication, and on SUSE the YaST2 GUI contains modules for configuring either an LDAP client or server. The OpenLDAP administration commands themselves are: ldapadd ldapcompare ldapdelete ldapmodify ldapmodrdn ldappasswd ldapsearch ldapwhoami Add entry Perform compare based on parameters Delete entry modify entry Rename (modifies RDN) entry Set password of entry Connect to server and searches based on parameters Perform LDAP whoami operation

11.4 NIS and NIS+

Linux contains full support for name switch service (NSS) and has support for the following NIS (and NIS+) maps: aliases, ethers, group, hosts, netgroups, networks, protocols, publickey, passwd, rpc, services, and shadow. It fully supports shadow passwords over NIS. Although NIS+ is supported, its use is not recommended on Linux, because LDAP is the favored solution in this space, and some NIS+ problems are still present. Linux configures in a similar manner as Solaris. For a complete NIS/NIS+ (and YP) how-to, refer to:
http://www.tldp.org/HOWTO/NIS-HOWTO/

164

Solaris to Linux Migration: A Guide for System Administrators

11.5 User ID and group ID differences

There are differences in the UIDs and GIDs used for the various users defined on the system by default. Table 11-1 lists UID differences. Table 11-2 on page 168 lists the GID differences.
Table 11-1 User ID differences User root daemon bin sys adm uucp nuucp smmsp Solaris UID:GID 0:1 1:1 2:2 3:3 4:4 5:5 9:9 25:25 Red Hat UID:GID 0:0 2:2 1:1 N/A 3:4 10:14 N/A 51:51 SUSE UID:GID 0:0 2:2 1:1 N/A N/A 10:14 N/A N/A Comment Superuser N/A N/A N/A Admin uucp admin uucp admin SendMail Message Submission Program Network admin Line printer admin Nobody No access user SunOS 4.x Nobody

listen lp nobody noaccess nobody4 sync shutdown halt mail news operator games

37:4 71:8 60001:60001 60002:60002 65534:65534 N/A N/A N/A N/A N/A N/A N/A

N/A 4:7 99:99 N/A N/A 5:0 6:0 7:0 8:12 9:13 11:0 12:100

N/A 4:7 65534:65533 N/A N/A N/A N/A N/A 8:12 9:13 N/A 12:100

Chapter 11. Users and groups

165

User gopher man ftp squid pvm named at postgres mysql ncsd mdom rpcuser wwwrun rpc amanda netdump rpm ntp canna irc mailman gdm

Solaris UID:GID N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

Red Hat UID:GID 13:30 N/A 14:50 23:23 24:24 25:25 N/A 26:26 27:27 28:28 N/A 29:29 N/A 32:32 33:6 34:34 37:37 38:38 39:39 N/A 41:41 42:42

SUSE UID:GID N/A 13:62 40:49 31:65534 N/A 44:44 25:25 26:26 60:2 N/A 28:28 N/A 30:8 N/A 37:6 104:104 N/A 74:65534 N/A 39:65534 72:67 50:15

Comment

Man pages viewer FTP user Squid proxy server Parallel processing pkg

Batch daemon PostgreSQL server mySQL server ncsd daemon Mailing list agent RPC service user WWW daemon Apache Portmapper RPC user Amanada backup suite netdump Package manager

Canna service users IRC daemon GNU mailing list mgr GNOME desktop

166

Solaris to Linux Migration: A Guide for System Administrators

User xfs mailnull apache wnn ldap vscan webalizer pop haldaemon vcsa snort sshd radvd

Solaris UID:GID N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

Red Hat UID:GID 43:43 47:47 48:48 49:49 55:55 N/A 67:67 N/A 68:68 69:69 N/A 74:74 75:75

SUSE UID:GID N/A N/A N/A N/A 76:70 65:103 N/A 67:100 N/A N/A 73:68 71:65 N/A

Comment X11 Font Server

Apache Wnn input server LDAP user Virus scanner Webalizer POP server HAL daemon Virtual console memory owner Snort network monitor Privilegeseparated SSH Router advertisement daemon Cyrus IMAP server Network monitor user mgetty fax spool System message bus Mail server Quagga routing suite Spam/virus pkg Radius user

cyrus pcap fax dbus postfix quagga exim radiusd

N/A N/A N/A N/A N/A N/A N/A N/A

76:12 77:77 78:78 81:81 89:89 92:92 93:93 95:95

96:12 N/A N/A N/A 51:51 101:101 N/A 102:102

Chapter 11. Users and groups

167

User dovecot ident htt stunnel dhcpd nfsnobody

Solaris UID:GID N/A N/A N/A N/A N/A N/A

Red Hat UID:GID 97:97 98:98 100:101 N/A N/A 65534:65534

SUSE UID:GID N/A N/A N/A 100:65534 103:65534 N/A

Comment IMAP/POP3 server

IIIMF Htt SSL tunnel daemon DHCP daemon

Table 11-2 Group ID differences Group root other bin sys adm uucp mail tty lp nuucp staff daemon sysadmin smmsp nobody noaccess nogroup disk Solaris 0 1 2 3 4 5 6 7 8 9 10 12 14 25 60001 60002 65534 N/A Red Hat 0 N/A 1 3 4 14 12 5 8 N/A N/A 2 N/A N/A 99 N/A N/A 6 SUSE 0 N/A 1 3 N/A 14 12 5 7 N/A N/A 2 N/A N/A 65533 N/A 65534 6

168

Solaris to Linux Migration: A Guide for System Administrators

Group mem www kmem wheel news man shadow dialout audio floppy games cdrom slocate console utmp squid pvm named at postgres mysql nscd mdom rpcuser gopher rpc public

Solaris N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

Red Hat 8 N/A 9 10 13 15 N/A N/A N/A 19 20 N/A 21 N/A 22 23 24 25 N/A 26 27 28 N/A 29 30 32 N/A

SUSE N/A 8 9 10 13 N/A 15 16 17 19 40 20 N/A 21 22 N/A N/A 44 25 26 N/A N/A 28 N/A N/A N/A 32

Chapter 11. Users and groups

169

Group video netdump rpm ntp canna dip mailman xok gdm trusted xfs modem mailnull apache wnn ftp smmsp lock ldap maildrop man pkcs11 sshd webalizer haldaemon snort vcsa

Solaris N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

Red Hat N/A 34 37 38 39 40 41 N/A 42 N/A 43 N/A 47 48 49 50 51 54 55 N/A N/A N/A N/A 67 68 N/A 69

SUSE 33 N/A N/A N/A N/A N/A 67 41 N/A 42 N/A 43 N/A N/A N/A 49 N/A N/A 70 59 62 64 65 N/A N/A 68 N/A

170

Solaris to Linux Migration: A Guide for System Administrators

Group ntadmin sshd radvd pcap fax dbus postfix postdrop quagga exim radiusd dovecot ident users htt quaggavt vscan dump nfsnobody

Solaris N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

Red Hat N/A 74 75 77 78 81 89 90 92 93 95 97 98 100 101 102 N/A N/A 65534

SUSE 71 N/A N/A N/A N/A N/A 51 N/A 101 N/A 102 N/A N/A 100 N/A N/A 103 104 N/A

Chapter 11. Users and groups

171

172

Solaris to Linux Migration: A Guide for System Administrators

12

Chapter 12.

Monitoring and performance

This chapter provides information about the commands available to monitor the utilization of, and display the statistics for, the physical and logical resources of your system. The parameters for the commands are generally the same on Solaris and Linux. Many of the commands provide information about several resources. The output from the commands might differ significantly. The physical components of a Linux server are its: Processors (CPUs) Real memory (RAM) Physical media, such as disk drives and floppy disks Network devices The logical components of a Linux server are its: Virtual memory (SWAP) Software RAID drives Logical volume groups and logical volumes File systems System and user processes, such as daemons and applications

Copyright IBM Corp. 2006. All rights reserved.

173

Certain tools are installed by default. You should review these packages and install them if your installation requires the tools provided by the packages: For RHEL: sysstat package: sar, iostat sysreport package: sysreport (system information collector) For SLES: sysstat package: sar, iostat siga package: siga (System Information GAthering) Many of the commands discussed in this chapter have options that allow them to be run iteratively. The output from these commands will eventually scroll off the screen. You can save the output several different ways. For example: Execute the command in a script subshell. Redirect the output to a file with the redirect operator > or >>. Redirect the output to a file with the tee command and view it at the same. Execute the command with the watch command. Here are three more points to consider: Log files showed be reviewed on a regular basis. You can use the dmesg command to display the current messages in the kernels ring buffer. The /proc file system contains system-wide and process-specific configuration and state information. There are configuration files associated with some of the commands in this chapter. Refer to the documentation for each command to find the location and format of its configuration file. Linux kernel parameters can be displayed with the sysctl -a command. Setting kernel parameters is outside the scope of this chapter.

174

Solaris to Linux Migration: A Guide for System Administrators

12.1 Processors
Processor configuration information is in The /proc/cpuinfo file. Table 12-1 describes the differences in CPU monitoring commands.
Table 12-1 CPU monitoring commands Solaris iostat mpstat psrinfo, sar -r ps sag top uptime vmstat w Linux iostat mpstat procinfo ps isag top uptime vmstat w Comments Display I/O statistics Display individual processor statistics in a multiprocessor system Display system status for CPU and memory usage Per process CPU utilization and time Display sar data graphically Display top CPU user for running processes and CPU and memory statistics Load average for the past 1, 5 and 15 minutes Display virtual memory statistics Display who is currently logged in

In addition to these tools, a profiling suite known as OProfile is available. OProfile is a system-wide profiler for Linux systems, capable of profiling all running code without too much extra impact on system load. It consists of a kernel driver and a daemon for collecting sample data and several post-profiling tools for turning data into information. This tool ships with RHEL4. In SUSE, the kernel is enabled for it (the driver portion) and the daemon is available from:
http://oprofile.sourceforge.net/

12.2 Real and virtual memory

Physical and virtual memory are monitored by the same commands. The /proc/meminfo file contains detailed information concerning memory usage. Refer to Table 12-2 on page 176 for a description of memory monitoring commands. In addition, Linux provides the GNOME System Monitor, a GUI for monitoring resources.

Chapter 12. Monitoring and performance

175

Table 12-2 Real and virtual memory monitoring commands Solaris sar -r psrinfo sag sar top vmstat Linux free procinfo isag sar top vmstat Comments Display amount of free and used memory Display system status for CPU and memory usage Display sar data graphically Display, collect, or store system activity status Display top CPU user for running processes and CPU and memory statistics Display virtual memory statistics

12.3 Physical media, software RAID, LVM, and file systems

Physical media must be partitioned and formatted with a file system unless it will be used as a raw device. For this reason, we differentiate between physical media and file systems in the section. Commands to monitor disk partitions, which are analogous to Solaris slices, will be listed as physical media monitoring commands. Software RAID and LVM provide complementary functions. LVM does not replace RAID, nor does software RAID replace LVM.

Task table: Physical media monitoring

Table 12-3 illustrates the physical media monitoring commands.
Table 12-3 Physical media monitoring commands Solaris iostat sag sar vmstat Linux iostat isag sar vmstat Comments Display I/O statistics Display sar data graphically Display, collect, or store system activity status -a and -d parameters valid on Linux kernels 2.5.x and later

Task table: Software RAID

Linux provides a different set of utilities for software RAID devices and logical volume management. This section contains commands specific to monitoring

176

Solaris to Linux Migration: A Guide for System Administrators

software RAID devices. Table 12-4 illustrates software RAID monitoring commands.
Table 12-4 Software RAID monitoring commands Solaris metastat metadb Linux cat /proc/mdstat lsraid (SUSE only; deprecated in Red Hat, use mdadm instead) mdadm sgraidmon (SUSE only) Comments Display status information for software RAID devices List and query MD devices

mdmonitord N/A

Manage and monitor software RAID devices Monitor SCSI devices for hot-removal and hot-insertion events

12.3.1 Logical volume groups and logical volumes

Linux provides a different set of utilities for software RAID devices and logical volume management. This section contains commands specific to monitoring logical volume groups and logical volumes. The commands to manage logical volumes, groups, and devices are in these directories: SUSE: /sbin Re Hat: /usr/sbin Both Red Hat and SUSE provide LVM2. SUSE also provides the Enterprise Volume Management System (EVMS). The Linux 2.6 kernel prevents multiple volume managers from managing the same device. EVMS will display information about LVM2-managed devices by default. EVMS can be configured to exclude information about these devices.

Task table: Logical volume monitoring

Table 12-5 illustrates logical volume monitoring commands.
Table 12-5 Logical volume monitoring commands Solaris metastat metastat metadb Linux lvdisplay lvs pvdisplay Comments Display logical volume attributes Report logical volume information Display physical volume attributes

Chapter 12. Monitoring and performance

177

Solaris metadb metastat metastat

Linux pvs vgdisplay vgs

Comments Report physical volume information Display volume group attributes Report volume group information

12.3.2 File systems

We differentiate file systems from physical devices in this chapter, because a file system can extend beyond a single device. Examples of file systems that extend beyond a single device are hardware and software RAID drives and logical volumes.
Table 12-6 File system monitoring commands Solaris df du repquota N/A Linux df du repquota stat Comments Display file system disk space usage information Display individual file disk space usage

12.4 Network
Table 12-7 illustrates some network activity monitoring commands. Additional network performance information is available in /proc/slabinfo.
Table 12-7 Network monitoring commands Solaris N/A ifconfig N/A netstat nfsstat Linux ethereal ifconfig mrtg netstat nfsstat Comments Graphical network protocol analyzer. Available for Solaris, but not installed by default. Configure network device. Monitor traffic load. Available for Solaris, but not installed by default. Display network statistics, routing information, connections, and so on. Display NFS statistics.

178

Solaris to Linux Migration: A Guide for System Administrators

Solaris ping sag sar snoop spray tcpdump

Linux ping isag sar N/A N/A tcpdump, tethereal

Comments Send ICMP echo request packets to network host. Display sar data graphically. The -n parameter on Linux can be used to report different types of network activity. Display network packets. Test network connectivity using a UDP protocol. Dump network traffic.

12.5 System and user processes

Each process has a directory containing its state data. The directory is /proc/nnnn, where nnnn is the process ID (PID) number. Table 12-8 lists some process monitoring commands.
Table 12-8 Process monitoring commands Solaris ipcs ldd proc tools ldd ps pstree sag sar truss top vmstat Linux ipcs ldd lsof ltrace ps pstree isag sar strace top vmstat Comments Display interprocess communication facilities status Display shared library dependencies Display open files Display dynamic library calls Display process status Display process inter-dependencies in a tree format Display sar data graphically Display, collect, or store system activity status Trace system calls and signals Display top CPU user for running processes and CPU and memory statistics Display virtual memory statistics

Chapter 12. Monitoring and performance

179

180

Solaris to Linux Migration: A Guide for System Administrators

13

Chapter 13.

Backup and restore

This chapter describes some basic UNIX backup and compression tools and the differences in usage between Solaris and Linux. We discuss the following topics: Common backup tools Compression tools File system backup and restore File system snapshots AMANDA

Copyright IBM Corp. 2006. All rights reserved.

181

13.1 Common backup tools

Performing backups is a critical component of system maintenance. UNIX has several options available for the system administrator to use. Across UNIX, the most common tools available for backup are: tar cpio dd pax These commands operate the same way on Linux as they do on Solaris, although there might be some differences in functionality. Refer to 13.3, File system backup and restore on page 183 for a highlight of the additional GNU tar functions on Linux.

13.2 Compression tools

System administrators use compression tools as a quick way to recover file system space when it is at a premium. They also allow the system administrator or user to keep copies, possibly multiple copies, of earlier versions of files without taking up as much space. In addition, administrators can use these tools to compress files before writing them to an archive device. Whatever the usage might be, these are convenient tools for use in the system administrators day-to-day life cycle.

Common tools
The following tools are some of the tools available for all three operating systems discussed in this book: bzip2, bunzip2, bzcat, bzip2recover compress, uncompress gzip, gunzip, gzcat gznew (recompress .Z files to .gz files) gzexe (compress executable files in place) pack, pcat, unpack zip, zipcloak, zipnote, zipsplit There are other programs that come with some of these tools that do specific functions. For example, gzgrep provides the ability to grep within a compressed

182

Solaris to Linux Migration: A Guide for System Administrators

file and bzless, which enables a user to page a compressed file. For Solaris, some of these other programs are only available in Solaris 9, or as part of the Companion CD installation. For more details about how to use each tool, refer to their operating system man pages.

13.3 File system backup and restore

This section discusses the unique backup and restore tools used in Solaris and their functional equivalents in Linux.

Task table
Table 13-1 illustrates the basic file system backup and restore commands.
Table 13-1 File system backup commands File system backup task Full or incremental backup Restore Solaris ufsdump usfrestore Linux dump (for ext2/3 file system only) tar restore (for ext2/3 file system only) tar

13.3.1 Solaris overview

Solaris comes with a useful tool for doing file system backups called ufsdump. A limitation to this tool is that it only works on the UFS file system. This tool enables you to do full and incremental backups based on the parameter options given to the tool. The corresponding tool for recovery is ufsrestore. The disadvantage of any backup tool is that the file, or file system, being stored in the backup needs to be quiescent in order to avoid having partially updated files or file systems stored in the backup. ufsdump is an all-inclusive tool as far as being able to do incremental backups. However, you can also use pax if you want the archive to be portable across different versions of UNIX or platforms, or both.

13.3.2 Linux overview

Linux comes with the dump command that is specific to the ext2 and ext3 file systems. The program works at the inode-level and does not change the

Chapter 13. Backup and restore

183

last-access date stamps. This tool is very useful for doing a full backup. The corresponding tool to this is the restore command. The GNU tar command that comes in Linux has the following useful features: --atime-preserve: Option to not change access times on dumped files. --checkpoint: Print directory names while read the archive. -f, --file=[HOSTNAME:]<device>: Use archive file or device <device>. -F, --info-script=<file>, --new-volume-script=<file>: Run the script <file> at the end of each archive or tape. -g, --listed-incremental=<file>: Create, list, or extract new GNU-format incremental backup file. -h, --dereference: Do not dump symlinks; dump the files to which they point. -j, --bzip2, --bunzip2: Filter the archive through bzip2. -l, --one-file-system: Stay in local file system when creating an archive. -N, --after-date=<DATE>, --newer=<DATE>: Store only files newer than <DATE>. -V, --label=<NAME>: Create archive with volume name <NAME>. -Z, --compress, --uncompress: Filter the archive through compress. -z, --gzip, --ungzip: Filter the archive through gzip. For more information about the GNU tar command, refer to the man pages or run tar --help for a more updated list of available options.

13.4 File system snapshots

Using file system snapshots is another way to archive files while still allowing read/write activity on the file system. Performance degradation is only affected during the snapshot creation phase and when users want to update files after the snapshot is done. Care should be taken when the file system is large, because the snapshot size is correspondingly large.

Overview of Solaris task

Solaris comes with the fssnap command to create and manage a file system snapshot. After taking a snapshot, you can store the file system in a backup using the backup tool of our choice.

184

Solaris to Linux Migration: A Guide for System Administrators

Linux task
Linux does not come with a file system snaphot as Solaris does. The closest thing it has is the dump command, as referred to in the previous section. There is a script with about the same functionality as fssnap. It was originally written by Mike Rubel and is available at:
http://www.mikerubel.org

Look for the Rsync snapshot article.

13.5 AMANDA
AMANDA, the Advanced Maryland Automated Network Disk Archiver, is a public domain utility developed at the University of Maryland. It is a backup system that enables the administrator of a LAN to set up a single master backup server to back up multiple hosts to a single large capacity tape drive. AMANDA uses native dump or GNU tar facilities and can back up a large number of workstations running multiple versions of UNIX. Recent versions can also use SMB to back up Microsoft Windows hosts. For more information about this tool, see:
http://www.amanda.org/

Chapter 13. Backup and restore

185

186

Solaris to Linux Migration: A Guide for System Administrators

14

Chapter 14.

Security and hardening

This chapter discusses the differences in security and hardening configuration between Solaris and Linux. This chapter does not instruct the you about how to lock down a system, because those settings typically differ depending on how the system is used and what governance it must follow. We discuss the following topics: Patching and security updates Security hardening Securing and removing services Kernel tuning for security Logging Access control lists PAM umask SSH IPSec and IKE Kerberos Warning banners Firewalls

Copyright IBM Corp. 2006. All rights reserved.

187

14.1 Patching and security updates

The security of a system is only as good as its weakest link. While it is important for systems to be properly secured when they are initially installed and put on a network, it is imperative that each system be regularly patched against new security vulnerabilities. Solaris, Red Hat, and Novell SUSE are all supported though vendor services that release security patches, alerts, and software upgrades at the following locations: Solaris:
http://sunsolve.sun.com/pub-cgi/show.pl?target=security/sec

Red Hat:
https://www.redhat.com/apps/support/

SUSE:
http://www.novell.com/linux/security/securitysupport.html

14.2 Security hardening

Security hardening can be as important with Linux as it is with Solaris or any other type of UNIX-like operating system. Many out-of-the-box installations of Linux have numerous features that are disabled, but to really harden a system before putting it online, consider using a hardening tool.

14.2.1 Hardening tools

Most Solaris administrators are familiar with Yet Another Solaris Security Package, YASSP (http://www.yassp.org/) and the Solaris Security Toolkit. These tools are used to more efficiently harden a Solaris system. In Linux, one of the most popular tools used to implement system hardening is called Bastille.

Bastille Linux is an open source project aimed at hardening a Linux system by

changing some of the systems settings based on security best practices. For more information about Bastille, see:
http://www.bastille-linux.org

Bastille is a useful tool for system administrators to discover and tighten Linux security settings. A graphical user interface enables administrators to choose security settings appropriate for their environment. Bastille automatically implements changes based on selections provided by the user.

188

Solaris to Linux Migration: A Guide for System Administrators

14.2.2 Auditing: ASET/BSM

As we discussed in the last section, Linux can be hardened using a tool called Bastille. Many Solaris system administrators are familiar with using the Automated System Enhancement Tool (ASET) and Basic Security Module (BSM) as ways to set basic security parameters, keep those settings configured appropriately, and monitor security levels. On the Linux platform, Bastille can certainly help in creating a secure Linux configuration, but for more enhanced auditing functionality with features similar to ASET and BSM, look at Security Enhanced Linux (SE-Linux). SE-Linux is a special Red Hat package that was created by the National Security Agency (NSA). It provides role-based access control and additional auditing features. For more information about SE-Linux, see:
http://www.nsa.gov/selinux/

14.3 Securing and removing services

Both Solaris and Linux have many capabilities that are necessary in one environment but not necessarily in another. Additionally, some services are safe to run in one environment, but that same service will open the system to attack in a different environment. This being the case, system administrators are always trying to allow only the necessary services to run on their systems to prevent unnecessary security exposure. This section focuses on a few key differences between the two environments that enable the system administrator to identify and disable unnecessary services.

14.3.1 inetd/xinetd
In Solaris, the more traditional Internet services are enabled or disabled in the file /etc/inetd.conf. As we discuss in 7.8.1, inetd-based versus xinetd-based network services on page 117, Linux uses a slightly different method than Solaris when it comes to inetd services. From a security perspective, it is important to ensure that all unnecessary services are disabled and TCP wrappers are set to monitor all enabled services that can be wrapped. Table 14-1 on page 190 shows the different locations of the inetd configuration files.

Chapter 14. Security and hardening

189

Table 14-1 inetd configuration files Services Configuration locations Solaris /etc/inetd.conf Red Hat /etc/xinetd.d/ /etc/xinetd.conf SUSE /etc/xinetd.d/ /etc/xinetd.conf

To disable an xinetd service, edit the appropriate file in the /etc/xinetd.d/ directory by ensuring that the line disable = yes is present in the file. Example 14-1 provides an example of how to disable the Telnet service.
Example 14-1 Sample xinetd entry for a service # default: on # description: The telnet server serves telnet sessions; it uses \ # unencrypted username/password pairs for authentication. service telnet { flags = REUSE socket_type = stream wait = no user = root server = /usr/sbin/in.telnetd log_on_failure += USERID disable = yes }

To enable a service, change the entry to disable = no.

14.3.2 TCP wrappers

When enabling any inetd services in Linux, TCP wrappers add an extra layer of access control the same way it does with Solaris. In most Linux distributions, TCP wrappers are installed and enabled by default. As in Solaris, TCP wrappers access control is managed through two configuration files: /etc/hosts.allow /etc/hosts.deny This file is used to set the systems and protocols that are allowed to connect to this host. This file is used to set the systems and protocols that are not allowed to connect to this host.

In Solaris, TCP wrappers are enabled by adding the following line to the /etc/default/inetd file:
ENABLE_TCPWRAPPERS=YES

190

Solaris to Linux Migration: A Guide for System Administrators

This is not necessary in Linux because TCP wrappers are typically enabled during installation. In Solaris, each service to be wrapped must be specified as such in the /etc/inetd.conf file. Example 14-2 shows how to wrap the Telnet service in the /etc/inetd.conf file.
Example 14-2 Wrapping the Telnet service in /etc/inetd.conf telnet stream tcp nowait root /usr/sbin/tcpd in.telnetd

Linux TCP wrappers are automatically applied to all xinetd services as long as TCP wrappers are compiled in at installation time. For detailed configuration information, check the following resources. Chapter 15, TCP Wrappers and xinetd, in Red Hat Linux 9: Red Hat Linux Reference Guide, available at:
http://www.redhat.com/docs/manuals/linux/RHL-9-Manual/ref-guide/ch-tcpwrapp ers.html

See also 7.9, TCP wrappers on page 123

14.3.3 FTP
FTP historically is a protocol has had a lot of documented security issues associated with its use, for example, the broadcasting of unencrypted user names and passwords during authentication. Nevertheless, FTP is a very popular and useful internetworking protocol. This section discusses methods for running FTP on Linux in a reasonably secure manner. Implementations of the FTP protocol are generally similar across all platforms. However, there are different versions of the server and several differences between the Solaris and Linux versions of the FTP server. Solaris currently installs wu-ftp as its default FTP server. Red Hat and SUSE both use Very Secure FTP (vsftpd). vsftpd is a GPL-licensed FTP server for UNIX systems and is included in Linux. It is considered secure, extremely fast, and stable. Although it is small in size, vsftpd has a lot of functionality. Some of the more popular features of vsftpd include: Virtual IP configurations Virtual users Ability to chroot users into their home directories

Chapter 14. Security and hardening

191

Run either stand-alone or from inetd Written to reduce the risk of buffer overflow attacks Powerful per-user configurability Reduce the risk of DoS attacks by using bandwidth throttling IPv6 support Support for SSL encryption One of the most important security-related differences between Solaris and Linux FTP is the location of the ftpusers file. This file is used to control access to the FTP server. Solaris: /etc/ftpd/ftpusers Linux: /etc/ftpusers With all of these advanced features, the following references should be helpful in learning and configuring vsftp:
http://vsftpd.beasts.org/vsftpd_conf.html http://vsftpd.beasts.org/ http://www.vsftpdrocks.org/

14.4 Kernel tuning for security

Solaris kernel tuning to modify behavior of TCP for security purposes can be manually implemented using the ndd command. The problem with manual configuration at the command prompt is that the changes will not survive a reboot. Because of this, the ndd commands can be set more permanently by placing them in a boot-time shell script, such as /etc/init.d/inetinit. Similarly, Linux allows for manual kernel tuning at the command prompt using the sysctl command. But for persistent configuration changes, you need to modify the /etc/sysctl.conf file. If changes are made to the /etc/sysctl.conf file, the changes will not become active until the system is rebooted or the command is executed:
# sysctl -wp /etc/sysctl.conf

Here we provide some examples of kernel tuning in both Solaris and Linux. For socket queue defense against SYN attacks: Solaris: /usr/sbin/ndd -set /dev/tcp tcp_conn_req_max_q 1024 Linux: sysctl -w net.ipv4.tcp_max_syn_backlog-1280

192

Solaris to Linux Migration: A Guide for System Administrators

Redirects (IP redirects can be used to modify the routing table on a remote host): Solaris:
/usr/sbin/ndd -set /dev/ip ip_ignore_redirect 1 /usr/sbin/ndd -set /dev/ip ip_send_redirects 0

Linux:
sysctl -w net.ipv4.conf.all.send_redirects=0 sysctl -w net.ipv4.conf.all.accept_redirects=0

As you can see, Linux uses the sysctl command to modify kernel parameters at run time. These kernel parameters are controlled by the files in the /proc/sys/ directory. For more detailed information, check the sysctl man page online at:
http://linuxcommand.org/man_pages/sysctl8.html

Another good resource about this topic is the UNIX IP Stack Tuning Guide at:
http://www.cymru.com/Documents/ip-stack-tuning.html

14.5 Logging
Logging in Linux is similar to Solaris with a few exceptions. For the most part, Solaris logs syslog messages to the /var/adm/messages file unless otherwise specified in the /etc/syslog.conf file. Linux by default logs many of its syslog messages in numerous files in /var/log/ by default. As with Solaris, these are speicified in the /etc/syslog.conf file. Table 14-2 shows many of the default Linux log files found in /var/log/.
Table 14-2 Security logging facilities Solaris Main log file authpriv cron logs Mail messages /var/cron/log /var/adm/messages Red Hat /var/log/messages /var/log/secure /var/log/cron /var/log/maillog /var/log/mail /var/log/mail.info /var/log/mail.warn /var/log/mail.err SUSE /var/log/messages

Boot messages

/var/log/boot.msg

/var/log/boot.log

For more information about logging, see 17.4, Logs on page 220.

Chapter 14. Security and hardening

193

14.6 Access control lists

Access control lists (ACLs) are managed in much the same manner in Linux as they are in Solaris. The commands chacl, getfacl, and setfacl should be familiar to Solaris and Linux administrators alike. Because much of the Linux ACL functionality is the same as Solaris, you should be able to manage the transition with little direction. For further review of the more technical details of Linux ACLs, refer to: POSIX ACLs on Linux:
http://www.suse.de/~agruen/acl/linux-acls/online/

Red Hat ACLs:

http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-guide/ ch-acls.html

14.7 PAM
The default mechanism used for user authentication in Linux uses Pluggable Authentication Modules (PAM). PAM provides centralized authentication for common services such as login, FTP, Telnet, and SSH. PAM is implemented using dynamic load libraries (DLLs). Authentication can be customized for PAM-aware services using PAM configurations files. Figure 14-1 illustrates PAM operations.

194

Solaris to Linux Migration: A Guide for System Administrators

Figure 14-1 PAM operations

Authentication for each PAM-aware application is configured by a configuration file in the /etc/pam.d directory (the file name indicates the service it configures). PAM modules implement user authentication; by default, PAM modules are located in the /lib/security directory. Some PAM modules read global configuration files located in the /etc/security directory. Note: Solaris PAM configuration occurs in the /etc/pam.conf file, and Linux uses the /etc/pam.d directory. A PAM configuration file consists of lines in the form:
module-type control-flag module-path arguments

We describe each field in the following sections.

14.7.1 PAM module types

Table 14-3 on page 196 shows four types of authentication mechanisms (referred to as module types).

Chapter 14. Security and hardening

195

Table 14-3 Authentication mechanisms Type auth account Description Used to authenticate a user based on user name and password Used to grant or deny access based on account policies (such as time of day) password session Used to authorize user password updates Used to apply settings before a session is established

Note: More than one line can be specified for any PAM module type. This is referred to as module stacking. In this case, each module is executed in first-to-last order. Final authentication depends on the PAM control flags specified.

14.7.2 Limiting superuser login to secure terminals

Superuser access should always be restricted to a secure terminal. Network login by root should be denied. This can be enforced using the pam_securetty.so module. If enabled when a superuser logs in, the pam_securetty.so modules deny access unless the login originates from a terminal device listed in the /etc/securetty file. Example 14-3 shows the default secure device defined for SLES8 on a zSeries platform.
Example 14-3 The SLES8 default /etc/securetty file # This file contains the device names of tty lines (one per line, # without leading /dev/) on which root is allowed to login. ttyS0

If superuser authority is required for a network session, log in as a non-privileged user and then issue the su command to switch to superuser. Alternatively, consider using sudo to delegate authority. The pam_securetty.so module is not enabled for SSH in the default SLES8 installation. It might be desirable to deny superuser SSH network access even though the data is encrypted. This can help prevent an attacker from guessing the root password. See Table 14-4 on page 197.

196

Solaris to Linux Migration: A Guide for System Administrators

Table 14-4 Limiting superuser login Solaris Allow/deny root login /etc/default/login Red Hat /etc/securetty SUSE /etc/securetty

14.7.3 Restricting user login

At times, it might be useful to restrict non-superuser logins (for example, when performing system maintenance). The pam_nologin.so module can be used to deny access to the system. When enabled, pam_nologin.so checks for the existence of the /etc/nologin file. If found, pam_login.so denies access and displays the contents of the file to the user. An alternate nologin file can be specified by supplying the file=pathname PAM argument. Note: The nologin mechanism does not affect the superuser login. Superusers can log in even if the /etc/nologin file exists.

14.8 umask
The function and usage of umask is the same between Solaris and Linux. The configuration locations are a little different, as shown in Table 14-5.
Table 14-5 Controlling default umask Solaris umask /etc/profile Red Hat /etc/bashrc SUSE /etc/profile

14.9 SSH
The Secure Shell (also known as SSH) is a secure network access protocol. Traditional network access protocols (Telnet, FTP, and the r-command suite, such as rlogin, rcp, and rsh) must be considered inherently insecure. Unencrypted data (including user IDs and passwords) is transmitted over the network and can be captured by malicious third parties. SSH provides a secure alternative for remote access, file transfer, and remote command execution. SSH was originally developed by Tatu Ylnen (a researcher at the Helsinki University of Technology) as a method to prevent password-sniffing attacks. His initial implementation (known as SSH1) was released as free software. The SSH-1 protocol was formalized as an Internet Engineering Task Force (IETF) Draft. Subsequent improvements resulted in a new SSH-2 protocol originally

Chapter 14. Security and hardening

197

implemented as the SSH2 commercial product available from SSH Communications Security, Ltd. (founded by Ylnen in 1995). OpenSSH is an open source implementation of Secure Shell based on the last free version of SSH1. It supports both the SSH-1 and SSH-2 protocols. OpenSSH is provided as part of most distributions of Linux. For details about OpenSSH, refer to:
http://www.openssh.com

Note: When we use the term SSH in the remainder of this chapter, we are referring specifically to the OpenSSH implementation.

14.10 IPSec and IKE

In this section, we discuss the differences in IPSec/IKE implementation between Red Hat and SUSE.

Red Hat IPSec

The Red Hat Enterprise Linux implementation of IPSec uses IKE for sharing keys between hosts across the Internet. The racoon keying daemon handles the IKE key distribution and exchange. Implementing IPSec requires that the ipsec-tools RPM package be installed on all IPSec hosts (if using a host-to-host configuration) or routers (if using a network-to-network configuration). The RPM package contains essential libraries, daemons, and configuration files to aid in setup of the IPSec connection, including: /lib/libipsec.so A library that contains the PF_KEY trusted key management socket interface between the Linux kernel and the IPSec implementation used in Red Hat Enterprise Linux. setkey manipulates the key management and security attributes of IPSec in the kernel. This executable is controlled by the racoon key management daemon. For more information about setkey, refer to the setkey(8) man page. The IKE key management daemon, used to manage and control security associations and key sharing between IPSec-connected systems. This daemon can be configured by editing the /etc/racoon/racoon.conf file. For more information about racoon, refer to the racoon(8) man page.

/sbin/setkey

/sbin/racoon

198

Solaris to Linux Migration: A Guide for System Administrators

/etc/racoon/racoon.conf

The racoon daemon configuration file used to configure various aspects of the IPSec connection, including authentication methods and encryption algorithms used in the connection. For a complete listing of directives available, refer to the racoon.conf(5) man page.

Configuring IPSec on Red Hat Enterprise Linux can be done through the Network Administration Tool or by manually editing networking and IPSec configuration files. For more information about using the Network Administration Tool, refer to the Red Hat Enterprise Linux System Administration Guide, available at:
http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/security-guide/ ch-vpn.html

SUSE IPSec
SLES uses the Road Warrior server as its VPN tool. Road Warrior server is a VPN server that accepts connections from any client with a valid and signed CA certificate. Perform the following three steps to set up a Road Warrior server: 1. Create a server certificate on the CA management computer. 2. Import a certificate on the server computer. 3. Set up a connection on the server. For details about how to configure both the IPSec server and client on SLES, refer to Chapter 26 of the document available at:
http://www.novell.com/documentation/sles9/pdfdoc/sles_9_admin_guide/sles_9_ admin_guide.pdf

14.11 Kerberos
We discuss the differences in Kerberos implementations between Red Hat and SUSE in this section.

Red Hat Kerberos

To configure a basic Red Hat Kerberos server, follow these steps: 1. Ensure that clock synchronization and DNS are running. 2. Install the krb5-libs, krb5-server, and krb5-workstation packages on the KDC system. 3. Edit the /etc/krb5.conf and /var/kerberos/krb5kdc/kdc.conf configuration files to reflect your realm name and domain-to-realm mappings.

Chapter 14. Security and hardening

199

4. Create the kerberos database using the following command:

/usr/kerberos/sbin/kdb5_util create -s

5. Edit the /var/kerberos/krb5kdc/kadm5.acl file. 6. Start Kerberos using the following commands:
/sbin/service krb5kdc start /sbin/service kadmin start /sbin/service krb524 start

7. Add principals for your users using the addprinc command with kadmin. 8. Verify that the server will issue tickets by running kinit to obtain a ticket and store it in a credential cache file. For more information about how to configure a Kerberos server, see:
http://www.redhat.com/docs/manuals/linux/RHL-9-Manual/ref-guide/s1-kerberos -server.html

The Red Hat kerberos distribution has the ability to run through the configuration of services within xinetd. Example 14-4 shows the contents of a typical xinetd configuration for some kerberized services (from /etc/xinetd.d/).
Example 14-4 xinted service entries krb5-telnet

service telnet { flags = REUSE socket_type = stream wait = no user = root server = /usr/local/krb5/sbin/telnetd server_args = -X KERBEROS_V4 -a valid log_on_failure += USERID disable = no } kshell service kshell { flags socket_type wait user server

= = = = =

REUSE stream no root /usr/local/krb5/sbin/kshd

200

Solaris to Linux Migration: A Guide for System Administrators

server_args = -5 -c -A disable = no } klogin service klogin { flags socket_type wait user server server_args disable } eklogin service eklogin { flags socket_type wait user server server_args disable } gss-ftp service ftp { flags socket_type wait user server server_args disable }

= = = = = = =

REUSE stream no root /usr/local/krb5/sbin/klogind -5 -c no

= = = = = = =

REUSE stream no root /usr/local/krb5/sbin/klogind -5 -c -e no

= = = = = = =

REUSE stream no root /usr/local/krb5/sbin/ftpd -a no

SUSE Kerberos
SUSE uses the Heimdal Kerberos implementation. This functions similarly to other Kerberos implementations. For configuration and administration details of both the Key Distribution Center (KDC) and the Kerberos client, see Chapter 26

Chapter 14. Security and hardening

201

of the SUSE Linux Enterprise Server 9 Administration and Installation Guide, available at:
http://www.novell.com/documentation/sles9/

14.12 Warning banners

Warning banners are used both prior to, and after successful login. Many organizations have specific legal verbiage that must be placed in these warnings, but from a security perspective, it is common practice to remove all references to the operating system name and version from the banners to prevent an attacker from easily identifying a systems potential weaknesses. Table 14-6 shows the paths to the common locations for the control files remaining the same between Solaris and Linux.
Table 14-6 Locations of motd and issue files Solaris Warning at login prompt Warning after successful login /etc/issue /etc/motd Red Hat /etc/issue /etc/motd SUSE /etc/issue /etc/motd

14.13 Firewalls
Both Linux and Solaris include firewall services. Linux uses the open source GPL Netfilter, and Solaris 9 includes SunScreen 3.2, a Sun-branded, non-open source firewall. Both firewalls offer standard filtering such as blocking traffic by port or protocol, stateful inspection, and logging, and both offer NAT and IP masquerading capabilities. Note: Both firewalls have the ability to run as a host-based firewall to protect only the host on which it is installed. Consider that there are two general types of firewalls: Those that protect networks and hosts on networks Those that protect individual hosts

202

Solaris to Linux Migration: A Guide for System Administrators

The Linux firewall can be configured to act as both a network-based firewall with stateful inspection and packet filtering, and as an individual firewall protecting only the host on which it is installed. The Linux firewall is made up of three basic components: Tables: Tables can be thought of as a set of firewall rules. Chains: Chains are groups of rules found within a table. The default chains are FORWARD, INPUT, and OUTPUT. Policies: Policies are the default actions taken on a packet when it does not match any of the chains. The policies for packets are DROP, REJECT, and ACCEPT. The primary command facility used to add and remove rules to or from a Linux firewall is iptables. Table 14-7 shows some of the more useful iptables command options. For a more exhaustive list, see the iptables man page.
Table 14-7 iptables command options Command iptables -A iptables -D iptables -I (capital i) iptables -R iptables -L iptables -F iptables -N iptables -h Description Append a rule to the end of the selected chain Delete one or more rules from the selected chain Insert a rule into the selected chain as a specific rule number Replace a rule in the selected chain List all of the rules in the selected chain Flush the selected chain (delete all rules) Create a new user-defined chain Help

For more detailed information about how to configure firewall services for Linux, see the following links: Red Hat:
http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/security-guide/ http://www.redhat.com/docs/manuals/linux/RHL-9-Manual/install-guide/s1-fire wallconfig.html

Chapter 14. Security and hardening

203

Novell SUSE:
http://www.novell.com/documentation/sles9/ http://www.novell.com/documentation/suse91/suselinux-adminguide/html/ch19.h tml

204

Solaris to Linux Migration: A Guide for System Administrators

15

Chapter 15.

Linux high availability overview

This chapter describes high availability features for Linux to better enable migration from a Solaris high availability system, such as SunCluster. We discuss the services available and considerations for the migration environment and provide pointers to more information. This chapter introduces and discusses the following topics: Introduction to Linux-HA Linux-HA services Linux migration Cluster environment considerations Significant portions of this chapter are from the High Availability Linux project Web site, available at:
http://linux-ha.org

Copyright IBM Corp. 2006. All rights reserved.

205

15.1 Introduction to Linux-HA

Linux high availability (Linux-HA) provides basic high-availability (fail-over) capabilities for Linux. It provides monitoring of cluster nodes and applications, and provides a sophisticated dependency model with a rule-based resource placement scheme. When faults occur, or a rules change occurs, the user-supplied rules are then followed to provide the desired resource placement in the cluster.1 For much more detailed information about Linux-HA, including configuration details, tutorials, and technical papers, refer to the High Availability Linux project at:
http://linux-ha.org

15.1.1 Migration scenarios

Linux-HA can be deployed in several different migration environments: Single node HA system: Provides stand-alone application monitoring and restart Multiple virtual machines (single physical machine): Adds OS crash protection and also provides for rolling upgrades of OS and applications Multiple physical machines: Adds protection against hardware failures Split-site (stretch) clusters: Adds protection against site failures

15.1.2 Some features of Linux-HA

Some of the features and functionality offered by Linux-HA include1: Works on all known Linux variants and platforms, supplied with many distributions including Novell SUSE starting with SLES9 (not yet included in Red Hat). Sub-second server failure detection. Sophisticated knowledge of service interdependencies, always starts and stops resources quickly and correctly. Policy-based resource placement puts services exactly where you want them based on dependencies, user-defined rules, and node attributes. Fine-grained resource management can be based on user-defined criteria. Time-based rules allow for different policies depending on time. Easy-to-understand and deploy init-based application management; most services require no application scripts for basic management.
1

From http://linux-ha.org

206

Solaris to Linux Migration: A Guide for System Administrators

Resource groups provide simple, easy-to-use service management for straightforward environments. Active fencing mechanism (STONITH) provides strong data integrity guarantees, even in unusual failure cases. Provides stronger integrity guarantees than simple quorum tie-breakers. Open source implementation avoids vendor lock-in and provides great flexibility, stability, responsiveness, and testing. Works on stand-alone machines to provide easy-to-use, init-script-based service monitoring and restart.

15.2 Linux-HA services

In this section, we describe the different system services that work together to provide the Linux-HA functions.

15.2.1 Heartbeat
The Heartbeat program is one of the core components of Linux-HA. It performs death-of-node detection, communications, and cluster management in one process. Heartbeat monitors node death and includes built-in resource monitoring for all types of resources. Resources are automatically restarted on failure.1 Heartbeat also provides strongly authenticated, locally ordered, multicast messaging.

15.2.2 Cluster Resource Manager (CRM)

The Cluster Resource Manager (CRM) is the brain of Linux-HA. It maintains the resource configuration, decides what resources should run where, decides how to move from the current state into the desired state. In addition, it supervises the Local Resource Manager (LRM) in accomplishing all these things.1 CRM functionality: Uses heartbeat for communication Receives membership updates from the Consensus Cluster Membership Directs the work of and receives notifications from the Local Resource Manager Tells the STONITH daemon when and what to reset Logs using the logging daemon

Chapter 15. Linux high availability overview

207

15.2.3 Consensus Cluster Membership (CCM)

Consensus Cluster Membership (CCM) provides strongly connected consensus cluster membership services. Ensures that every node in a computed membership can talk to every other node in this same membership.

15.2.4 Local Resource Manager (LRM)

The Local Resource Manager is basically a resource agent abstraction. It starts, stops, and monitors resources as directed by the CRM. It has a plug-in architecture and can support many classes of resource agents. The types currently supported include:1 Open Cluster Framework (OCF). For more information, refer to:
http://opencf.org

Heartbeat (release 1) style resource agents. Linux Standards Base (LSB) (normal init script) resource agents. STONITH. Resource agent interface for instantiating STONITH objects to be used by the STONITH daemon.

15.2.5 STONITH daemon

The STONITH daemon provides cluster-wide node reset facilities using the improved release 2 STONITH API. The STONITH library includes support for about a dozen types of C STONITH plug-ins and native support for script-based plug-ins, allowing scripts in any scripting language to be treated as full-fledged STONITH plug-ins. The STONITH daemon locks itself into memory, and the C based plug-ins are preloaded by the stonith daemon so that no disk I/O is required for a STONITH operation using C methods. The STONITH daemon provides full support for asymmetric STONITH configurations, allowing for the possibility that certain STONITH devices can be accessible from only a subset of the nodes in the cluster.1

15.3 Linux migration

There are several sources of additional information to study after deciding to migrate from the Solaris environment to one using Linux-HA. There is a tutorial explaining the basics of how to set up the various components of Linux-HA available on the Linux-HA Web site:
http://linux-ha.org/GettingStarted

208

Solaris to Linux Migration: A Guide for System Administrators

For more information about how to get started configuring Heartbeat, refer to the documentation available at the Linux-HA Web site:
http://linux-ha.org/GettingStartedWithHeartbeat

This page is particularly good because it shows various example configurations for many popular cluster environments.

Kernel
Linux-HA runs well on any distribution using Linux kernel version 2.6 or later. This includes all current Red Hat and Novell SUSE kernels. There are also no kernel dependencies or hooks.

15.4 Cluster environment considerations

Linux-HA clustering is designed to recover from single-system faults. For more complete and up-to-date information about how to implement HA in a clustered environment, refer to the sources mentioned in 15.3, Linux migration on page 208.

15.4.1 Data sharing arrangements

Linux-HA has no special shared disk requirements. It supports (at least) the following data sharing configurations:1 No shared data Replication (DRBD, or application-specific) SCSI RAID controllers supporting clustering (IBM ServeRAID, ICP Vortex) External RAID units: SCSI, Fibre Channel, any kind The only requirements on shared disk is that it support mount and umount. More specifically, it does not rely on SCSI reservations (or their equivalent).

15.4.2 IBM ServeRAID

Certain of the IBM ServeRAID SCSI RAID controllers provide support for access to a given RAID volume by either one of two machines at a time. In this environment, two ServeRAID controllers can be connected to the same SCSI array, and then either (but not both) of the two machines can access the data. For more specifics about the supported IBM ServeRAID controllers and their configuration, refer to the Linux-HA Web site:
http://linux-ha.org/ServeRAID

Chapter 15. Linux high availability overview

209

210

Solaris to Linux Migration: A Guide for System Administrators

16

Chapter 16.

Shell scripting
This chapter discusses the common shells that system administrators use for their day-to-day work. This chapter includes the following topics: Overview of the shell environment Public Domain Korn shell Moving from ksh to bash

Copyright IBM Corp. 2006. All rights reserved.

211

16.1 Overview of the shell environment

Shells are standard command interpreters for the UNIX system. Commands are entered through the terminal or from a file. This is not a complete description of what a shell is. We discuss only the most commonly used shells in this section.

16.1.1 Solaris shell environments

Solaris 8, when installed, comes with the following shells: sh ksh csh rsh rksh Bourne shell Korn shell C shell Restricted Bourne shell Restricted Korn shell

There are other shells that can be installed from the Companion CD. Solaris 9, in addition to the previous shells, also comes with: bash tcsh zsh Bourne-Again shell Tenex C shell Another shell option that includes a combination of features found in bash, ksh, and tsch

The default shell provided when a new user is created, and the shell parameter is not used, is sh (Bourne shell).

16.1.2 Linux shell environments

The Bourne-Again Shell (bash) is the default shell for Linux. The bash shell consists of the features of sh and ksh, plus a bit more. For more information about bash, see the man pages. Linux also comes with other shells: csh, tcsh, zsh, and so on. Most of these shells are dynamic executables that rely on shared library files. In a disaster scenario when these libraries are not available, the shell cannot be started. In Linux, a file called /bin/ash.static is a statically linked binary that does not rely on any external shared libraries. This shell can be used as the shell when the system is in a recovery mode and when the normal shells do not work as intended.

212

Solaris to Linux Migration: A Guide for System Administrators

For a description about how to use this file, refer to Chapter 17, Troubleshooting on page 215.

16.2 Public Domain Korn shell

The Linux equivalent of the Korn shell (ksh) is the Public Domain Korn shell (pdksh). Although it is intended to be a POSIX-compliant Korn shell, it is not fully conforming at this time. The bash shell has most of the features and functionality of ksh, and is the recommended migration path in the Linux environment.

16.3 Moving from ksh to bash

Migrating your sh and ksh scripts to bash is typically a straightforward process. However, there is a minor difference to be aware of. If you are using the read statement in your script, as shown in Example 16-1, the second line echo results in a blank line. This is a known issue in bash and applies to both single and multiple variable reads.
Example 16-1 Sample read statement > echo A B C | read x y z > echo $x $y $z > _

Here are some suggested workarounds for the read command: Work around 1: Using echo, as shown in Example 16-2
Example 16-2 Using a redirected echo to a read > > A > read x y z < <(echo A B C) echo $x $y $z B C _

Chapter 16. Shell scripting

213

Work around 2: Using awk, as shown in Example 16-3

Example 16-3 Using awk > x=echo A B C | awk { print $2 } > echo $x B > _

For more information about using bash, refer to the following Web site:
http://www.gnu.org/software/bash/

214

Solaris to Linux Migration: A Guide for System Administrators

17

Chapter 17.

Troubleshooting
This chapter describes some differences in the troubleshooting process of Linux and Solaris. We discuss the following topics: Troubleshooting the booting process Core files Crash dumps Logs Permissions: File access problems Printing File systems Packages root password recovery Network System and user processes Diagnostic and debugging tools

Copyright IBM Corp. 2006. All rights reserved.

215

17.1 Troubleshooting the booting process

Sometimes a system will not boot properly. It might hang at a point in the boot process or reboot continuously.

Solaris boot troubleshooting

In Solaris, you can perform interactive boots, device reconfiguration boots, single user boots, and verbose boots in order to correct a faulty system that does not boot anymore. If the system is severely damaged, and the file systems are corrupted, you can boot from an external device such as CD/DVD or the network in order to repair the system.

Linux boot troubleshooting

In Linux, you can also perform single user boots or verbose boots for maintenance purposes using the already installed kernels, or if the local kernels or initrd images are corrupted or file systems are broken, you can boot from a CD, DVD, USB flash, or network in rescue mode in order to repair the system. To boot in single user mode, you need to add 1 at the end of kernel parameters at GRUB prompt, and the init will use that level instead of the default found in the /etc/inittab configuration file. To boot in verbose mode, you have to remove rhgb quiet from the kernel parameters at the GRUB prompt in Red Hat, or just press Esc when booting in Novell SUSE. If your /etc/inittab file is corrupted and you cannot boot in any run level, you can use the emergency kernel parameter that will tell init to ignore the /etc/inittab file and launch the default shell after prompting for the root password. If there is a major problem with your installed shared libraries, you might find that the init program or default shell are not working because they are dynamically linked to the respective corrupted libraries. In this case, you can boot the system in a very minimal single user mode by using init=/bin/ash.static as the kernel parameter, and it will run this static compiled shell instead of the init program. Then, you can use several other static binaries to correct your system. For more information about possible kernel parameters, refer to the man page for bootparam, and check your vendor documentation. For rescue booting, you just need to boot from an installation CD, DVD, USB flash, or network and follow the on-screen instructions. Some vendors also provide a rescue mode CD to be used only for rescue.

17.2 Core files

Core files are created when an application or process terminates abnormally. These core files are used by the developers to determine the cause of abnormal

216

Solaris to Linux Migration: A Guide for System Administrators

termination and to troubleshoot the respective program that generated the core file.

Solaris management of core files

In Solaris, core files are managed with the coreadm command. You can have a per-process core file path, which is enabled by default, and it creates a core file in the current directory. In addition, you can create a global core file path, which is disabled by default, and it creates another core file in the global directory you specify. The per-process core is created with read/write rights only for the owner of the process; the global core is created with read/write rights for the root only. You can also expand the default core file name to include useful information such as the PID, EUID, and EGID. Setuid programs do not produce a core by default, but this behavior can be changed with the coreadm command. Then, you can use the pstack, pmap, pldd, pflags, and pcred tools to show information about the core file. You can inspect core files with the mdb utility.

Linux management of core files

In Linux, we do not have a per-process core files and a global core file as in Solaris, but a single core file and its generation is disabled by default. To enable core file generation, you need to use the ulimit command to change the maximum accepted core file size from 0 to something more reasonable.This setting affects only the programs started from that shell. To set this option permanently, you can modify an initialization script such as /etc/profile or .bash_profile. To customize the core file name and path, use the sysctl -w command to dynamically change the value of the kernel.core_pattern attribute, or if you need a permanent setup, you can add your settings to the /etc/sysctl.conf file and run sysctl -p. For more information about setting core file names and path, refer to the man pages for proc, and search for core_pattern. There is also a kernel.core_uses_pid attribute that is obsolete and is maintained only for compatibility reasons. If you need PID in your core file names, use kernel.core_pattern instead with the %p pattern. You should set kernel.core_uses_pid to 0 using the same methods described earlier. Example 17-1 shows how to enable core file generation and modification of the default core pattern to store the core files in /core/core.program_name.pid.
Example 17-1 Example of core file setup ulimit -c 1000000 sysctl -w kernel.core_uses_pid=0 sysctl -w kernel.core_pattern=/core/core.%e.%p

Chapter 17. Troubleshooting

217

Table 17-1shows the differences between the Solaris and Linux core file patterns.
Table 17-1 Core file patterns Core pattern definition Process ID Effective user ID Effective group ID Executable file name System node name (uname -n) Machine name (uname -m) Time stamp Number of signal causing dump Literal % Solaris %p %u %g %f %n %m %t N/A %% Linux %p %u %u %e %h N/A %t %s %%

17.3 Crash dumps

When a system crashes, it can save an image of the physical memory at the time of crash on the default dump device for further analysis. System crashes can be due to from software problems or hardware problems.

Solaris management of crash dump files

In Solaris, when a system crashes, it writes a copy of the physical memory to the dump device. Then, after reboot, the savecore command is executed to save the crash dump files from the dump device to the configured savecore directory in the form of unix.X and vmcore.X files, where X is the dump sequence number. The crash dump files are used by Sun support to diagnose the problem. The default directory for saving crash dump files is /var/crash/hostname. You can inspect crash dump files with the mdb utility. The default dump device is the swap space of the system, but it can be changed to a separate partition. You can generate a crash dump of a running system with the savecore -L command if you configured a separate dump partition. Managing the crash dump configuration and content can be modified with the dumpadm command.

Linux management of crash dump files

The management of crash dump files is handled differently in Red Hat and SUSE.

218

Solaris to Linux Migration: A Guide for System Administrators

Crash dump files in Red Hat

Red Hat has two options for managing crash dump files: diskdump for disk-based dump files and netdump for network-based dump files. The diskdump is a standard service in the /etc/init.d/diskdump directory and has the /etc/sysconfig/diskdump configuration file. This service is disabled by default, and you need to manually configure and activate it. When a crash occurs, a dump is stored in the partition specified in the /etc/sysconfig/diskdump configuration file. Then, use the savecore utility to read the data from the dump partition. This creates a vmcore file in a directory named /var/crash/127.0.0.1-timestamp. The vmcore crash dump file can be analyzed with the crash utility that is provided in a separate package. For network-based dump, Red Hat uses the netdump package on the client machine that will generate the dump, and the netdump-server package on a reliable server that will receive and host the dump files. The netdump is a standard service in /etc/init.d/netdump and has the /etc/sysconfig/netdump configuration file. This service is disabled by default, and you need to manually configure and activate it. In order to use this service, you need to configure the /etc/netdump.conf file and start the netdump-server service on a reliable server in your network that will receive the oops data and memory dumps from the client and store them in the default /var/crash directory. You can instruct netdump-server to execute certain scripts upon receiving a dump data. The dump data can be inspected with the crash utility that is provided in a separate package. For more information, check the diskdumpctl, diskdumpfmt, savecore, crash, netdump, and netdump-server man pages and the vendor documentation.

Crash dump files in Novell SUSE

SUSE manages crash dump files using the Linux Kernel Crash Dump (KKCD) utility that provides the ability to collect, save, and analyze information about the kernel when the system crashes. After LKCD is configured in the /etc/sysconfig/dump file, a crash dump is created whenever a kernel panic or kernel oops occurs, or on demand by using a specific key combination. The default settings for the dump is to pick up the dump headers and kernel pages only. In addition, the default dump device is the network. When a kernel panic or oops occurs, the following steps are performed: 1. Saves system memory to the dump device. 2. Reboots the system. 3. Runs the LKCD configuration, which prepares the system for the next crash.

Chapter 17. Troubleshooting

219

4. Creates dump files on the permanent dump storage area from the dump device. The resulting dump can be analyzed with the lcrash tool that is part of this package. And last but not least, LCKD provides for both local and network storage of dumps. For more information about the Linux Kernel Crash Dump utility, refer to the following Web site and the and SUSE documentation:
http://lkcd.sourceforge.net

Table 17-2 shows the main differences between Solaris and Linux crash dump management.
Table 17-2 Crash dump management Topics Disk crash dump utilities Disk crash config files Network crash dump utilities Network crash config files Crash analyzing tools Crash dump default store directory Solaris dumpadm savecore /etc/dumpadm.conf N/A N/A mdb /var/crash RHEL /etc/init.d/diskdump savecore /etc/sysconfig/diskdump /etc/init.d/netdump /etc/init.d/netdump-server /etc/sysconfig/netdump /etc/netdump.conf crash /var/crash SLES lkcd /etc/sysconfig/dump lkcd /etc/sysconfig/dump lcrash /var/log/dump

17.4 Logs
System logs are very useful for troubleshooting. The logs are the first thing to check if something is going wrong. Even if the system is stable and running just fine, it is good to inspect the system logs periodically. There might be an indication of symptoms that will potentially affect the availability of the system.

Logs in Solaris
Solaris uses the syslogd daemon to record various system warnings and errors. The syslogd daemon is configured by default to send the most critical messages to the console in addition to logging them to files. Most system messages are

220

Solaris to Linux Migration: A Guide for System Administrators

stored by default in the /var/adm/ directory, and the file to check is /var/adm/messages. The system logs are rotated by the logadm (in Solaris 9) and newsyslog (in Solaris 8) command, which has an entry in the root crontab file. The syslogd configuration file is /etc/syslog.conf and it can contain macros and logic blocks. It also has a /etc/default/syslogd configuration file for changing the daemon behavior globally. To see the system diagnostic messages from the boot time, use the dmesg command.

Logs in Linux
Linux also uses the syslogd daemon to record various system warnings and errors. The syslogd daemon is configured by default to send the most critical messages to the console in addition to logging them to files. Most system messages are stored by default in the /var/log/ directory, and the file to check is /var/log/messages. The system logs are rotated by the logrotate program, which has an entry in the /etc/cron.daily/ directory and the configuration is stored in the /etc/logrotate.conf file. The syslogd configuration file is /etc/syslog.conf and it has a syntax similar to Solaris with the exception that it cannot contain macros and logic blocks. It also has a /etc/sysconfig/syslog configuration file for changing the daemon behavior globally. In Linux, there is also a klogd kernel log daemon, which has no equivalent is Solaris. It intercepts and logs Linux kernel messages from /proc/kmsg and logs it directly or through the syslogd daemon. For more information, refer to the man page of klogd. To see the system diagnostic messages from the boot time, use the dmesg command or look in /var/log/boot.log (in Red Hat) or /var/log/boot.msg (in SUSE). The default behavior for SLES is to split the messages for all facilities into files based on the priority of the messages. This is set in the /etc/syslog.conf file and described the syslog.conf man page. So, unless auth and cron have specific entries in the /etc/syslog.conf file, their messages will be directed to files based on the priority of the messages.

Chapter 17. Troubleshooting

221

17.5 Permissions: File access problems

In this section, we discuss problems accessing files and file systems. There are many reasons why a file cannot be accessed. File access problems can be as simple as a bad PATH environment variable or as complex as a non-existent UID. File systems might not be mounted or might be mounted read-only.

17.5.1 Problem: Command not found

The system-wide PATH environment variable is set in /etc/profile. Each user ID can use the system-wide PATH as it is defined or can modify it in their login profile script. Login profile script names are different for each shell. A users login shell is defined in /etc/passwd. Refer to the man page for the shell to determine the login profile name. Table 17-3 shows file location commands. You can display the current PATH with this command:
echo $PATH

The locate command is different from the find command. First, the locate command can be used to find any file or directory, not just executable files. Second, locate is much faster than the find command because it uses a file name database. The updatedb command is used to initialize and maintain the file name database. You can run the updatedb command at any time to update the file name database. If you want to schedule the updatedb command to run one or more times daily, pick times when there is less activity on the server. The updatedb command can run for a long time as it searches the file systems on the server. Refer to the man page for updatedb for configuration options.
Table 17-3 File location commands Task Display $PATH settings Search $PATH for first occurrence of executable filename Search file system or directory tree for filename Update the file name database Search the file name database for filename Display mounted file systems and mount options Solaris echo $PATH which Linux echo $PATH which

find N/A N/A mount

find updatedb locate mount

222

Solaris to Linux Migration: A Guide for System Administrators

17.5.2 Problem: File access

File permissions are based on owner, group, and guest access permissions. These are defined in a files access control list (ACL). The simplest way to display file permissions is with the ls -l command. The chown, chgrp, and chmod are the same for Solaris and Linux. Refer to the corresponding man pages for each command for more information. The stat command can give you more detailed information about a file. The stat command reports such items as the access permissions, the last time the file was accessed or modified, and the size of the file. There might be times when a file has been opened by another application and you can only open it in read-only mode. Linux provides the lsof and fuser commands to list the open files and the PID associated with the file, as shown in Table 17-4.
Table 17-4 File access commands Task Display file name, permissions, and owner Display detailed information for a file or file system Display open files Display processes using files or sockets Solaris ls -l N/A proc tools fuser Linux ls -l stat lsof fuser

17.6 Printing
This section covers basic troubleshooting of print subsystems. For more detailed information about managing print queues and print jobs, refer to your Linux distributions administration guides. Table 17-5 lists basic troubleshooting commands.
Table 17-5 Print service troubleshooting commands Print service troubleshooting task Display print queue status Display settings of a print queue Display print service status information Solaris lpstat lpc lpadmin lpstat Linux lpc lpq lpadmin lpoptions lpstat

Chapter 17. Troubleshooting

223

Print service troubleshooting task Display printing config information from the system config database Display available devices or drivers

Solaris lpget N/A

Linux N/A lpinfo

17.6.1 Troubleshooting remote printer connectivity

The technique used to debug network connectivity to a remote printer is to try to make a connection to the printer port. There are several commands that can be used to verify network connections. We demonstrate the telnet and netcat commands in this section. The default port for the lpd service is 515. The default port for the cupsd service is 631. Use either of the following commands to test network connections to remote printers:
telnet hostname portnumber netcat -v -z hostname portnumber

Both commands will return connection refused if no services are registered with the supplied port number. Any other output will mean that a service is registered with the supplied port number. After you have determined the problem is not with the network, troubleshoot the printer locally at the server.

17.6.2 Troubleshooting local printers

This section covers troubleshooting locally attached printers. The basic procedure to troubleshoot a print service is to review the error log and fix any problems noted in the log. Perform the following steps: 1. Stop the print service. 2. Back up, rename, or delete the print services error logs. 3. For CUPS, set LogLevel to debug in the configuration file /etc/cups/cupsd.conf. 4. Start the print service. 5. Retry the action or operation that causes the problem. 6. Correct any problems noted in the error log.

224

Solaris to Linux Migration: A Guide for System Administrators

LPD
Access permissions are located in /etc/lpd.perms.

CUPS
Access permissions and other configuration parameters are in /etc/cups/cupsd.conf. Log files are in the /var/log/cups directory. CUPS provides online manuals at the following location. The Software Administrators Manual contains a troubleshooting section along with installation and configuration sections.
http://localhost:6(2)31/documentation.html

17.7 File systems

If your server crashes, or has somehow powered off without actually flushing the buffer and unmounting the file systems, you will have some dirty file systems and might end up with inconsistencies in those file systems. That is why on each boot, the initialization scripts run fsck on selected file systems from /etc/vfstab in Solaris or /etc/fstab in Linux. If they were not unmounted properly, the fsck program will scan for inconsistencies. Usually, the fsck program runs in a non-interactive mode, automatically fixing minor inconsistencies. However, if there are some major problems, it switches to manual mode and lets the operator make the decisions. If you are in a situation where you need to manually fix the file system with fsck, it is a good idea to back up the partition before any attempt to fix the file system, especially if you have valuable data on it. In both operating systems, you can run fsck -y to automatically fix any type of problem, be it minor or major. In Solaris 9 and all current Linux distributions, the default file system is journalized unless the administrator chooses something else. This means that even if the system crashes, or is abruptly powered off, the mounted file systems will not be dirty because of this transactional log that is the journal. In this case, it is still a good practice to do a forced fsck -f at scheduled intervals, just to be sure that everything is fine. This is because the journalization will not prevent a kernel or other software bug from causing a file system inconsistency. One other problem with disks is that they can develop bad blocks. If this is the case, you can scan for bad blocks in Linux by using the badblocks command, and then you can instruct fsck to update the bad block list of the existing file system.

Chapter 17. Troubleshooting

225

Important: In Linux, it is a good practice to use the specific fsck.filesystemtype program instead of the generic fsck, or if you use the generic fsck, always use the -t fstype switch.

17.7.1 Remote file systems

You might notice that a remote file system has not mounted, or that it has not mounted properly. Consider the following troubleshooting methods. At the server: Check /etc/exports for access permission. Verify that the rpcbind/portmap service is running. NFSv2 or NFSv3 require the rpcbind service in Solaris and the portmap service in Linux. Verify that the nfsd or nfsserver process is running. Display registered portmap processes. Use the showmount command to display exported subdirectories. At the client: Check /etc/fstab for default mount options. Check /etc/mtab for current mount options. Ping the target server. Table 17-6 shows some of the commands used in this process.
Table 17-6 Remote file system troubleshooting commands Task Display mounted file systems Display detailed information about mounted file systems Display NFS servers export list Display rpcbind/portmap server PID Display registered rpcbind/portmap processes Solaris df mount showmount -e hostname pgrep rpcbind rpcinfo -p hostname Linux df mount showmount -e hostname pgrep portmap rpcinfo -p hostname

226

Solaris to Linux Migration: A Guide for System Administrators

17.7.2 Software RAID

In Linux, the status of software RAID partitions is maintained in /proc/mdstat. If you find that a software RAID array is damaged: 1. Partition a new drive identically to the damaged drive. 2. Shut down your Linux system and power off if necessary. 3. Replace the defective drive with the new drive. 4. Boot the server. 5. Issue the raidhotadd command.

17.7.3 Logical volumes

LVM2 is the default Linux logical volume management utility. Basic operations on volume groups (vg*) and logical volumes (lv*) are vgcreate/lvcreate, vgextend/lvextend, and vgreduce/lvreduce. Table 17-7 lists troubleshooting tasks and the associated command.
Table 17-7 Logical volume troubleshooting tasks Logical volume troubleshooting task Check volume group metadata Scan disk drives and rebuild caches Activate/deactivate volume groups Display volume group status Scan disk drives for logical volumes Activate/deactivate logical volumes Display logical volume status Solaris metadb -i metastat N/A N/A metastat N/A N/A metastat Linux vgck vgscan vgchange vgdisplay lvscan lvchange lvdisplay

17.8 Packages
This section covers troubleshooting software packages. We describe the following tasks in this section: removing packages, upgrading packages, and verifying the installation of packages. Table 17-8 on page 228 illustrates commands used to accomplish these tasks.

Chapter 17. Troubleshooting

227

Table 17-8 Package troubleshooting tasks Package management tasks Remove software package Upgrade/install a package Verify correct installation Verify digest and digital signature Solaris pkgrm pkgadd pkgchk N/A Linux rpm -e rpm -U rpm -F rpm -V rpm -K

Solaris package troubleshooting

The Solaris package format is based on the System V Interface Definition for Application Binary Interface (ABI) and has tools for managing these packages. These management tools include the ability to add and remove packages, check for consistency, and display package information. If system administrators use the pkgadd and pkgrm tools to manage packages, the tools update the software product database accordingly. You can use pkgchk command to perform a consistency check on an installed package and pkgparam to see the parameters of an installed package. Other useful information is found in the /var/sadm/install/contents file and the /var/sadm/pkg/ directory.

Linux package troubleshooting

Linux provides the rpm command for all aspects of package management.

Testing and verifying installed packages

The -K parameter is used to verify digests and digital signatures. The -V parameter is used to test or verify an installed package. Up to eight tests are run against the package. The results for the tests, an attribute marker, and the file name from the package are all returned by the command. For example, if you check the cron package, you might see output similar to that shown in Example 17-2.
Example 17-2 rpm command -V output myhost:~ #rpm -V cron-3.0.1-920-1 .M...... c /etc/crontab .M...... /usr/bin/crontab

In this case, the file mode test failed because the file mode is different on the file system when compared to the file mode in the package. This particular test compares permissions and the file type. All other tests passed because they returned a single . . The letter c is the attribute marker. It tells us that /etc/crontab is a configuration file.

228

Solaris to Linux Migration: A Guide for System Administrators

Refer to the man page for the rpm command for a detailed list of options and parameters.

Upgrading installed packages

The -U and -F parameters are used to upgrade packages. The -F parameter upgrades a package only if it is already installed. The -U parameter installs the package if it is not installed; otherwise, it upgrades the installed version and then removes all other versions after the upgrade completes. SUSE provides a debug option with the command yast2 online_update. For the full syntax of the command, see SUSE Linux Enterprise Server Installation And Administration.

Removing installed packages

The -e parameter is used to erase or remove installed packages. The --nodeps parameter is useful if you do not want to check for dependencies when removing the package.

Database consistency
You might encounter a situation where the RPM database becomes unusable or corrupt. It can be difficult, if not impossible, to upgrade packages if their corresponding entries in the database cannot be accessed. Two parameters are available for RPM database management: --initdb is used to create a new RPM database. --rebuilddb is used to rebuild RPM database indexes from existing, installed RPM package headers.

17.9 root password recovery

There are cases when the root password is lost. This section explains how to change the root password in such cases.

Solaris root password recovery

In Solaris, perform the following steps to change a lost root password: 1. Stop the system, or do a power off/on cycle. 2. Boot in single user mode from a CD-ROM/DVD or from a network boot/install server. 3. Mount the root (/) file system in read/write mode. 4. Remove the root password from the /etc/shadow file and save the file. 5. Unmount the root file system.

Chapter 17. Troubleshooting

229

6. Reboot the system. 7. The root password now is not set. Press Enter at the password prompt when you login as root.

Linux root password recovery

In Linux, you have two options for recovering a lost root password. The first option is to add init=/bin/bash as kernel parameters at GRUB prompt, and the kernel will launch the bash shell instead of the init process. This works only if you can modify the boot parameters, that is, GRUB has no password, or has a known password for configuring boot parameters. The second option is to boot in rescue mode from a CD, DVD, USB flash media, or from the network. To boot in rescue mode, follow the instructions presented on screen when you boot from CD/DVD media. This works even if local GRUB is password protected and you do not know the GRUB password. You can also change the GRUB password if needed by editing its configuration file. The rest of steps are the same in both password recovery options: 1. Mount the root (/) file system in read/write mode if not already mounted by the rescue program. 2. chroot in the new mounted root (/) file system. 3. Use the passwd command to change the root password. 4. exit from the chroot shell. 5. Unmount the root file system. 6. Reboot the system. Tip: In Red Hat, it is enough to boot in single user mode, change to root password with the passwd command, and reboot the server.

17.10 Network
Most of the networking tools used for network troubleshooting are the same in both Solaris and Linux, with some minor differences as illustrated in Table 17-9 on page 231.

230

Solaris to Linux Migration: A Guide for System Administrators

Table 17-9 Troubleshooting network problems Task Display interface settings Display interface status and statistics Configure interface Check various network statistics Check DNS resolver Check name services config Display kernel network parameters Configure kernel network promontories Check for network link Query DNS Check routing table Check ARP entries Testing for connectivity Check IP path Capturing network packets Solaris ifconfig netstat -i ifconfig netstat more /etc/resolv.conf more /etc/nsswitch.conf ndd /dev/ip (\?)parameter ndd /dev/tcp (\?)parameter ndd -set driver parameter ndd driver link_status nslookup dig netstat -r arp -a ping traceroute snoop Linux ip ifconfig netstat -i ifconfig ip ifconfig netstat more /etc/resolv.conf more /etc/nsswitch.conf sysctl -a | grep ^net sysctl -w var=value ethtool mii-tool nslookup dig netstat -r route arp ping traceroute tcpdump tethereal ethereal

Note: In Linux, a network interface is considered up by the kernel even if the network link is not present, so this is the first thing to check if you have network problems.

17.11 System and user processes

Both Solaris and Linux provide the ps and top commands (top is optionally installable on Solaris from the Companion CD), used to display the state of

Chapter 17. Troubleshooting

231

running processes. The ps command gives you a point in time snapshot of the system processes. The top command gives you an iterative, interactive display of system processes. Whether you use the ps or top command, important columns in the output of these commands to monitor are %CPU, STAT, START or STIME, and TIME. Table 17-10 provides an overview of the command sets used to troubleshoot system and user processes.
Table 17-10 Troubleshooting system and user processes Task Trace system calls and signals Library trace Print shared library dependencies Report IPC status Display and manage system resources available to a user or process List open files and sockets and their owning user IDs and PIDs Identify which user or process is using a file or socket Send signals to processes List current processes List current processes based on name and other attributes List current processes in an interactive, iterative format Report system activity Display virtual memory statistics Display I/O statistics Display system error log Solaris truss N/A ldd ipcs ulimit proc tools: pfiles, pmap, ptree, and so on fuser kill pkill ps pgrep top (on companion CD) sar vmstat iostat dmesg Linux strace ltrace ldd ipcs ulimit lsof fuser kill pkill ps pgrep top sar vmstat iostat dmesg

17.12 Diagnostic and debugging tools

Certain tools are installed by default. Review these packages and install them if your installation requires the tools provided by the packages. The base package

232

Solaris to Linux Migration: A Guide for System Administrators

name for some packages will be the same on Red Hat Enterprise Linux and Novell SUSE Linux Enterprise Server. The version might be different. The format of the following lists is: base package name* : command For Red Hat Enterprise Linux: sysreport* : sysreport hwbrowser* : hwbrowser (graphical tool) pciutils* : lspci oprofile* : opreport gdb* : gdb strace* : strace statserial* : statserial For Novell SUSE Linux Enterprise Server: siga* : siga pciutils* : lspci gdb* : gdb strace* : strace ksymoops* : ksymoops statserial* : statserial Refer to the man pages and RPM package descriptions for information regarding these commands.

Memtest86
SUSE provides a memory test application available only on x86 systems called memtest86. It is available from the boot screen when booting from the CD-ROM. This RAM test runs an endless read and write loop. You should let this test run for several hours. Reboot the system to terminate the test. For more information about this application, refer to the following Web site:
http://www.memtest86.com

Chapter 17. Troubleshooting

233

234

Solaris to Linux Migration: A Guide for System Administrators

Part 3

Part

IBM eServer platforms

In the following chapters, we provide an overview of Linux system administrator tasks and issues specific to the IBM Eserver platforms.

Copyright IBM Corp. 2006. All rights reserved.

235

236

Solaris to Linux Migration: A Guide for System Administrators

18

Chapter 18.

IBM eServer xSeries hardware platform specifics

This chapter covers the differences in the migration process as they relate to IBM Eserver xSeries and Eserver BladeCenter server hardware and describes some differences in setting up and managing these systems. This chapter describes the distinguishing differences in the hardware for the xSeries server platforms and BladeCenter platforms in the following areas: Installation considerations Hardware and device differences

Copyright IBM Corp. 2006. All rights reserved.

237

18.1 Installation considerations

Both Red Hat and Novell SUSE Linux distributions are architected to exploit the capabilities of the Eserver xSeries and BladeCenter servers. The xSeries and BladeCenter server platforms are both Red Hat and Novell SUSE certified. Before you install Linux, we recommend that you check the BIOS and firmware levels on the following items and update them to the most current revision as part of your installation procedure for the xSeries server: System BIOS Remote Supervisor Adapter firmware and BIOS Onboard diagnostics Additional devices if installed, such as ServeRAID adapters and FAStT Fibre Channel host adapters For the latest BIOS and firmware code, refer to:
http://www.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-4JTS2T

Follow the installation instructions provided with each package or those included with your server platform regarding updating firmware levels. For a complete driver listing of the most common server components, refer to:
http://www.ibm.com/pc/support/site.wss/DRVR-MATRIX.html

For more information about the booting and installation process, refer to Chapter 8, Boot and system initialization on page 125.

18.1.1 xSeries
Before starting an installation, make sure that any disk subsystems, such as RAID arrays, have already been configured. For more information about how to configure these systems, refer to the documentation included with the disk subsystem hardware. In general, the xSeries server platform is installed in much the same manner as other IA32-based architecture platforms. The xSeries specifics for installation starts with the selection of the boot device. If the system does not boot from the installation device that contains the Linux distribution, you need to enter the boot setup device screen at startup. Follow the instructions after initially powering on the system for what keys to press to enter the boot device configuration or refer to the system documentation.

238

Solaris to Linux Migration: A Guide for System Administrators

18.1.2 BladeCenter
All the blades share the CD-ROM, floppy drive, keyboard, video, and mouse. There are two I/O selection buttons on the front of each blade that allow for assignment of the CD-ROM and floppy drive and also for the keyboard, video, and mouse (KVM). There is also a power button on each blade that is protected by a hinged plastic flap. After a blade is powered up, pressing the CD-ROM or the KVM button on that blade selects the corresponding resource for that blade. On the blade that is currently connected to the CD-ROM or the KVM, the corresponding I/O selection button lights up solid green. Sharing the CD-ROM for all the blades is a limitation of installing on multiple blades. Using the CD-ROM, you can install individual blades one at a time; however, that process is time-consuming if you install more than one or two blades. As an alternative, install one blade and configure it to be a network-installed server. Subsequent operating system installations are then performed from that server. To read more about how to do this installation process, refer to the IBM Redbook IBM Eserver BladeCenter, Linux, and Open Source: Blueprint for e-business on demand, SG24-7034, available at:
http://www.redbooks.ibm.com/abstracts/sg247034.html

18.2 Hardware and device differences

Accessing the xSeries server and BladeCenter hardware for initial configuration and monitoring is done through special management hardware included with the system. This special management hardware depends on the purchased server options, but can include the integrated the Baseboard Management Controller (BMC), the Remote Supervisor Adapter II (RSA II), or the BladeCenter Management Module. Access the management hardware using the available user interfaces such as the Web interface, Management Processor Command Line Interface (MPCLI), Advanced Settings Utility (ASU), and the OSA System Management Bridge (SMBridge) Utility, as well as the ability to telnet or ssh directly to some models. For much more in-depth information about the xSeries and BladeCenter management hardware and service processors, refer to the IBM Redbook IBM Eserver xSeries and BladeCenter Server Management, SG24-6495, available at:
http://www.redbooks.ibm.com/abstracts/sg246495.html

Chapter 18. IBM eServer xSeries hardware platform specifics

239

IBM Director
For complete server management of xSeries and BladeCenter systems, the recommended tool is IBM Director. With IBM Director, you have complete access to the management hardware in addition to other management tasks such as event management, inventory, and deployment. IBM Director is available to xSeries platform customers from:
http://www.ibm.com/pc/support/site.wss/MIGR-57057.html

18.2.1 xSeries
This section provides functional details for IBM Eserver xSeries Baseboard Management Controller and Remote Supervisor Adapter II.

Baseboard Management Controller (BMC)

Many xSeries servers have service processors integrated onto the system board. These provide different levels of monitoring and alerting depending on the type and functionality of service processor.

Functionality
The integrated BMC might have all or part of the following functionality: Monitoring of system voltages Battery voltage monitor System temperature monitors Fan speed control Fan tachometer monitor Power Good signal monitor System ID and planar version detection System power control System reset control NMI detection SMI detection and generation Serial port text redirection Remind button detection System LEDs control (power, HDD activity, alert, and so on) Control of Lightpath LED

240

Solaris to Linux Migration: A Guide for System Administrators

For more information about the full functionality of the BMC, refer to the xSeries system documentation that came with the server.

External connection
The BMCs communicate through one of the integrated Ethernet adapters on the server. To communicate with the BMC, attach a standard Ethernet cable.

Configuration
There are two different methods for configuring the BMC. The BMC configuration can be done through the System Setup in BIOS (pressing F1 at boot time) and accessing BMC settings through the Advanced Options menu. Using System Setup in BIOS is the recommended method. After the server network settings are configured, you can use the Management Processor Assistant (MPA) task in IBM Director to configure the other required settings such as user IDs, passwords, and alert destinations in the BMC. The second configuration method uses the bmc_cfg.exe utility. This utility is located on the BMC firmware update diskette or CD. You can only run bmc_cfg by exiting to a MS-DOS-based program after booting the server from the bootable BMC management firmware update diskette or CD.

Remote Supervisor Adapter II (RSA II)

The Remote Supervisor Adapter II (RSA II) is an advanced management adapter for xSeries server platforms. It provides many options for alerting, monitoring, and remote management of xSeries servers. It replaces the earlier Remote Supervisor Adapter (RSA or RSA I) and the Advanced System Management PCI Adapter (ASMA).

Functionality
Some of the features and functionality of the RSA II include: Automatic notification and alerts Continuous health monitoring and control Event log LAN and Advanced System Management (ASM) interconnect remote access Operating system failure screen capture Remote media Remote power control Server console redirection

Chapter 18. IBM eServer xSeries hardware platform specifics

241

Connections
The RSA II supports various connections available on the card and on breakout connectors. Refer to Figure 18-1 for connector locations. There are variations on the location of the connectors available depending on the model of adapter and xSeries server. For specific connector locations on your adapter, refer to the system documentation that came with the xSeries server hardware.

Figure 18-1 xSeries Remote Supervisor Adapter II (RSA II)

The RSA II has the following connectors (the numbers refer to Figure 18-1): Video connector (1 in Figure 18-1): The RSA II contains an additional video subsystem on the adapter. When the RSA II is installed in a server, it automatically disables the on-board video. The servers monitor should be connected to the RSA II video connector. 10/100 Ethernet connector (2): For connection to a 10 Mbps or 100 Mbps Ethernet-based client LAN or management LAN. Power connector (3): Access to RSA II is possible if the server is powered down when the external power supply is used (supplied when the adapter is purchased as an option). Connect the power supply to a different power source as the server (for example, a separate UPS). Mini-USB connector (4): This port provides the ability for a remote keyboard and mouse when using the remote control feature. Connect this to a USB port of the server, except for the following servers where the cable should not be connected (the USB signal is transmitted inside these servers): xSeries 225 xSeries 226 xSeries 365 Eserver 326

242

Solaris to Linux Migration: A Guide for System Administrators

Breakout connector (5): To use RSA II as a focal point for an ASM network or for connecting a modem, a breakout cable is supplied, which has the ASM and serial connections. The breakout cable has two serial connectors (earlier RSA II adapters had only one serial port) and two RJ45 connectors for daisy chaining the ASM RS-485 interconnect network. 20-pin connector for the connection to the servers motherboard (6): Figure 18-1 on page 242 shows the connector on the planar where the supplied cable should be connected.

Accessing the RSA II

RSA II provides access to the server for the system administrator, even when the server is down. Monitoring alerts and management functions are performed directly with the RSA II service processor through one of the servers Gigabit Ethernet ports. The Advanced System Management (ASM) network is for interconnectivity of older service processors with the RSA II. The RSA II can work as a gateway for the ASM network. It can provide access to the connected management processors (that have no Ethernet connection) from the LAN and is the gateway for the management processors to the LAN. For more information about how to configure the RSA to act as a gateway to the ASM, refer to the IBM Redbook IBM Eserver BladeCenter, Linux, and Open Source: Blueprint for e-business on demand, SG24-7034, available at:
http://www.redbooks.ibm.com/abstracts/sg247034.html

Management Processor Command line Interface (MPCLI)

The Management Processor Command Line Interface (MPCLI) is a management tool for xSeries servers running Linux. The system management functions are provided from a command line interface (CLI) that connects to some models of the RSA II service processor in the server. Refer to the server documentation for specific information about this support. Using this CLI, you can access and set a wide range of information about the health, configuration, communication, and state of your system. These functions are available after you install the CLI and make a connection to the service processor. You can use the MPCLI on a remote server provided you know the IP address of the remote service processor and have a valid user ID and password. For more information about how to configure user IDs and passwords for this service, refer to the IBM Redbook IBM Eserver BladeCenter, Linux, and Open Source: Blueprint for e-business on demand, SG24-7034, available at:
http://www.redbooks.ibm.com/abstracts/sg247034.html

Chapter 18. IBM eServer xSeries hardware platform specifics

243

MPCLI supports three methods of communicating with a service processor: In-band communication using a device driver Out-of-band communication using an IP connection Out-of-band communication using an RS-485 interconnect For more information and to download the MPCLI, refer to following link:
http://www.ibm.com/pc/support/site.wss/MIGR-54216.html

Web interface
The Remote Supervisor Adapter II and BladeCenter Management Module have a built-in Web server that enables users to access these service processors using a Web browser. The browser must be Java enabled, support JavaScript 1.2 or later, and have a Java 1.4.1 or later plug-in installed. For best results when using the Web browser, ensure that your monitors resolution is at least 800x600 and at least 256 colors.

18.2.2 BladeCenter
The BladeCenter system provides shared resources to all the blades, such as power, cooling, system management, network connections, CD-ROM, floppy, keyboard, video, and mouse. The use of common resources allows the blades to be smaller and reduces the need for cabling. BladeCenter consists of a rack-mounted chassis. The front of BladeCenter supports 14 blade server slots and has a CD-ROM drive, USB port, and floppy drive. The back of the chassis has slots for two blower modules, four power modules, four network modules, and a Management Module.

BladeCenter Management Module

The BladeCenter Management Module has similar capabilities to the RSA II. There are some additional BladeCenter-specific features such as the integrated KVM switch. The BladeCenter Management Module acts like a global RSA II for all the installed blade servers in the chassis. See Remote Supervisor Adapter II (RSA II) on page 241 for more information about the features and functionality of the BladeCenter Management Module. The Management Module manages the BladeCenter chassis itself with all its networking modules (for example, Gigabit Ethernet or SAN) and all blade servers installed in the chassis. The Management Module has an integrated KVM switch and an integrated network switch for internal IP connections to all the modules, such as Ethernet switch modules (ESMs) and Fibre Channel switch modules, to

244

Solaris to Linux Migration: A Guide for System Administrators

manage the blade servers. Additionally, it acts like a RSA II for every installed blade server. For more information about how to connect and configure the Management Module, refer to Management Module configuration on page 264.

18.3 References
The following links provide more specific information about the Eserver xSeries and BladeCenter server hardware: From this chapter: The latest BIOS and firmware code is available at:
http://www.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-4JTS2T

For a complete driver listing of the most common server components, refer to:
http://www.ibm.com/pc/support/site.wss/DRVR-MATRIX.html

The IBM Redbook IBM Eserver BladeCenter, Linux, and Open Source: Blueprint for e-business on demand, SG24-7034, available at:
http://www.redbooks.ibm.com/abstracts/sg247034.html

The IBM Redbook IBM Eserver xSeries and BladeCenter Server Management, SG24-6495, available at:
http://www.redbooks.ibm.com/Redbooks.nsf/RedbookAbstracts/sg246495.html

IBM Director is available to xSeries platform customers from:

http://www.ibm.com/pc/support/site.wss/MIGR-57057.html

For more information and to download the MPCLI, refer to the following link:
http://www.ibm.com/pc/support/site.wss/MIGR-54216.html

Additional resources: More information about the IBM Eserver xSeries server platforms:
http://www.ibm.com/servers/eserver/xseries

For more information about the IBM BladeCenter server platform:

http://www.ibm.com/servers/eserver/bladecenter

For more information about the IBM Director management application:

http://www.ibm.com/servers/eserver/xseries/systems_management/ibm_director

Chapter 18. IBM eServer xSeries hardware platform specifics

245

246

Solaris to Linux Migration: A Guide for System Administrators

19

Chapter 19.

IBM POWER technology hardware platform specifics

In this chapter, we provide an overview of Linux system administrator tasks and issues specific to the IBM POWER5 technology for IBM Eserver and BladeCenter JS20 hardware platforms. We discuss the following topics: Planning Booting and system initialization Linux installation eServer i5 and eServer p5 virtualization POWER technology platform service and productivity tools References and further readings In this chapter, we provide guidance for basic system administration. Refer to the references listed in 19.6, References and further readings on page 304, for more POWER technology-based, platform-specific system administration information.

Copyright IBM Corp. 2006. All rights reserved.

247

19.1 Planning
This section describes issues that system administrators should consider before beginning the installation of an Eserver i5, Eserver p5, or Eserver BladeCenter JS20.

19.1.1 IBM eServer i5 and eServer p5

There are two general paths available for installing Linux on IBM Eserver i5 Eserver p5 systems depending on your configuration choice of a stand-alone system or a partitioned system.

Monolithic (stand-alone)
In a monolithic installation, Linux owns the entire server and all of its resources. This installation is only common for Eserver p5 systems. A monolithic install does not require any POWER technology-specific preinstallation planning. Note: The Eserver p5 system is initially preconfigured for monolithic operation.

Hosted (partitioned)
In a hosted installation, Linux runs in a partition, along with other instances of Linux. By using logical partitioning (LPAR), a single physical system can be divided into multiple logical partitions, each running their own operating system image. The Eserver i5 and Eserver p5 both support hosted installations.

Concepts for Linux logical partitions

The purpose of this information is to familiarize you with the hardware and software required for Linux logical partitions. Logical partitioning is the ability to make a server run as though it were two or more independent servers. When you logically partition a server, you divide the resources on the server into subsets called logical partitions. Processors, memory, and input/output devices are examples of resources that you can assign to logical partitions. You can install software on the logical partition, and the logical partition runs as an independent logical server with the processor, memory, and I/O resources that you have allocated to it.

248

Solaris to Linux Migration: A Guide for System Administrators

You must use tools to partition Eserver i5 and Eserver p5 servers. The tool that you use to partition each server depends on the server model and the operating systems and features that you want to use: Hardware Management Console (HMC) The Hardware Management Console (HMC) is a hardware appliance that connects to the server firmware. You use the HMC to specify to the server firmware how you want resources to be allocated among the logical partitions on the managed system. You also use the HMC to start and stop the logical partitions, update the server firmware code, manage IBM Eserver Capacity on Demand, and transmit service information to service and support if there are any hardware problems with your managed system. Figure 19-1 illustrates the logical partitions and the server firmware on the IBM Eserver hardware. The server firmware is code stored in system flash memory on the server. The server firmware directly controls the resource allocations on the server and the communications between logical partitions on the server. The HMC connects with the server firmware and specifies how the server firmware allocates resources on the server.

Figure 19-1 HMC illustration

If you use a single HMC to manage a server, and the HMC malfunctions or becomes disconnected from the server firmware, the server continues to run, but you will not be able to change the logical partition configuration of the server or manage the server. If desired, you can attach an additional HMC to act as a backup and to provide a redundant path between the server and IBM service and support. Partitioning using the HMC is supported on all IBM Eserver i5, IBM Eserver p5, IBM System p5, and IBM Eserver OpenPower server models, although some models require you to enter an Advanced POWER Virtualization activation code before you can partition the server. Integrated Virtualization Manager The Integrated Virtualization Manager is a browser-based system management interface for Virtual I/O Server that enables you to create and

Chapter 19. IBM POWER technology hardware platform specifics

249

manage Linux logical partitions on a single IBM Eserver p5 or IBM System p5 server or Linux logical partitions on a single IBM Eserver OpenPower server. The Integrated Virtualization Manager is supported only on select IBM Eserver p5, IBM System p5, and OpenPower server models. Virtual I/O Server is software that provides virtual storage and shared Ethernet resources to the other logical partitions on the managed system. Virtual I/O Server is not a general purpose operating system that can run applications. Virtual I/O Server is installed on a logical partition in the place of a general purpose operating system and is used solely to provide virtual I/O resources to other logical partitions with general purpose operating systems. You use the Integrated Virtualization Manager to specify to Virtual I/O Server how these resources are assigned to the other logical partitions. To use the Integrated Virtualization Manager, you must first install Virtual I/O Server on an unpartitioned server. Virtual I/O Server automatically creates a logical partition for itself, which is called the management partition for the managed system. The management partition is the Virtual I/O Server logical partition that controls all of the physical I/O resources on the managed system. After you install Virtual I/O Server, you can configure a physical Ethernet adapter on the server so that you can connect to the Integrated Virtualization Manager from a computer with a Web browser. Figure 19-2 on page 251 illustrates Virtual I/O Server in its own logical partition, and the Linux logical partitions that are managed by the Virtual I/O Server logical partition. The browser on the PC connects to the Integrated Virtualization Manager interface over a network, and you can use the Integrated Virtualization Manager to create and manage the logical partitions on the server.

250

Solaris to Linux Migration: A Guide for System Administrators

Figure 19-2 Virtual I/O server illustration

Virtual Partition Manager The Virtual Partition Manager is a feature of IBM i5/OS V5R3 that enables you to create and manage one i5/OS logical partition and up to four Linux logical partitions on a single IBM Eserver i5 server. To use the Virtual Partition Manager, you must first install i5/OS on an unpartitioned server. After you install i5/OS, you can initiate a console session on i5/OS and use Dedicated Service Tools (DST) or System Service Tools (SST) to create and configure Linux logical partitions. i5/OS controls the resource allocations of the logical partitions on the server. Figure 19-3 on page 252 illustrates the i5/OS logical partition and the Linux logical partitions that are managed by the i5/OS logical partition. The twinaxial console connects to the i5/OS logical partition, and you can use DST or SST to create and configure the Linux logical partitions on the server.

Chapter 19. IBM POWER technology hardware platform specifics

251

Figure 19-3 Virtual partition manager illustration

When you use the Virtual Partition Manager to partition an Eserver i5 server, DST and SST are the only tools that you can use to create and manage the logical partitions. You cannot use iSeries Navigator to create or manage logical partitions on an Eserver i5 server. However, the console session that you use to access DST or SST can be initiated using either iSeries Operations Console (LAN or direct attach) or a twinaxial console. Although each logical partition acts as an independent server, the logical partitions on a server can share some kinds of resources with each other. The ability to share resources among many logical partitions enables you to increase resource utilization on the server and to shift the server resources to where they are needed. The following list illustrates some of the ways in which logical partitions can share resources. (For some server models, the features mentioned in this list are options for which you must obtain and enter an activation code.) Micro-Partitioning (or shared processing) enables logical partitions to share the processors in the shared processor pool. The shared processor pool includes all processors on the server that are not dedicated to specific logical partitions. Each logical partition that uses the shared processor pool is assigned a specific amount of processor power from the shared processor pool. If the logical partition needs more processor power than its assigned amount, the logical partition is set by default to use the unused processor power in the shared processor pool. The amount of processor power that the logical partition can use is limited only by the virtual processor settings of the logical partition and the amount of unused processor power available in the shared processor pool.

252

Solaris to Linux Migration: A Guide for System Administrators

Dynamic logical partitioning enables you to move resources to, from, and between running logical partitions manually without shutting down or restarting the logical partitions. This enables you to share devices that logical partitions use occasionally. For example, if the logical partitions on your server use an optical drive occasionally, you can assign a single optical drive to multiple logical partitions as a desired device. The optical drive would belong to only one logical partition at a time, but you can use dynamic logical partitioning to move the optical drive between logical partitions as needed. On servers that are managed using the Integrated Virtualization Manager, dynamic logical partitioning is supported only for the management partition. Dynamic logical partitioning is not supported on servers that are managed using the Virtual Partition Manager. Virtual I/O enables logical partitions to access and use I/O resources on other logical partitions. For example, virtual Ethernet enables you to create a virtual LAN that connects the logical partitions on your server to each other. If one of the logical partitions on the server has a physical Ethernet adapter that is connected to an external network, you can configure the operating system of that logical partition to connect the virtual LAN with the physical Ethernet adapter. This enables the logical partitions on the server to share a physical Ethernet connection to an external network.

Hardware requirements for Linux logical partitions

During the process of planning for logical partitions, you must decide how you want to configure hardware resources. Each Linux logical partition on an IBM Eserver hardware system requires the following minimum hardware resources: One dedicated processor or 0.1 processing unit 128 MB memory One storage adapter (physical or virtual) Approximately 1GB of space provided by one of the following storage options: Internal storage using SCSI adapters and drives attached within the system External storage (SAN) using SAN adapters and drives in an external storage unit Virtual storage using a virtual SCSI adapter and storage in a different partition One network adapter (physical or virtual)

Chapter 19. IBM POWER technology hardware platform specifics

253

Logical partition planning tasks

Use the following planning tasks: 1. Identify the hardware requirements for each logical partition based on the hardware configuration of the server. 2. Identify whether the partitions will communicate with other partitions, servers, or workstations using physical or virtual Ethernet connections. 3. Design and validate your partition configuration using the IBM LPAR Validation Tool (LVT), available at:
http://www.ibm.com/servers/eserver/iseries/lpar/

The LVT provides you with a validation report that reflects your system requirements while not exceeding logical partition recommendations. 4. If your machine is an Eserver i5, decide if you want i5/OS to provide Linux logical partitions. 5. Plan for Linux software licensing in a partitioned environment by reading and understanding the license agreement for your Linux distribution.

19.1.2 IBM eServer BladeCenter JS20

This section discusses planning considerations associated with the implementation of IBM Eserver BladeCenter JS20. Specifically, it covers network planning.

Network planning
Successful installation of BladeCenter JS20 requires that you have a clear plan for how you will use the various networking capabilities of the BladeCenter infrastructure. This plan should address the following questions: What network connectivity is needed for the blade servers to support the applications installed on them? What network connectivity is needed to manage the BladeCenter, input/output (I/O) modules, and blade servers? What virtual local area networks (VLANs) are required for each LAN Switch I/O Module to provide the network connectivity established previously? What IP subnet will be used for each VLAN and how will IP addresses be allocated to each device on the VLAN? Will IP addresses be assigned statically or dynamically using DHCP? What host names will be used for each network interface? How will host names be resolved to IP addresses?

254

Solaris to Linux Migration: A Guide for System Administrators

Are there requirements for a high-performance, low-latency interconnection network? Where multiple BladeCenter chassis are installed, how will they be interconnected? The following topics explore common network planning requirements that we expect to arise in the installation of BladeCenter JS20s.

Minimal network requirements

At a minimum, most BladeCenter JS20 environments have the following network requirements: A dedicated hardware management subnet: It is used to access both the Management Module and management interfaces of I/O modules that are installed in each BladeCenter chassis. A Serial over LAN (SoL) subnet internal to each BladeCenter chassis that supports the SoL remote text console function: This is always implemented using a LAN Switch I/O Module installed in I/O module bay 1. A subnet connected to each BladeCenter JS20: It is used to install and manage the operating system on the blade server. One or more subnets connected to each BladeCenter JS20: They are used by applications installed on the blade server to communicate to other systems. Figure 19-4 on page 256 illustrates how these requirements can be provided in a logical network view.

Chapter 19. IBM POWER technology hardware platform specifics

255

Figure 19-4 Network logical view

The following four sections describe each of the logical networks illustrated in Figure 19-4.

256

Solaris to Linux Migration: A Guide for System Administrators

Hardware management subnet

We recommend that you establish a dedicated hardware management subnet. The 10/100BaseT Ethernet interface on the Management Modules installed in each BladeCenter chassis provides the gateway to this subnet for each BladeCenter chassis. You need to install external LAN switches or hubs to interconnect the 10/100BaseT Ethernet interface on the Management Modules of each BladeCenter chassis with external management systems. You use this subnet to access the Management Module Web interface and command line interface (CLI). You also use this subnet to access the Web interface and CLI of I/O modules. System management applications such as IBM Director or IBM Cluster Systems Management also use this subnet to communicate with the hardware management functions of the BladeCenter infrastructure. Restrict access to this subnet to those management systems, system administrators, and operators who have responsibility for managing the BladeCenter infrastructure. You need to allocate multiple IP addresses to each BladeCenter chassis on this subnet, including: One IP address for the external interface of the Management Module in each BladeCenter chassis One IP address for the internal interface of the Management Module in each BladeCenter chassis One IP address for the management interface of each I/O module in each BladeCenter chassis Note: Although the logical network view illustrated in Figure 19-4 shows the I/O Management Module interfaces connecting directly to the hardware management subnet, they are physically connected through the Management Module, which acts as gateway to those interfaces. The Management Module performs a proxy Address Resolution Protocol (ARP) function to make it appear as though the I/O module management interfaces are attached to the hardware management subnet. It is possible to use a different subnet for the Management Module internal network interface and I/O module management interfaces. However, we do not recommend this configuration.

Chapter 19. IBM POWER technology hardware platform specifics

257

Serial over LAN subnet

The SoL remote text console function requires a subnet and underlying VLAN that is implemented by a LAN Switch I/O Module installed in I/O module bay 1 of the BladeCenter chassis. This subnet and VLAN is entirely internal to each BladeCenter chassis and should not be externally accessible. If you use the 4-Port Gigabit Ethernet Switch Module or the Nortel Networks Layer 2-7 Gigabit Ethernet Switch Module, the VLAN uses VLAN ID 4095. If you use the Cisco Systems Intelligent Gigabit Ethernet Switch Module, you can choose the VLAN ID. You need to assign a unique range of IP addresses to this subnet for use by the SoL remote text console function. Important: One IP address is required for each blade server. You need to specify only the starting IP address within the range of IP addresses that you assign into the Management Module. The Management Module then automatically assigns consecutive IP addresses from the starting address that you provide to each blade server that you have installed.

Operating system management subnet

We expect most environments that use the BladeCenter JS20 to rely on the network installation procedure to install the operating systems. The operating system management subnet is used to support both the initial installation and subsequent management of the operating systems installed on BladeCenter JS20s. This subnet is implemented using a VLAN provided by the LAN Switch I/O Modules installed in I/O module bay 2 of each BladeCenter chassis. Alternatively, you can implement it using a Pass-Thru I/O Module installed in I/O module bay 2 that is connected to an external LAN switch.

Application subnet
The primary reason you install BladeCenter JS20s is to support applications. Many applications have requirements to communicate with other systems. You use one or more application subnets for this purpose. The primary application subnet is implemented using a VLAN provided by the LAN Switch I/O Modules installed in I/O module bay 1 of each BladeCenter chassis. The same LAN Switch I/O Module is used to support the SoL subnet and VLAN. Where different BladeCenter JS20s participate in different applications and there is a requirement to separate application traffic, you might need to define multiple

258

Solaris to Linux Migration: A Guide for System Administrators

application subnets and VLANs. Each BladeCenter JS20 is connected to the appropriate application subnet based on the applications that are installed on the blade server. If application communication requirements with other systems are complex, you can install an additional pair of Gigabit Ethernet interfaces on each BladeCenter JS20. You do this using the Gigabit Ethernet Expansion Card in conjunction with compatible I/O modules installed in I/O module bays 3 and 4.

19.2 Booting and system initialization

This section describes the booting and system initialization process of Eserver i5, Eserver p5, and JS20 blade. For simplicity, we only cover booting and installing from the CD in this section. For additional methods, refer to your hardware and Linux distribution installation guides.

19.2.1 IBM eServer i5 and eServer p5

The initial step of powering on the server depends on whether the system is monolithic or hosted.

Monolithic (stand-alone)
There are three ways to power on a monolithic Eserver p5: Power switch The Advanced System Management Interface (ASMI) through a serial connection The ASMI Web interface Only the first and second option are available when the machine is initially unpacked from the box and plugged in. The third option requires the configuration of the Hardware Management Console (HMC) network port using the ASMI serial connection. For more information about the ASMI, refer to Managing the Advanced System Management Interface (ASMI), available at:
http://publib.boulder.ibm.com/infocenter/eserver/v1r3s/index.jsp

Tip: Use ASMI to set the systems time and date before beginning any Linux installations. This will help avoid potential problems during Linux installations. After powering on, a monolithic system proceeds to progress through a series of hardware checks. The LED panel on the front of the machine will flash various codes to indicate this progression.

Chapter 19. IBM POWER technology hardware platform specifics

259

If the machine is booted with an ASCII terminal connected to its serial port, these same LED codes will display on the ASCII terminal. Nothing will show on a console connected to the graphics card during this process.

Hosted (partitioned)
The servers HMC network port must be configured using the ASMI serial connection. For more information, refer to Managing the Advanced System Management Interface (ASMI), available at:
http://publib.boulder.ibm.com/infocenter/eserver/v1r3s/index.jsp

Tip: Use ASMI to set the systems time and date before beginning any Linux installations. This will help avoid potential problems during Linux installations. The Hardware Management Console must be configured and the server attached before powering up. Refer to Managing the Hardware Management Console (HMC), available at:
http://publib.boulder.ibm.com/infocenter/eserver/v1r3s/index.jsp

When connected to an HMC, there are three ways to power on Eserver i5 and Eserver p5 systems: The HMC The Advanced System Management Interface (ASMI) through a serial connection The ASMI Web interface Restriction: When connected to an HMC, Eserver i5 and Eserver p5 will not power up through the power switch. We recommend using the HMC to power on a hosted system into the partition standby mode. This mode powers the machine to a level that allows partition creation and manipulation. To power on a managed system to this mode, complete the following steps: 1. In the Navigation area, expand the Server and Partition folder. 2. Click the Server Management icon. 3. In the Contents area, select the managed system. 4. From the menu, click Selected Power On. 5. Select Standby power-on mode and click OK. After powering on, a hosted system proceeds to progress through a series of hardware checks. The LED panel on the front of the machine and the panel

260

Solaris to Linux Migration: A Guide for System Administrators

display for the machine in the HMC window will flash various codes to indicate this progression. The system status panel displays Standby after the power on process has completed.

System management services (SMS) menu

This section describes the basic steps involved in booting a monolithic system or partition to the point where the system management services menu becomes available. The SMS menu enables the system administrator to configure, among other things, the boot options for the system or partition. There are two ways to boot into the SMS menu: Preboot: This involves using the ASMI power on options. During boot: This involves using the console or HMC terminal. We only cover the second method in this section. For information about the first method, refer to Managing the Advanced System Management Interface (ASMI), available at:
http://publib.boulder.ibm.com/infocenter/eserver/v1r3s/index.jsp

Monolithic (stand-alone)
During the power-on process, the console (ASCII or graphical) will eventually display information about the final stage of hardware checking. This stage is where the SMS menu becomes available. To access the SMS menu, the usual method involves pressing the F1 key after the keyboard icon appears for a graphical console (see Figure 19-5 on page 262), or pressing the 1 key after the word keyboard appears for a ASCII terminal (see Figure 19-7 on page 264). Consult your hardware guide to confirm the correct graphical key sequence. The menu opens after the hardware check process completes. Note: If prompted for a password, the default password is admin.

Chapter 19. IBM POWER technology hardware platform specifics

261

Figure 19-5 Graphical SMS example

Hosted (partitioned)
When the system is in standby mode, the next step involves creating and booting a partition. For steps for creating partitions, as well as example partition configurations, refer to Partitioning for Linux, available in the IBM Eserver Information Center:
http://publib.boulder.ibm.com/infocenter/eserver/v1r3s/index.jsp

In hosted mode, the partition to install must have the CD/DVD drive in its partition profile. To assign the CD/DVD drive to a partition: 1. Go to Server and Partition Server Management. 2. Open the server and partition to which you want to install. 3. Right-click the profile you want to use for installation and select Properties. 4. In the Logical Partiton Profile Properties window, go to the Physical I/O tab. 5. From the Managed system I/O devices window, select the Other Mass Storage Controller from the bus where it is attached. 6. Click Add as required to assign this CD/DVD drive to the partition. The result should look similar to the example in Figure 19-6 on page 263.

262

Solaris to Linux Migration: A Guide for System Administrators

Figure 19-6 Adding a CD device example

To boot an Eserver p5 partition: 1. Right-click the partition. 2. Select Activate. 3. Select the Open terminal option. 4. Click OK. To boot an Eserver i5 partition: 1. Right-click the partition. 2. Select Open Terminal Window. 3. Vary on the partition from i5/OS. During the power on process, the terminal console eventually displays information about the final stage of hardware checking. This stage is where the SMS menu becomes available. To access the SMS menu, press the 1 key after the word keyboard appears (see Figure 19-7 on page 264). The menu appears after the hardware check process completes.

Chapter 19. IBM POWER technology hardware platform specifics

263

Figure 19-7 Text SMS example

Boot device selection

When in the SMS menu, select the Select Boot Options option. In that menu, specify Select Install or Boot a Device. There, select CD/DVD, and then the bus type (IDE or SCSI). If you are uncertain, you can select to view all devices. This scans all available buses for boot devices, including network adapters and hard drives (will not display empty or non-bootable drives). Finally, select the device containing the installation CD.

19.2.2 eServer BladeCenter JS20

We cover the following topics in this section: Management Module configuration LAN Switch I/O Module configuration Blade server configuration Providing a console for the JS20 blades

Management Module configuration

This section describes the steps to set up the BladeCenter Management Module so that you can work with BladeCenter JS20s. Advanced configuration of the Management Module is beyond the scope of this section.

264

Solaris to Linux Migration: A Guide for System Administrators

The primary setup task for the Management Module is to assign IP addresses, which are necessary to communicate with the Management Module Web and command line interfaces. The Management Module has two network interfaces, an external interface (eth0) and an internal interface (eth1). The external interface is accessible through the 10/100BaseT connector on the Management Module. The internal interface is connected to the management interfaces of all the installed I/O modules that support such interfaces. This includes all switch I/O modules. The default behavior of a new Management Module is to request an IP address for the external interface through DHCP. If the Management Module does not receive a valid response from a DHCP server within two minutes of powering on, it uses a default static IP address of 192.168.70.125 with the default subnet mask of 255.255.255.0. Note: The default host name is MMxxxxxxxxxxxx, where xxxxxxxxxxxx is the burned-in medium access control (MAC) address. The default IP address for the internal interface is statically assigned to be 192.168.70.126 with a subnet mask of 255.255.255.0. You can reset the IP addresses of a Management Module that was previously configured back to the factory defaults by using the IP reset button on the Management Module. For the procedure for doing this, see IBM Eserver BladeCenter Management Module Users Guide, available at:
http://www.ibm.com/pc/support/site.wss/MIGR-45153.html

Configuring the Management Module

Use the following procedure to set the IP addresses for the Management Module external and internal interfaces: 1. Connect the Management Module to an isolated private Ethernet network. Also connect a workstation that has a Web browser to the same isolated private Ethernet network. 2. Configure a static IP address for the workstation Ethernet interface that is in the same subnet as the Management Module default IP addresses. For example, we used a static address of 192.168.70.100 with a subnet mask of 255.255.255.0 for the workstation Ethernet interface. Restriction: Do not use addresses in the range of 192.168.70.125 through 192.168.70.130. These addresses conflict with the default addresses assigned by the Management Module.

Chapter 19. IBM POWER technology hardware platform specifics

265

3. Connect to the Management Module Web interface by pointing your Web browser on the workstation to:
http://192.168.70.125

4. Enter a valid user ID and password. The factory default configuration of a Management Module defines a user ID named, USERID, with a password of PASSW0RD. The 0 in the password is a zero. In production environments, consider changing these defaults. 5. From the Management Module Web interface, select MM Control Network Interfaces. 6. The form window shown in Figure 19-8 on page 267 opens. Enter the desired external and internal IP addresses, subnet masks, and default gateway for the Management Module. Then, click Save to store the new IP addresses.

266

Solaris to Linux Migration: A Guide for System Administrators

Figure 19-8 IP address configuration

7. Restart the Management Module. In the Management Module Web interface, select MM Control Restart MM. Then, click the Restart button on the displayed form. You are prompted to confirm the restart before it occurs. 8. Remove the Management Module from the isolated private Ethernet network and connect it to the network you will use to manage BladeCenter. You can now connect to the Management Module Web and command line interfaces using the IP address that you assigned to the Management Module external network interface.

Chapter 19. IBM POWER technology hardware platform specifics

267

Now, consider performing other Management Module setup tasks, such as: Setting the Management Module date and time so that log entries have useful time stamps. Defining user IDs and passwords for the system administrators and operators who will manage the BladeCenter. Alternatively, you can configure the Management Module to use a Lightweight Directory Access Protocol (LDAP) directory for this purpose. Configuring the Management Module to send alerts to management systems through SNMP, or system administrators using e-mail through Simple Mail Transfer Protocol (SMTP). Enabling the use of Secure Sockets Layer (SSL) to securely access the Management Module Web interface. Enabling the use of Secure Shell (SSH) to securely access the Management Module command line interface (CLI). For additional information about how to perform these tasks, refer to the IBM Eserver BladeCenter Management Module Users Guide, available at:
http://www.ibm.com/pc/support/site.wss/MIGR-45153.html

LAN Switch I/O Module configuration

This section explains the basic set up for LAN Switch I/O Modules so that you can work with BladeCenter JS20s. The advanced configuration of LAN Switch I/O Modules is beyond the scope of this section. The primary setup tasks for LAN Switch I/O Modules are to: 1. Assign IP addresses to the LAN Switch I/O Module management interfaces. 2. Enable the LAN Switch I/O Module external Ethernet ports. When a new LAN Switch I/O Module is first installed, the Management Module assigns a default IP address to the management interface of the LAN Switch I/O Module. The default IP address is chosen based on the I/O module bay where the LAN Switch I/O Module is installed. Those I/O modules installed in I/O module bays 1, 2, 3, and 4 are assigned IP addresses 192.168.70.127, 192.168.70.128, 192.168.70.129, and 192.168.70.130, respectively. Set the IP address of each LAN Switch I/O Module management interface and enable the external ports: 1. Connect to the Management Module using the Web browser interface as explained in Configuring the Management Module on page 265.

268

Solaris to Linux Migration: A Guide for System Administrators

2. From the Management Module Web interface, select I/O Module Tasks Management. 3. In the form shown in Figure 19-9, scroll down to the entry for the LAN Switch I/O Module that you want to configure. Enter the IP address, subnet mask, and gateway that you want to assign to the LAN Switch I/O Module management interface. Click Save to activate the new IP address.

Figure 19-9 LAN Switch I/O Module IP address

4. You are prompted to confirm that you want to change the IP address. 5. Select the Advanced Management link for the LAN Switch I/O Module. 6. Scroll down to the Advanced Setup section of the displayed form, as illustrated in Figure 19-10 on page 270. For the External ports field, select Enabled from the drop-down list and then click Save.

Chapter 19. IBM POWER technology hardware platform specifics

269

Figure 19-10 LAN Switch I/O Module setup

At this point, the LAN Switch I/O Module management interface has an IP address. In addition, the external ports on the LAN switch module are enabled so that they can be used to communicate with blade servers. The SoL remote text console function of the Management Module depends on a VLAN provided by a LAN Switch I/O Module installed in I/O module bay 1. This VLAN is automatically provided by the 4-Port Gigabit Ethernet Switch Module and the Nortel Networks Layer 2-7 Gigabit Ethernet Switch Module using VLAN 4095.

270

Solaris to Linux Migration: A Guide for System Administrators

If you are using a Cisco Systems Intelligent Gigabit Ethernet Switch Module in I/O module bay 1, you must manually configure this VLAN. A procedure is documented in the Cisco Systems Intelligent Gigabit Ethernet Switch Module for the IBM Eserver BladeCenter Installation Guide. This Cisco guide describes how to manually set up the VLAN that is needed to support the SoL remote text console function and is available at:
http://www.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-57858

Blade server configuration

Minimal configuration is required for each BladeCenter JS20 prior to installing an operating system. The two main configuration tasks are: 1. Assigning a name to the blade server 2. Setting the blade server boot sequence The easiest way to perform these tasks is through the Management Module Web interface.

Assigning names to blade servers

Set the name of each blade server in a BladeCenter chassis by using these steps: 1. From the Management Module Web interface, select Blade Tasks Configuration. 2. In the Blade Information section of the form, as shown in Figure 19-11 on page 272, type the name that you want to assign to each blade server. 3. Scroll down the form and select Save to save the names that you have assigned to each blade server.

Chapter 19. IBM POWER technology hardware platform specifics

271

Figure 19-11 Blade Server naming

Setting the boot sequence

Set the boot sequence of each blade server by using this procedure: 1. From the Management Module Web interface, select Blade Tasks Configuration. 2. Scroll down the form to the Boot Sequence section to see the current boot sequence for all the blade servers. 3. If a blade server does not have the correct boot sequence, you can change the boot sequence by selecting the blade server name, which opens the form shown in Figure 19-12 on page 273. 4. Set the desired boot sequence using the drop-down lists and click Save to set the new boot sequence.

272

Solaris to Linux Migration: A Guide for System Administrators

The correct boot sequence depends on the method that you plan to use to install an operating system on the blade server. To install Linux from a CD, set the CD-ROM first.

Figure 19-12 Boot Sequence of blades

Providing a console for the JS20 blades

The BladeCenter JS20 differs from other blade servers that are available for the BladeCenter, in that it does not provide an interface to a keyboard, video monitor, and mouse (KVM) console. Therefore, you must set up the Serial over LAN (SoL) remote text console function to provide a console for a BladeCenter JS20. Use of SoL is optional for other types of blade servers that support KVM consoles. The SoL remote text console function involves several different components in the BladeCenter infrastructure and works as follows: 1. Using a Telnet or SSH client, connect to the BladeCenter Management Module CLI. This is usually through an external management network that is connected to the Management Modules 10/100BaseT Ethernet interface. 2. From the Management Module CLI, initiate a SoL remote console session to the desired blade server. 3. The Management Module uses a private VLAN provided by the LAN switch module in I/O module bay 1 to transport the SoL data stream to the Ethernet interface of the target blade server.

Chapter 19. IBM POWER technology hardware platform specifics

273

4. The Ethernet controller of the target blade server passes the SoL data stream received from the private VLAN to the blade system management processor (BSMP), which manages the text console for the blade server.

Configuring Serial over LAN

Before you attempt to configure the SoL remote text console function, verify that you have all the prerequisites in place: A supported LAN switch module installed in I/O module bay 1 This switch module is used to provide a VLAN that connects the Management Module to the first Ethernet interface on each blade server. The correct firmware levels This is important if you install a BladeCenter JS20 in an existing BladeCenter chassis that might have back-level firmware in the Management Module or LAN switch modules. A reliable Telnet or SSH client A network connecting the Telnet or SSH client to the BladeCenter Management Module external 10/100BaseT Ethernet interface An identified range of IP addresses that will be used by the Management Module to communicate with the BSMP on each blade server through the private VLAN The easiest way to configure the SoL remote text console function is through the Management Module Web interface as explained in the following steps. To configure the SoL remote text console function, use the following procedure: 1. From the Management Module Web interface, select Blade Tasks Serial Over LAN. 2. In the right pane, scroll down until you see the Serial Over LAN Configuration section (see Figure 19-13 on page 275). Complete the following tasks: a. From the Serial over LAN list, select the Enabled option. b. Leave the value for SoL VLAN ID at the default (4095) if you have either a 4-port Gigabit Ethernet Switch Module or a Nortel Networks Layer 2-7 Gigabit Ethernet Switch Module installed in I/O module bay 1. If you have a Cisco Systems Intelligent Gigabit Ethernet Switch Module installed in I/O module bay 1, set the VLAN ID to the same value you used when you manually defined the VLAN that supports SoL. c. Enter the start of the IP address range that will be used by the Management Module to communicate with the BSMP on each blade server.

274

Solaris to Linux Migration: A Guide for System Administrators

d. Leave the values for Accumulate timeout, Send threshold, Retry count, and Retry interval at their defaults (5, 250, 3, and 250). e. Leave the values for User Defined Keystroke Sequences at their defaults. f. Click Save.

Figure 19-13 Serial Over LAN Configuration

3. Restart the Management Module: a. In the Management Module Web interface, select MM Control Restart MM. b. Click the Restart button on the displayed form. c. You are prompted to confirm the restart before it occurs. 4. After the Management Module has restarted and you have reconnected to the Management Module Web interface from your Web browser, enable the SoL remote text console for each blade server. a. Select Blade Tasks Serial Over LAN from the Management Module Web interface. b. In the right pane, scroll down until you see the Serial Over LAN Status section (see Figure 19-14 on page 276).

Chapter 19. IBM POWER technology hardware platform specifics

275

c. Select the blade servers you want to enable for SoL. You can choose all of them by selecting the check box at the top of the table. d. Click the Enable Serial Over LAN link under the table. You might need to scroll down to see this link.

Figure 19-14 Serial Over LAN Status

e. After a few seconds, the window refreshes. In the SOL column of the table, verify that each blade server has a status of Enabled. The configuration of the SoL remote text console function is now complete.

Using Serial over LAN

Access the SoL remote text console function through the Management Module CLI. The CLI is documented in the IBM Eserver BladeCenter Management Module Command-Line Interface Reference Guide, available at:
http://www.ibm.com/support/docview.wss?uid=psg1MIGR-54667

You can connect to the Management Module CLI using either a Telnet or SSH client. The Management Module can support up to 20 concurrent CLI

276

Solaris to Linux Migration: A Guide for System Administrators

connections. This is sufficient to support the concurrent use of a SoL remote text console to each blade server in a full BladeCenter chassis. At the same time, it supports six additional CLI connections for other administrative activities. Restriction: You can only have one active SoL remote text console connection to each blade server. The Management Module CLI is context sensitive. When you issue CLI commands, you can accept the current context or override the context using the -T option provided on most commands. In our examples, we use the -T option to make it clear on which entity the command is operating. You can change the context for the Management Module CLI by using the env -T command. To use the SoL remote text console function, first connect to the Management Module from a Telnet or SSH client. You are prompted to enter a user ID and password. The default user ID is USERID and the default password is PASSW0RD, where 0 in the default password is a zero. Consider changing the defaults in a production environment. There are several different commands that you can use to access the SoL remote text console function. Table 19-1 lists the most useful commands.
Table 19-1 Management Module commands for SoL Command console Description Open a SoL remote text console for the blade server. This command fails if another SoL remote text console is already open for the blade server. Terminate any existing SoL remote text console for the blade server and open a SoL remote text console for the blade server. Reset the blade server, and then open a SoL remote text console for the blade server. This is functionally equivalent to the boot -c command when used in a blade server context. Power on the blade server and then open a SoL remote text console for the blade server. Power on the blade server and then open a SoL remote text console for the blade server. If the blade server is already powered on, power it off first, and then power on.

console -o

boot -c reset -c power -on -c power -cycle -c

Chapter 19. IBM POWER technology hardware platform specifics

277

Here are some examples of how to use these commands. To open a SoL remote text console to the blade server in bay 3, use the following command:
console -T system:blade[3]

To reset the blade server in bay 2 and then open a SoL remote text console to the blade server, use the following command:
boot -c -T system:blade[2]

To terminate an active SoL remote text console, press the Esc key followed by an open parenthesis (Shift+9 on U.S. keyboards). When the SoL remote text console ends, you return to the Management Module CLI prompt.

19.3 Linux installation

This chapter explains how to install Red Hat Enterprise Linux (RHEL) and Novell SUSE Linux Enterprise Server (SLES) on IBM Eserver i5, Eserver p5, and Eserver BladeCenter JS20 using the CDs and configure the YaBoot OpenFirmware boot loader. For information about performing a network install of either distribution, refer to Chapter 6, Installing Linux, in the IBM Redbook The IBM Eserver BladeCenter JS20, SG24-6342:
http://www.redbooks.ibm.com/abstracts/sg246342.html

This section is applicable to Eserver BladeCenter JS20 and the Eserver i5 and Eserver p5 platforms.

19.3.1 eServer i5 and eServer p5

After configuring the system to boot from the CD-ROM using the SMS menu, insert the first CD from your RHEL or SLES distribution. Select Normal Mode Boot, followed by selecting 1. Yes to exit System Management Services. Refer to 19.3.3, Red Hat on page 279, or 19.3.4, Novell SUSE on page 280, for additional information about installing Linux on Eserver i5 and Eserver p5 systems or a partition.

19.3.2 eServer BladeCenter JS20

After configuring the system to boot from the CD-ROM using the BladeCenter Management Module, insert the first CD from your RHEL or SLES distribution. Select Blade Tasks Power/Restart and connect to the console using SoL. Refer to 19.3.3, Red Hat on page 279, or 19.3.4, Novell SUSE on page 280, for additional information about installing Linux on a BladeCenter JS20 blade.

278

Solaris to Linux Migration: A Guide for System Administrators

19.3.3 Red Hat

For information about installing Linux on Eserver i5 and Eserver p5 systems or a partition or a JS20 blade server, refer to Installation Guide for the IBM POWER Architecture, available at:
http://www.redhat.com/docs/manuals/enterprise/

These are the steps for performing a basic install of RHEL on Eserver i5 and Eserver p5 systems or a partition or a JS20 blade server from a CD: 1. At the yaboot prompt (boot:), press Enter to begin the installation. Note: Refer to Additional RHEL installation boot options on page 280 for information about providing additional options, such as using VNC. 2. In the CD Found window, click Skip if you do not want to test the CD media, and go to step 3. Otherwise, click OK if you want to test the CD media before you proceed with the installation. a. In the Media Check window, click Test to test the CD currently in the CD-ROM drive. b. If the media check result is PASS, click OK and proceed with the installation from that CD. c. If the CD is ejected after the media check, reinsert the CD into the CD-ROM drive. d. Click Continue. 3. In the Red Hat Enterprise Linux AS welcome window, click OK. 4. Select your language; then click OK. 5. In the Disk Partitioning Setup window, select Autopartition. If you get a warning that says the drive will be initialized and all data will be lost, select Yes. 6. In the Partitioning window, you can see how the devices will be allocated by the Linux installation. Click OK. Be aware that you will need approximately 3 GB allocated for the root (/) partition. Important: The Eserver i5 and Eserver p5 installations require /sda1 to be a PReP boot partition. Changing the size of or removing this partition will cause the system to fail during boot. 7. Configure the appropriate network settings for your environment.

Chapter 19. IBM POWER technology hardware platform specifics

279

8. When your network configuration is complete, the Firewall and SELinux windows open. Click OK if you want to enable the default configurations. 9. In the Language Support window, select any additional languages that you want to use on this Linux installation, and then click OK. 10.In the Time Zone Selection window, select System clock uses UTC and select the appropriate time zone for your system. Click OK. 11.Set the root password for your system. Click OK. 12.Select what packages to install in the Package Defaults window. Click OK. 13.In the Installation to begin window, click OK. 14.The installation process prompts you to change the CD several times. After the installation is complete, the system will inform you to remove the CD and reboot. Note: The SMS menu should be invoked after reboot on Eserver i5 and Eserver p5 systems to ensure that the boot device order has the installed hard drive as the first device. JS20 blade servers should have the hard disk listed as second in the boot order. If not, update this using the BladeCenter Management Module.

Additional RHEL installation boot options

For the complete list of boot and kernal options for the RHEL installation program, refer to the following Web site:
http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/ppc-multi-insta ll-guide/ap-bootopts.html

19.3.4 Novell SUSE

For information about installing Linux on Eserver i5 and Eserver p5 systems or a partition, refer to the SUSE Linux Enterprise Server 9 Architecture-Specific Information for IBM POWER, available at:
http://www.novell.com/documentation/sles9/index.html

These are the steps for performing a basic installation of SLES on Eserver i5 and Eserver p5 systems or a partition from a CD: 1. At the prompt (boot:), press Enter to begin the installation. Note: Refer to the Additional SLES install boot options on page 282 for information about providing additional options, such as using VNC.

280

Solaris to Linux Migration: A Guide for System Administrators

2. At the prompt, What type of terminal do you have?, select 4) X Terminal Emulator (xterm). 3. After you have read the License Agreement, select I Agree. 4. Select your language. Click Accept. 5. Select New Installation. 6. Select Partitioning. 7. In the Suggested Partitioning window, select Create custom partition setup, and then click Next for a list of the available hard drives. Important: Eserver i5 and Eserver p5 installations require /sda1 to be a PReP boot partition. Changing the size of or removing this partition will cause the system to fail during boot. 8. Select the drive you want to use, and then click Next. 9. To create a full system partition, click Use entire hard disk. 10.A warning appears; select Yes, install to begin the installation. During this part of the installation procedure, you might need to change the CD. Follow the prompts. At the end of this phase of the installation, the system restarts. Note: The SUSE Linux Enterprise Server installer automatically sets the SMS boot order for the Eserver i5 and Eserver p5 systems, so the partition restarts from the correct device. The boot order for a JS20 blade server should be changed using the BladeCenter Management Module. 11.After the boot: prompt appears, the system starts automatically after several seconds. 12.When the installation window reopens, you are prompted to create a password for root. Enter the password for the root user twice, as prompted. Click Next. 13.In the Network Configuration window, select Change. 14.Select Network Interfaces. 15.In the Network cards configuration window, select Change under Ethernet card 0. 16.Select Edit. 17.Select and enter the settings appropriate for your environment, such as IP address, host name and name server, and routing. Click Next. 18.In the Network cards configuration overview window, click Finish. Click Next.

Chapter 19. IBM POWER technology hardware platform specifics

281

19.In the Test Internet Connection window, select No, Skip This Test. Click Next. 20.In the Service Configuration window, make sure that the Common Name and the Server Name fields have the correct values for your environment. Click Next. 21.For User Authentication Method, select Local (/etc/passwd). Click Next. 22.For LDAP Client Configuration, click Next. 23.In the Add a New LDAP User window, if you want to create a new user, do so here. Otherwise, do not create a new user. Click Next, and select Yes if you are warned about Empty user login. 24.After you have read the Release Notes window, click Next. 25.In the Hardware Configuration window, click Next. 26.When you see the Installation Completed window, click Finish. 27.After all of the settings are configured, the system exits the installation procedure and a login prompt is displayed.

Additional SLES install boot options

A partial list of options that can be added to the boot prompt (boot:) command line before beginning installation include: Install: URL (nfs, ftp, hd, and so on) Specifies the installation source as a URL. Possible protocols include cd, hd, nfs, smb, ftp, http, and tftp. The URL syntax corresponds to the common form used in Web browsers, for example: nfs://<server>/<directory> ftp://[user[:password]@]<server>/<directory> Netdevice: <eth0> The Netdevice: keyword specifies the interface that linuxrc should use, if there are several Ethernet interfaces available. HostIP: <10.10.0.2> Specifies the IP address of the host. Gateway: <10.10.0.128> This specifies the gateway through which the installation server can be reached, if it is not located in the subnetwork of the host. Proxy: <10.10.0.1> The Proxy: keyword defines a proxy for the FTP and HTTP protocols.

282

Solaris to Linux Migration: A Guide for System Administrators

ProxyPort: <3128> This specifies the port used by the proxy, if it does not use the default port. Textmode: <0|1> This keyword enables starting YaST in text mode. VNC: <0|1> The VNC parameter controls the installation process through VNC, which makes the installation more convenient for hosts that do not have a graphical console. If enabled, the corresponding service is activated. Also see the VNCPassword keyword. VNCPassword: <password> This sets a password for a VNC installation to control access to the session. UseSSH: <0|1> This keyword enables access to linuxrc through SSH when performing the installation with YaST in text mode. SSHPassword: <password> This sets the password for the user root to access linuxrc. Insmod: <module parameters> This specifies a module the kernel should load and any parameters needed for it. Module parameters must be separated by spaces. AddSwap: <0|3|/dev/hda5> If set to 0, the system does not try to activate a swap partition. If set to a positive number, the partition corresponding to the number is activated as a swap partition. Alternatively, specify the full device name of a partition.

19.3.5 YaBoot OpenFirmware boot loader

The POWER technology-based platform uses the YaBoot boot loader, which is loaded by OpenFirmware after identifying the PReP boot partition on the disk. Only disks containing the PReP boot partition can be selected in SMS. YaBoot starts and scans all the primary disk partitions, searching for its own configuration file (usually /etc/yaboot.conf) and a kernel image to load. YaBoot unfortunately cannot scan software mirrored md devices or LVM logical volumes. It also contains a built-in knowledge of only some file systems: ext2 ext3 reiserfs fat

Chapter 19. IBM POWER technology hardware platform specifics

283

For information about configuring the YaBoot bootloader, refer to your distributions POWER technology-specific documentation and the official YaBoot Web site located at:
http://penguinppc.org/bootloaders/yaboot/

19.4 eServer i5 and eServer p5 virtualization

The IBM POWER5 technology-based processor provides some unique virtualization capabilities that you can use in a Linux-only solution. Some of the capabilities that we mention in this section include: Shared processing Virtual SCSI Virtual Ethernet Virtual console Shared processing allows logical partitions to share the processors in the shared processor pool. The shared processor pool includes all processors on the server that are not dedicated to specific logical partitions. Each logical partition that uses the shared processor pool is assigned a specific amount of processor power from the shared processor pool. Virtual SCSI consists of a virtual SCSI server and multiple virtual SCSI clients. The virtual SCSI server is configured to serve block devices to corresponding virtual SCSI clients. All of the configuration is done on the virtual SCSI server side, while the virtual SCSI client detects the devices as regular SCSI devices when the virtual SCSI client kernel module is loaded. The virtual SCSI client can also rescan the virtual SCSI bus if new SCSI devices have been configured on the virtual SCSI server. Virtual Ethernet enables you to create a separate virtual local area network (VLAN) that can be shared among multiple partitions. Virtual Ethernet adapters are created on each partition that is to be part of the VLAN. One adapter can also be configured as a trunk adapter, which can then be bridged to a physical adapter in the partition. This allows a route to an external network for other clients connected to the VLAN. Virtual consoles consist of virtual console servers and their corresponding virtual console clients. Each Linux partition contains two virtual console servers by default, with the HMC providing the default console client for each partition. However, a Linux virtual console client can also connect to this virtual console server using the /dev/hvcs device. Typically, one partition contains all of the virtual console clients for the remaining client partitions in a system.

284

Solaris to Linux Migration: A Guide for System Administrators

Table 19-2 shows the supported virtual functions found in both Eserver-supported Linux distributions and the IBM VIO Server.
Table 19-2 Supported virtual I/O functions Virtual I/O function Shared processing Virtual SCSI server Virtual SCSI client Virtual Ethernet server Virtual Ethernet client Virtual console server Virtual console client RHEL 4 Yes No Yes Yes Yes Yes Yes SLES 9 Yes Yes Yes Yes Yes Yes Yes IBM VIO Server Yes Yes N/A Yes N/A Yes N/A

Although IBM VIO Server is based on IBM AIX 5L, it is quite different. You only use the VIO Server commands; some standard AIX 5L features are not present, and some typical AIX 5L setups are not allowed, such as Logical Volume Manager (LVM)-level mirroring of VIO client logical volumes. For in-depth information about this and POWER virtualization, refer to the IBM Redbook Advanced POWER Virtualization on IBM System p5, SG24-7940. The following topics describe how to create profiles for a virtual I/O server and client partition and configure them in Linux: Create a virtual I/O server partition profile Create a client partition profile Configure Linux on a virtual I/O server Configure Linux on a virtual client partition

19.4.1 Create a virtual I/O server partition profile

The virtual I/O server is the most important partition you will create because it provides the virtual adapters to the other client partitions. After the system is connected to the HMC, you can proceed with partitioning it. This section covers the process of creating the virtual I/O server partition using an HMC. 1. To begin, right-click Partitions under the system name (from the HMC) and select Create Logical Partition to start the Create Logical Partition Wizard. 2. Select Virtual I/O server and assign the partition the name, as shown in the example in Figure 19-15 on page 286.

Chapter 19. IBM POWER technology hardware platform specifics

285

Figure 19-15 Create Logical Partition Wizard

3. Take the default (No) for the Workload Management Groups function. 4. Assign a name to the profile. 5. Select the amount of memory you want to assign to the partition. 6. Select at least one Dedicated processor. 7. Assign physical I/O devices, such as CD/DVD drives and physical network adapters, to the partition. 8. Click Next in the I/O Pools window, because this function is not supported in Linux. 9. Select Yes, I want to specify virtual I/O adapters, as shown in Figure 19-16 on page 287.

286

Solaris to Linux Migration: A Guide for System Administrators

Figure 19-16 Create Logical Partition Profile - Virtual I/O Adapters

10.The next window enables you to create all of the virtual I/O adapters you will use in the virtual I/O server. See Figure 19-17 for an example.

Figure 19-17 Create Logical Partition Profile - Create Virtual I/O Adapters

Chapter 19. IBM POWER technology hardware platform specifics

287

 For the virtual Ethernet adapter, select Trunk adapter, as shown in Figure 19-18, and then click OK to return to the previous menu. In order to bridge two networks, one adapter in the virtual network must be designated as the trunk adapter. The trunk adapter manages all frames with unknown MAC addresses and routes them to the physical network. This virtual Ethernet adapter will be bridged with the VIO servers physical Ethernet adapter to an external network.

Figure 19-18 Virtual Ethernet Adapter Properties

288

Solaris to Linux Migration: A Guide for System Administrators

 For the virtual SCSI, leave Server selected for the Adapter Type (as shown in Figure 19-19) because this partition provides the disk or block devices to the client partitions. You can also leave Any remote partition and slot can connect selected because you do not know which remote partitions this slot should map to yet. After you create the client partitions, you can come back and change this to the correct client partition and slot number.

Figure 19-19 Virtual SCSI Adapter Properties

Click OK to create the adapter and return to the Create Virtual I/O Adapter window. Serial adapters are not created until the client partitions are created later.

Chapter 19. IBM POWER technology hardware platform specifics

289

11.Figure 19-20 is an example of what you should see after creating your virtual SCSI and Ethernet adapters.

Figure 19-20 Example of created virtual I/O adapters

12.Click Next in the Create Virtual Adapters window. 13.Click Next in the Power Controlling Partition window, because this function is not supported in Linux.

290

Solaris to Linux Migration: A Guide for System Administrators

14.Select desired optional settings and Normal under the Boot modes section of the Optional Settings menu. Click Next. 15.Click Finish in the Profile Summary window.

19.4.2 Create a client partition profile

Creating client partitions is similar to creating the virtual I/O server, except the client partitions only use virtual I/O adapters. This section covers the process of creating a client partition using an HMC. Perform the following steps: 1. To begin, right-click Partitions under the system name (from the HMC) and select Create Logical Partition to start the Create Logical Partition Wizard. 2. Select AIX or Linux and assign the partition the name. 3. Select the amount of memory you want to assign to the partition. 4. Unlike the in the virtual I/O server partition, assign shared processors to the client processors. Make sure to select Shared from the Processors menu. Then, click Next. 5. The remaining processors need to be divided among the client partitions. The processing unit settings shown in Figure 19-21 illustrate an example of a client partition. Refer to the IBM Redbook Advanced POWER Virtualization on IBM System p5, SG24-7940, for more information about shared processors.

Figure 19-21 Example of client profile processor settings

Chapter 19. IBM POWER technology hardware platform specifics

291

Click Next after configuring the processor settings. 6. Usually, there are no physical devices to assign because the client partitions only use virtual I/O. Click Next in the I/O window to continue. 7. Select Yes, I want to specify virtual I/O adapters, as shown in Figure 19-16 on page 287. 8. The next window enables you to create all of the virtual I/O adapters you will use in the client. See Figure 19-17 on page 287 for an example of this window. For the virtual Ethernet adapter, you can leave all the defaults, as shown in Figure 19-22. This is the virtual LAN that will be bridged to the physical network.

Figure 19-22 Client Virtual Ethernet Adapter Properties

Click OK to create the adapter and return to the Create Virtual I/O Adapter window.

292

Solaris to Linux Migration: A Guide for System Administrators

 For the virtual SCSI, select Client for the Adapter Type (as shown in Figure 19-23).

Figure 19-23 Client virtual SCSI Adapter Properties

Map the remote partition and slot number to the virtual SCSI server adapter slot you created for the virtual I/O server created earlier. For example, the first virtual SCSI adapter created for the virtual I/O server in 19.4.1, Create a virtual I/O server partition profile on page 285 used slot 4. So enter 4 in the Remote partition virtual slot number field for this partition. Then, click OK. Serial adapters are not created for client partitions. 9. Click Next in the Create Virtual Adapters window. 10.Click Next in the Power Controlling Partition window, because this function is not supported in Linux. 11.Select desired optional settings and Normal under the Boot modes section of the Optional Settings menu. Click Next. 12.Click Finish in the Profile Summary.

Chapter 19. IBM POWER technology hardware platform specifics

293

13.Return to the virtual I/O server profile to create the virtual serial client adapter. Start by right-clicking the default profile under the virtual I/O server partition and selecting Properties. 14.Go to the Virtual I/O tab and under Create adapters, select Serial and click Create. This launches a separate window to configure the properties of the virtual serial adapter. 15.Leave the type as Client and assign the Remote Partition to the corresponding client partition. Enter 0 for the Remote partition virtual slot number, as shown in Figure 19-24.

Figure 19-24 Virtual I/O server Virtual Serial Adapter Properties

16.You can now go through each of the virtual server SCSI adapters you created earlier and assign the correct remote partition to them. For each of the virtual Server SCSI adapters, select Only selected remote partition and slot can connect in the Connection Information section. 17.Mapping the slot numbers is a little trickier. You must assign the correct client partition and slot to its corresponding SCSI server slot. Figure 19-25 on page 295 shows an example of the mapping between a SCSI server adapter and an associated SCSI client adapter.

294

Solaris to Linux Migration: A Guide for System Administrators

Figure 19-25 Virtual SCSI client/server mapping example

The client and server configuration is now complete. The next steps involve installing and configuring Linux to recognize and use the virtual devices on the server and client.

19.4.3 Configure Linux on a virtual I/O server

In this section, you learn how to set up the services and devices the virtual I/O server will provide. First, you must install SLES9 SP1 or later on the virtual I/O server partition. SLES9 SP1 and later is currently the only IBM-supported

Chapter 19. IBM POWER technology hardware platform specifics

295

distribution that contains the virtual SCSI server driver. After installing SLES9 SP1 or later, install the bridge-utils and ckermit packages found in your SLES installation media.

Configure the Ethernet

This section describes the steps for configuring the physical and virtual Ethernet devices and the process involved in bridging theses devices together on the virtual I/O server. This exercise uses a dual-port Gigabit Ethernet adapter. In this example, one of the ports is configured to connect to the external network, while the second port is bridged to the virtual network and provides a route to the external network. Figure 19-26 is a graphical representation of what we are configuring.

Figure 19-26 Diagram of virtual LAN configuration

Ethernet device configuration

We recommend using YaST to configure the Ethernet devices: Configure the first physical device attached to the external network according to your LAN configuration. Do not configure the second physical or virtual Ethernet device.

Physical and virtual network bridging

Now that all of the Ethernet devices are configured, it is time to set up the bridging between the physical interface and the virtual interface. You do this by using tools found in the bridge-utils package and network scripts associated with the physical interface. Using the ifconfig command and the MAC address (from when you configured the interface) of the physical interface, you can determine

296

Solaris to Linux Migration: A Guide for System Administrators

which interface is your second physical interface. For this example, it happens to be eth2. You can also use the same method to determine the first virtual Ethernet interface, which is eth0 for this exercise. Example 19-1 shows both of these examples.
Example 19-1 ifconfig output eth2 Link encap:Ethernet HWaddr 00:0D:60:DE:01:A9 inet6 addr: fe80::20d:60ff:fede:1a9/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:794359 errors:0 dropped:0 overruns:0 frame:0 TX packets:16 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:56965484 (54.3 Mb) TX bytes:1124 (1.0 Kb) Base address:0xec00 Memory:b8100000-b8120000 Link encap:Ethernet HWaddr 2A:34:F0:00:10:02 inet6 addr: fe80::2834:f0ff:fe00:1002/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:794177 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 b) TX bytes:53775670 (51.2 Mb) Interrupt:184

eth0

Notice how neither interface has an IP address defined. Next, use the bridging tools available in the bridge-utils package to create a couple scripts that configure the bridge when eth2 is activated and take down the bridge when eth2 is deactivated. The network configuration files for SLES9 are located in /etc/sysconfig/network and are labeled ifcfg-eth-id-<MAC_ADDRESS>. Using the MAC address of eth2, you can identify which configuration file needs to be edited. For this exercise, the configuration file is: /etc/sysconfig/network/ifcfg-eth-id-00:0d:60:de:01:a9 The variable POST_UP points to a script that will run after the device is activated, while POST_DOWN points to a script that will run after the device has been deactivated. The two scripts are bridge-up.sh and bridge-down.sh and can be stored in /usr/local/bin or /etc/sysconfig/network/. Set these variables in your Ethernet configuration file. You should have entries that look similar to these in your configuration file:
POST_UP_SCRIPT='/etc/sysconfig/network/bridge-up.sh' POST_DOWN_SCRIPT='/etc/sysconfig/network/bridge-down.sh'

Chapter 19. IBM POWER technology hardware platform specifics

297

Example 19-2 shows a sample bridge-up.sh script with comments describing each line.
Example 19-2 Sample bridge-up.sh script #!/bin/sh # the ethX device that called this script CFG=$1 # the ethX device that called this script # source the eth config script to set some variables . /etc/sysconfig/network/ifcfg-$CFG # create the ethernet bridge /sbin/brctl addbr br0 # reset addresses for virtual ethernet (eth0) /sbin/ifconfig eth0 0.0.0.0 # reset address for physical etherent (eth2) /sbin/ifconfig eth2 0.0.0.0 # add the virtual ethernet to the bridge /sbin/brctl addif br0 eth0 # add the physical ethernet to the bridge /sbin/brctl addif br0 eth2 # bring up the bridge interface with the IP # address our physical interface was using /sbin/ifconfig br0 $IPADDR netmask $NETMASK

Example 19-3 shows a sample bridge-down.sh script with comments describing each line.
Example 19-3 Sample bridge-down.sh script #!/bin/sh # bring down virtual ethernet device eth0 ifconfig eth0 down # bring down physical ethernet device eth2 ifconfig eth2 down # delete eth0 from bridge interface br0 brctl delif br0 eth0 # delete eth2 from bridge interface br0

298

Solaris to Linux Migration: A Guide for System Administrators

brctl delif br0 eth3 # bring down bridge interface br0 ifconfig br0 down # remove bridge instance brctl delbr br0

Using the techniques described here, you should now have all three Ethernet interfaces configured in the virtual I/O server partition (see Figure 19-26 on page 296).

Configure the virtual console

The virtual console is provided by the IBM Hypervisor Virtual Console kernel module. You can load this on the virtual I/O server partition with the command:
# modprobe hvcs

Add this command to /etc/rc.d/boot.local so that it executes on every boot. Looking at your kernel messages in /var/log/messages, you will see something similar to the following:
Initializing IBM hvcs (Hypervisor Virtual Console Server) Driver vio_register_driver: driver hvcs registering HVCS: Added vty-server@30000010. HVCS: driver module inserted.

Note: Slots are in hexadecimal in the kernel log. For example, 0x10 above is slot 16. Using these mappings, you can create console connections on the virtual I/O server partition to the corresponding client partition. For this tutorial, we use the kermit serial program, which is part of the ckermit software package. The following command creates a console connection to the client:
# kermit -l /dev/hvcs0 -c

The device /dev/hvsc0 is the first device and maps to the first vty-server. If additional clients were configured, they would map accordingly. After you run the kermit command, you should see a screen similar to the following:
Connecting to /dev/hvcs0, speed 38400 Escape character: Ctrl-\ (ASCII 28, FS): enabled Type the escape character followed by C to get back, or followed by ? to see other options. ----------------------------------------------------

Chapter 19. IBM POWER technology hardware platform specifics

299

Configure the virtual SCSI server

Now that all the network services are up and running, the last part of setting up your virtual I/O server is configuring the SCSI disks that will be served to the client partition. The IBM virtual SCSI server module is called ibmvscsis. Add this module to /etc/rc.d/boot.local with the following line:
# modprobe ibmvscsis

This causes the module to load automatically on boot. The virtual SCSI server supports 64 target devices per bus and eight buses per virtual adapter. It defaults to one target and one bus per virtual adapter when the IBM virtual SCSI module is loaded. The virtual SCSI adapter supports multiple block devices, including an entire SCSI disk, a partition of a SCSI disk, and a shared file mounted over a loopback device. Modify a virtual SCSI adapter's settings through the sysfs file system (see /sys on page 90). The virtual SCSI adapters are located under /sys/devices/vio and all start with 300000xx, where xx is the slot number of the SCSI adapter. If the virtual I/O server partition has slot 4 configured as its first virtual SCSI adapter. It is represented in the sysfs file system under /sys/devices/vio/30000004/. To modify the number of buses this adapter supports, you must modify the setting /sys/devices/vio/30000004/num_buses. To modify the number of targets a certain bus supports, bus0 in this example, you modify /sys/devices/vio/30000004/bus0/num_targets. Here is an example of how to configure one adapter: 1. Set the number of targets to 1 for this bus:
# echo 1 > /sys/devices/vio/30000004/bus0/num_targets

2. Assign the block device /dev/sdb to this adapter:

# echo /dev/sdb > /sys/devices/vio/30000004/bus0/target0/device # echo b > /sys/devices/vio/30000004/bus0/target0/type

3. Do not enable read-only connect mode:

# echo 0 > /sys/devices/vio/30000004/bus0/target0/ro

4. Activate the device:

# echo 1 > /sys/devices/vio/30000004/bus0/target0/active

You can add a function to /etc/rc.d/boot.local to make setting up the virtual SCSI adapters easier, as shown in Example 19-4 on page 301.

300

Solaris to Linux Migration: A Guide for System Administrators

Example 19-4 Virtual SCSI function in boot.local #! /bin/sh # # Copyright (c) 2002 SuSE Linux AG Nuernberg, Germany. All rights reserved. # # Author: Werner Fink <[email protected]>, 1996 # Burchard Steinbild, 1996 # # /etc/init.d/boot.local # # script with local commands to be executed from init on system startup # # Here you should add things, that should happen directly after booting # before we're going to the first run level. # # load the Hypervisor Virtual Console driver modprobe hvcs # load the IBM Virtual SCSI server driver modprobe ibmvscsis function vscsisetup { if [ $# -ne 2 ]; then echo "vscsisetup: slot_number physical_device" exit 1 fi vslot=$1 physical_device=$2 if [ ! -e /sys/devices/vio/$vslot ]; then echo "vscsisetup: Error - Slot number $vslot does not exist" exit 1 fi if [ ! -e $physical_device ]; then echo "vscsisetup: Error - physical device $physical_device does not exist" exit 1 fi echo 1 > /sys/devices/vio/$vslot/bus0/num_targets echo $physical_device > /sys/devices/vio/$vslot/bus0/target0/device echo b > /sys/devices/vio/$vslot/bus0/target0/type echo 0 > /sys/devices/vio/$vslot/bus0/target0/ro echo 1 > /sys/devices/vio/$vslot/bus0/target0/active rc=`cat /sys/devices/vio/$vslot/bus0/target0/active` if [ $rc -ne '1' ]; then echo "IBMVSCSI Server configuration failed for device $vslot" fi; }

Chapter 19. IBM POWER technology hardware platform specifics

301

# assign block devices to virtual SCSI adapters # setup server slot 4 with /dev/sdb assigned to client partition vscsisetup 30000004 /dev/sdb

The virtual I/O server should be configured at this point. For this tutorial about virtualization, we do a CD installation on the client. In order to accomplish this, we must relocate the CD/DVD device to the client partition for installation. For additional installation methods, refer to Linux virtualization on POWER5: A hands-on setup guide on IBM developerWorks:
http://www.ibm.com/developerworks/edu/l-dw-linux-pow-virutal.html

Perform the following steps: 1. Power off the virtual I/O server partition. 2. Remove the CD/DVD device from the virtual I/O server partition profile. Refer to Hosted (partitioned) on page 262 for details. 3. Add the CD/DVD device to the client partition profile. 4. Power on the virtual I/O server partition and ensure that all the services and virtual functions are set up properly.

19.4.4 Configure Linux on a virtual client partition

This section describes the steps involved in installing Linux on a client partition using CDs. Perform the following steps: 1. Insert the first CD from the RHEL4 installation media into your Eserver i5 or Eserver p5 system. 2. Activate the client partition through the HMC. 3. Open a terminal to the client partition. 4. Use the SMS menus to boot from the CD. Refer to Hosted (partitioned) on page 262 for details. Both supported Linux distributions will automatically load virtual SCSI client and virtual Ethernet drivers before beginning the installation. You can now proceed to install the partition without any special virtualization related steps. Refer to 19.3.3, Red Hat on page 279, for information about installing RHEL4 and 19.3.4, Novell SUSE on page 280, for information about installing SLES9 on an Eserver i5 or Eserver p5 partition.

302

Solaris to Linux Migration: A Guide for System Administrators

19.5 POWER technology platform service and productivity tools

There are various hardware service diagnostic aids and productivity tools available for Eserver i5 and Eserver p5 and JS20 blade systems running Linux. These service and productivity tools perform the following functions: Base tools Access POWER technology-based platform features that are provided by the system's firmware Gather hardware inventory and microcode level Access system boot lists, LEDs, reboot policies, and so on from the operating system command line Communicate with an attached Hardware Management Console Service tools Analyze errors or events and perform actions to increase system availability or protect data integrity Communicate errors to an attached Hardware Management Console or to IBM Service Survey hardware and communicate the results to the IBM Machine Reported Product Data database Productivity tools Hot-plug add, remove, or replace PCI devices Dynamically add or remove processors or I/O slots from a running partition using an attached Hardware Management Console Table 19-3 describes some common tool packages that are available for Linux on POWER installations.
Table 19-3 Common packages available for Linux on POWER platforms Package librtas ppc64-utils lsvpd Description Contains a library that allows applications to access certain functionality provided by platform firmware. Enables a number of reliability, availability, and serviceability (RAS) features. Contains the lsvpd, lscfg, and lsmcode commands. These commands, along with a boot-time scanning script called update-lsvpd-db, constitute a hardware inventory system. Provides automatic analysis and notification of errors reported by the platform firmware.

diagela

Chapter 19. IBM POWER technology hardware platform specifics

303

19.5.1 Red Hat

All tools for RHEL4 on POWER technology-based installations are at the following Web sites: JS20 blade
http://techsupport.services.ibm.com/server/lopdiags/redhat/blades

Monolithic Eserver installation

http://techsupport.services.ibm.com/server/lopdiags/redhat/otherserver

Hosted Eserver installation

http://techsupport.services.ibm.com/server/lopdiags/redhat/hmcserver

19.5.2 Novell SUSE

All tools for RHEL4 on POWER installations are at the following Web sites: JS20 blade
http://techsupport.services.ibm.com/server/lopdiags/suselinux/blades

Monolithic Eserver installation

http://techsupport.services.ibm.com/server/lopdiags/suselinux/otherserver

Hosted Eserver installation

http://techsupport.services.ibm.com/server/lopdiags/suselinux/hmcserver

19.6 References and further readings

The following is a list of references for topics covered in this chapter. IBM Redbooks
http://www.ibm.com/redbooks

Linux on IBM Eserver i5 Implementation Guide, SG24-6388 The IBM Eserver BladeCenter JS20, SG24-6342 Advanced POWER Virtualization on IBM System p5, SG24-7940 IBM Information Center publications
http://publib.boulder.ibm.com/infocenter/eserver/v1r3s/index.jsp

Managing the Advanced System Management Interface (ASMI) Managing the Hardware Management Console (HMC) Managing your server using the Hardware Management Console

304

Solaris to Linux Migration: A Guide for System Administrators

 Partitioning for Linux Installing Linux Distribution guides Red Hat:Installation Guide for the IBM POWER Architecture
http://www.redhat.com/docs/manuals/enterprise/

Novel SUSE: SUSE Linux Enterprise Server 9 Architecture-Specific Information for IBM POWER
http://www.novell.com/documentation/sles9/index.html

IBM developerWorks
https://www.software.ibm.com/developerworks

Linux virtualization on POWER5: A hands-on setup guide

Chapter 19. IBM POWER technology hardware platform specifics

305

306

Solaris to Linux Migration: A Guide for System Administrators

20

Chapter 20.

IBM eServer zSeries and IBM System z hardware platform specifics

Linux, when run on IBM Eserver zSeries, IBM System z, or the IBM S/390 hardware platform, has a set of considerations above and beyond what needs to be considered for the Eserver xSeries and pSeries platforms. The S/390 and zSeries platforms have different capabilities that must also be considered. We use the term mainframe to refer to both platforms in this section unless a specific difference or capability must be identified. The majority of Linux management concepts, tools, and techniques are the same on all hardware platforms. This chapter describes resources, concepts, tools, and techniques specific to Linux and IBM mainframes. In this chapter, we discuss the following topics: Planning S/390 and zSeries overview Installation methods and techniques Booting, system initialization, and shutdown Device management

Copyright IBM Corp. 2006. All rights reserved.

307

Performance monitoring and tuning Troubleshooting and diagnostics S/390 and zSeries-specific Linux commands High availability Security References

308

Solaris to Linux Migration: A Guide for System Administrators

20.1 Planning
As with any server, you must plan an installation. You must define the resources your applications need to function optimally. These resources, referred to as a footprint, should be the starting point of your planning. With this information, you can size and configure your server requirements appropriately. You will need several resources to properly size, deploy, and tune Linux on a mainframe. We cover these at a high level in this section and refer you to other Web sites or documents for more in-depth information. The information in this section will help you prepare for a Solution Assurance Review (SAR). All Linux on mainframes deployments should have a SAR at the beginning of the project. Ask your IBM zSeries representative to host a SAR for you. This will be a question and answer session where your team and IBM can discuss the behavior of your applications, your requirements, and expectations, and help identify what hardware resources you will need to be successful in your deployment. Running Linux under z/VM in an LPAR is considered the best practice. Virtual Machine (VM) has long been recognized as a robust computing platform, spanning the family of S/390 and zSeries servers. z/VM provides a highly flexible test and production environment for enterprises deploying the latest e-business solutions. z/VM supports Linux, one of the world's leading open source operating systems. Within the VM environment, Linux images benefit from the ability to share hardware and software resources and use internal high-speed communications. z/VM offers an ideal platform for consolidating workloads on a single physical server, which enables you to run tens to hundreds of Linux images. Although you can run Linux in an LPAR without z/VM, this chapter assumes that you are considering hosting Linux images with z/VM. Refer to the following brief list of Internet references for Linux on mainframes: Linux for S/390 project This Web site contains useful links to IBM Redbooks, other related Web sites, how-tos, man pages, and more:
http://www.Linuxvm.org/

z/VM publications This Web site provides links to documents for specific z/VM releases and a link to another Linux library that has articles discussing specific workloads:
http://www.vm.ibm.com/pubs/

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

309

IBM Eserver zSeries and IBM System z This Web site provides links to documents for specific zSeries models and a link to another Linux library that has articles discussing solutions, success stories, and technical support:
http://www.ibm.com/servers/eserver/zseries/

z/VM overview This Web site provides an overview of the features available in current releases of z/VM:
http://www.vm.ibm.com/overview

A set of planning worksheets is available in IBM Redbook z/VM and Linux on zSeries: From LPAR to Virtual Servers in Two Days, SG24-6695.

20.1.1 Hardware resources

You need the following hardware components to deploy Linux on a mainframe. A more detailed description of these components follows in subsequent sections. zSeries logical partition (LPAR). Processors or CPUs: One Integrated Facility for Linux (IFL) minimum; two or more is recommended. Memory: Central and expanded at a 3:1 ratio. Direct Access Storage Devices (DASD, or disk drives): It is helpful to have DASD on different channel-path identifiers (CHPIDs). Open Systems Adapter (OSA) network cards: OSA-Express in Queued Direct I/O (QDIO) recommended. A computer that acts as a Network File System (NFS) or File Transfer Protocol (FTP) server.

20.1.2 Software resources

You need the following software and documentation to deploy Linux on a mainframe: z/VM installation media and documentation Linux installation media and documentation A workstation or desktop that has network access to the mainframe: A 3270 emulator: Attachmate Extra, Hummingbird Host Explorer or IBM Personal Communications for Windows desktops

310

Solaris to Linux Migration: A Guide for System Administrators

A 3270 emulator named x3270 (included for free in most distributions) for Linux desktops A Linux Secure Shell (SSH) client such as PuTTY (recommended) or TeraTerm for Windows desktops For Linux desktops, the SSH client, which is built-in

A remote login program:

20.1.3 Networking resources

You need the following TCP/IP resources: A TCP/IP address for z/VM One TCP/IP address for each Linux virtual server Associated TCP/IP information: Domain Name System (DNS) host name DNS domain DNS server TCP/IP address TCP/IP gateway TCP/IP subnet mask TCP/IP broadcast address (usually calculated from address and subnet mask) TCP/IP MTU size

20.1.4 Linux distributions and z/VM

We discuss two distributions throughout this section: Red Hat Enterprise Linux Advanced Server V4 and Novell SUSE Linux Enterprise Server V9. Each distribution has its own configuration limits. You can find this information at the following Web sites:
http://www.redhat.com/en_us/USA/rhel/details/limits/ http://www.novell.com/products/Linuxenterpriseserver/kernel_limits.html

Linux can run on the bare hardware of the mainframe without any special software or support required. The actual internals of the Linux OS are identical from platform to platform. The only changes that are made in the mainframe implementation of Linux relate to the instructions that the operating system passes to the hardware. As mentioned earlier, running Linux under z/VM in an LPAR is considered the best practice.

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

311

20.2 S/390 and zSeries overview

IBM Eserver zSeries hardware is designed for continuous availability, so zSeries systems offer a set of reliability, availability, and serviceability features (RAS). The zSeries architecture has two instruction units to compare every operation in order to discover possible errors. The main storage and the L2 cache uses Error Checking and Correction (ECC) technology. In addition, memory sparing is implemented to replaces failed memory units. CPU sparing is also available. A detailed description of the zSeries hardware can be found in the following IBM Redbooks: IBM Eserver zSeries 900 Technical Guide, SG24-5975 IBM Eserver zSeries 890 Technical Introduction, SG24-6310 Technical Introduction: IBM Eserver zSeries 800, SG24-6515 IBM Eserver zSeries 990 Technical Guide, SG24-6947 IBM System z9 109 Technical Guide, SG24-7124

20.2.1 Processing units

The multichip module (MCM) contains multiple processing units (PUs). On other platforms, these are commonly referred to as central processing units (CPUs). All processing units (PUs) are physically the same. At power-on reset (POR), distinct microcode that is loaded in the PU determines its processor type. The processor types pertinent to z/VM and Linux are: Central processor (CP) CPs are enabled to execute any operating system available on zSeries. Integrated Facility for Linux (IFL) IFLs are processors enabled specifically for Linux workloads. Both z/VM and Linux (either in LPAR mode, or as a z/VM) can execute on an IFL. Three other types of processors are available on zSeries platforms: System Assist Processor (SAP), Internal Coupling Facility (ICF) and zSeries Application Assist Processor (ZAAP). Linux does not use these processor types.

20.2.2 Memory
There are three different address space sizes for Linux distributions. 64 GB: 64-bit word lengths are available on Intel, AMD, POWER, and zSeries processors. A 64-bit word length, generally speaking, provides applications with a 64 GB address space with current technology.

312

Solaris to Linux Migration: A Guide for System Administrators

4 GB: 32-bit word lengths are available on Intel and earlier POWER or RISC processors. A 32-bit word length provides applications with a 4 GB address space. 2 GB: 31-bit address spaces are available on S/390 and zSeries processors. One bit of the 32-bit word length is reserved as a flag for the older 24-bit address space. This shorter 31-bit word length provides applications with a 2 GB address space (approximately). IBM S/390 only supports 31-bit address spaces. zSeries can run both 31-bit and 64-bit Linux guests. Because of the more limited address space in the 31-bit environment, we recommend that you deploy 64-bit Linux on zSeries. 31-bit applications are fully supported on 64-bit Linux on zSeries.

Central and expanded storage

IBM mainframes have two types of memory, referred to as storage: Central storage Central storage is the real memory that the operating systems and applications can access. Expanded storage Expanded storage is memory set aside for z/VM paging (swapping). In addition, virtual disks (VDISKs) can be created in the expanded storage. This reduces the amount of memory available for z/VM paging. The VDISK is analogous to a RAM disk on other hardware platforms.

20.2.3 The channel subsystem

I/O devices are attached to the zSeries system through the channel subsystem (CSS). Components of the channel subsystem include: Control unit A control unit (CU) provides the logic to operate and control an I/O device. An I/O device can be attached to one or more control units. Control unit function can be: Implemented in a separate physical unit Integrated inside the I/O device Integrated within the channel itself Channel A channel is a specialized processor that communicates with the control unit. Channels are located in the zSeries I/O cage and manage data movement between central storage and the control unit.

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

313

Subchannel One subchannel is dedicated to each I/O device accessed by the channel subsystem. The subchannel provides information about the attached I/O device to the channel subsystem (such as the CHPID, device status, and channel path availability). Subchannels are addressed using the system-unique 16-bit subchannel number. The actual number of available subchannels depends on the system model, but is limited to a maximum of 65,536 per system. Channel path A control unit is attached to the channel subsystem by one or more channel paths. Depending on the zSeries model and configuration, an I/O device can be accessed by as many as eight different channel paths. Types of channel paths supported on zSeries include: Enterprise Systems Connection (ESCON) Fibre Connection (FICON) Open Systems Adapter-2 (OSA-2) OSA-Express Channel paths are addressed using the system-unique 8-bit channel-path identifier (CHPID). The actual number of available channel paths depends on the system model, but is limited to a maximum of 256 per system. System assist processor (SAP) The SAP schedules an I/O operation. It finds an available channel path to the intended device and guarantees completion of the I/O operation. However, it is not involved in moving data between central storage and the channel.

ESCON
ESCON provides bidirectional serial bit transmission, in which the communication protocol is implemented through sequences of special control characters and through formatted frames of characters. The ESCON link data rate is 200 megabits per second (Mbps), which results in a maximum aggregate data rate of up to 17 megabytes per second (MBps). The maximum unrepeated distance of an ESCON link using multimode fiber optic cables is 3 km (1.87 miles) when using 62.5 micron fiber.

FICON
FICON provides all the strengths of ESCON, plus more. The link data rate has increased from 200 Mbps to 2 Gbps, with expected effective data transfer rates between 60 to more than 170 MBps. The FICON implementation enables full duplex data transfer, so data travels in both directions simultaneously, rather than

314

Solaris to Linux Migration: A Guide for System Administrators

the ESCON half-duplex data transfer. In addition, multiple concurrent I/Os can occur on a single FICON channel, a fundamental difference between FICON and ESCON.

zSeries FCP support

The Fibre Channel Protocol (FCP) feature is provided with zSeries servers equipped with a FICON card. A FICON card has two ports. Each port can serve only one LPAR to link to SCSI devices configured in an IBM TotalStorage Enterprise Storage Server (ESS) environment. Channel sharing is not implemented for this architecture. zSeries FCP support enables you to connect a Linux running on zSeries 800, zSeries 900, or zSeries 990 with industry-standard SCSI devices. This means that you are able to configure your Linux to use a Storage Area Network (SAN) environment.

20.2.4 Tape drives

The Linux installation program can be loaded from tape drives. The only requirement is that a tape subsystem must be attached to the mainframe and available to the Linux guest or LPAR. IBM supported devices such as the Enterprise Storage Server, Enterprise Tape System 3590, and Enterprise Tape Library 3494 are available. These devices can be attached through ESCON or FICON channels. Two types of storage controllers with Fibre Channel interfaces are currently available for zSeries: Those supporting the z/Architecture-specific FICON interface based on Channel Control Word (CCW) architecture, with the CKD command set for disk storage and similar sets of commands for tapes Those supporting the SCSI-based FCP protocol, with SCSI command sets specific to the device type, such as disk or tape Linux provides the mt command for control of tape devices.

20.2.5 Disk drives

Disk drives are attached to the mainframe through the channel subsystem. The protocols used to access the disk drives can be SCSI, ESCON, or Fibre Channel/FICON.

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

315

Disk drives are identified with the /dev/dasd prefix instead of /dev/sd (SCSI) or /dev/hd (IDE). The disk names are created in this order: /dev/dasda through /dev/dasdz, then /dev/dasdaa, and so on. Linux directories that contain disk drive information are /proc/scsi, proc/bus, and /proc/dasd. Disk drives can be accessed as full volumes or partitioned as z/VM minidisks. A special type of disk, the VDISK, is a non-persistent RAM disk that can be created in expanded storage. In many cases, the FICON implementation coexists with, and can reuse, a customers existing ESCON fiber cabling infrastructure. The FICON features on zSeries servers also support full fabric connectivity under the Linux operating system for the attachment of Small Computer System Interface (SCSI) devices using the Fibre Channel Protocol (FCP).

20.2.6 Network
Setting up real and virtual networking is described in Chapter 4 of Linux for IBM System z9 and zSeries, SG24-6694 and in Networking Overview for Linux on zSeries, REDP-3901. Red Hat and SUSE both provide specific instructions for identifying and configuring network connections.

HiperSockets
HiperSockets technology is available for IBM Eserver zSeries servers (z800, z890, z900, and z990) and IBM System z9 109. HiperSockets is an integrated any-to-any virtual TCP/IP network that provides interconnectivity between multiple LPAR or virtualized images under z/VM. With HiperSockets, you can connect any server running on the same zSeries and improve response time due to low latency. HiperSockets is a feature of the zSeries 800 and 900 processors. It provides a high-speed network connectivity method, which can be used to link operating systems running in logical partitions on a single zSeries processor. Note: Connectivity in a z/VM virtual network is limited to guests running in a single z/VM image. Virtual networks cannot be used for inter-LPAR communication.

Virtual networks
Three types of virtual networks are available to Linux guests: Virtual channel-to-channel (VCTC): VCTC networks provide point-to-point connectivity between guests without real channel allocation (as required for real channel-to-channel connectivity).

316

Solaris to Linux Migration: A Guide for System Administrators

A real channel-to-channel (CTC) adapter is used to connect a real mainframe to another using the channel protocol. z/VM provides the ability to define virtual channel-to-channel adapters so that users can connect virtual machines using the CTCA protocol. On z/VM, this is useful for connecting Linux virtual machines to other virtual machines that do not support inter-user communication vehicle (IUCV), such as VSE/ESA, OS/390, and z/OS. Virtual VCTC networks can also connect Linux virtual machines. Inter-user communication vehicle (IUCV): Point-to-point TCP/IP connections between Linux guests can be established using IUCV. IUCV is a VM-unique, virtual machine-to-virtual machine communication protocol. The Linux for S/390 and zSeries kernels include an IUCV driver, enabling you to connect two Linux virtual machines. Linux treats IUCV connections as any other TCP/IP network connection. This results in memory-speed networking between the connected Linux servers. The speed of the IUCV line is a factor of processor speed. The faster your zSeries server is, the faster your IUCV network is. When you upgrade to a faster real processor, you automatically increase the speed of your virtual network. IUCV is also supported by the z/VM TCP/IP stack, so you can use IUCV to connect Linux images to the VM TCP/IP stack. z/VM guest LAN and virtual switch: This type of network enables local area network (LAN) connectivity between z/VM guests. With VM guest LAN and VSWITCH technology, you can connect a large number of Linux servers to a real network using fewer real network adapters, exploiting the virtual networking capabilities of z/VM to essentially share the real network connections among your Linux servers. IBM recommends that you to implement VM guest LAN over point-to-point for your network connectivity. It also recommends the VSWITCH architecture over router virtual machine.

TPC/IP direct connection

The network interface your installation uses will, to a large extent, determine whether you can support all your Linux guests using direct connection. For example, an OSA-2 interface can only support 16 entries in its OSA Address Table per port, which means that each port can be shared by only eight TCP/IP stacks. The OSA-Express interface provides a very high level of shareability, with up to 4096 IP addresses per port available. The easiest way to directly connect Linux guests to the network is by sharing a port (or ports) on an OSA-Express interface. Each stack sharing the OSA port is directly attached to the network with an individual LAN IP address. Under z/VM, you can dedicate the required sets of device addresses to your Linux guests, giving them direct access to the OSA hardware.

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

317

Important: If you want the ability to move an OSA CHPID from LPAR to LPAR for recovery purposes, defining the CHPID as shareable is not a good idea. Doing so reduces the maximum usable capacity of the OSA by at least half, because you need to provide the same definitions on both the production and backup LPARs (to provide access for all the Linux guests that you recover). Defining the CHPID as reconfigurable is better in this case. If you must share the OSA with another LPAR and also use it for recovery, defining the CHPID as shareable is the only solution. You need to take into account the reduced capacity when designing your Linux guest connectivity. The candidate list is the list of LPARs that can have access to the CHPID. LPARs not defined in the candidate list cannot configure the CHPID online. The OSA-Express IP Assist function offloads the following protocol functions from the IP stack for processing in the adapter: Broadcast filtering ARP processing Building MAC and LLC headers Offloaded processing can reduce the cycles consumed by your Linux guests for network traffic, giving a slight performance improvement. In a single guest, the effect might not be significant, but in a z/VM LPAR with Linux guests generating a moderate-to-high volume of network traffic, we would expect an overall saving.

Stack-to-stack routing
When IP stacks are sharing an OSA, and one stack sends traffic destined to another stack sharing the same OSA port, the OSA sends the traffic directly to the destination stack. Network traffic processed in this way does not appear on the LAN. This action takes place transparently to the sending stack. This can provide a performance benefit when a number of guests send network traffic to each other, because it will be sent directly between the stacks rather than out across the network. If you are considering an OSA-sharing configuration for your large-scale Linux installation, you need to be aware of some points regarding this connection method. An IBM Technical Flash (document number Flash 10144) describes some considerations to be observed when sharing an OSA-Express port defined in QDIO mode. It details the number of LPARs or guests that can share an OSA (depending on the hardware type) and some zSeries code-level dependencies. In summary, a z900 with up-to-date microcode, or any z800, can support up to 80

318

Solaris to Linux Migration: A Guide for System Administrators

IP stacks on an OSA-Express port in QDIO mode. Other combinations can support considerably fewer than that. What is not clear is the level of performance that can be expected when the OSA is configured to its peak. Performance testing to date has focussed on throughput from a single stack, rather than saturation across a fully-configured OSA. To access this Flash, enter the Flash number 10144 at the following site:
http://www.ibm.com/support/techdocs/

20.2.7 Printers
There are two specific considerations for creating print servers on Linux on zSeries. Printers attached to and directly managed by z/VM that you can connect locally with S/390 and zSeries hardware platforms are not supported by CUPS. On these platforms, network printing is required. For information about configuring network printers, see Printing with Linux on IBM Eserver zSeries Using CUPS and Samba, REDP-3864. In general, a root file system of a single 3390-3 does not leave sufficient disk space for a print server. There are many possible configurations for laying out file systems. One suggested configuration is three 3390-3s, one for each of / (root), /usr, and /var.

20.2.8 Logical partition concepts

The Processor Resource/System Manager (PR/SM) is a standard feature of all zSeries central processor complexes (CPCs). This allows a CPC to be divided into multiple logical partitions (LPARs). LPARs allow workloads to be isolated in different system images, so you can run production work separately from test work, or even consolidate multiple servers into a single processor. A LPAR has the following properties: Each LPAR is a set of physical resources controlled by an operating system image. Physical processors and channels can be shared between multiple LPARs, or dedicated to single LPAR. In general, sharing processors maximizes mainframe system utilization. Physical storage used by an LPAR is dedicated to the LPAR.

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

319

20.2.9 Virtualization
The z/VM Control Program controls the virtual machine environment. Control Program provides each z/VM user with a working environment referred to as a virtual machine. A virtual machine is the functional equivalent a real system. Virtual machines share real resources: processors, storage, console, and input/output devices. The real resources will be managed with a user directory provided by z/VM. There does not need to be an actual correspondence between physical and logical resources for virtualization to be used. A guest operating system such as Linux running in the virtual machine sees the virtual resources as real hardware. No modifications are required to the operating system to operate in a virtual machine; the mapping of these virtual resources to real resources is handled by the virtualization software. With z/VM, you can grow your server workload vertically or horizontally. Vertical growth is realized when you add more system resources such as additional processing power to an existing virtual machine. You can give a Linux virtual machine more processor capacity, more virtual memory, more I/O devices and bandwidth, plus more networking bandwidth. Horizontal growth is a typical way to grow workloads. It is easily accomplished by adding another virtual machine, or two, or three, or more on z/VM. There is an added benefit in the efficient use and sharing of system resources. This makes adding more Linux virtual servers a particularly attractive option when you consider the limited scalability of a single Linux instance. The added scalability of Linux workloads is realized when you choose to exploit data-in-memory and sharing techniques of z/VM. Virtual disks in storage and minidisk cache can boost the performance of Linux servers on z/VM by making disk I/O operations execute at processor memory speeds, avoiding trips to the I/O subsystem, and waiting on spinning disks of storage. By sharing data and programming on shared, read-only minidisks, server deployment time can be reduced, as well as the maintenance effort required to replicate changes to data and programming among multiple distributed servers.

20.3 Installation methods and techniques

There are three basic installation methods for Linux on mainframes. These methods are: Installation as a native operating system. No other LPARs can be defined. All resources are dedicated to Linux. Linux cannot take advantage of virtualization techniques. This is not available on all IBM mainframes.

320

Solaris to Linux Migration: A Guide for System Administrators

Installation into a logical partition (LPAR). Linux in an LPAR is typically used when the single Linux server can process a large, critical workload. For example, SAP and IBM Lotus Domino servers are often run in an LPAR. Installation into a z/VM guest. Linux shares the LPARs resources with other z/VM guests. Virtualization techniques are available to Linux.

20.3.1 Preparing the z/VM guest resources

As with any other hardware platform, any other operating system, and any other application, the basic resources are the processors, RAM, network, and disk devices required to complete the workload. Because an LPAR contains non-Linux and Linux guests, your planning should include all guests that will run in the LPAR. Linux has been available on mainframes since the 2.2 kernel. During that time, best practices have been identified. For best practices for preparing z/VM guest resources, refer to:
http://www.ibm.com/servers/eserver/zseries/library/whitepapers/pdf/gm130647.pdf http://www.ibm.com/developerworks/linux/linux390/perf/

20.3.2 Server farms or cloned environments

z/VM setup for cloned environments is covered in more detail in several other sources, such as the following IBM Redbooks and Redpaper: z/VM and Linux on zSeries: From LPAR to Virtual Servers in Two Days, SG24-6695 Linux on IBM Eserver zSeries and S/390: ISP/ASP Solutions, SG24-6299 Linux on IBM Eserver zSeries and S/390: Cloning Linux Images in z/VM, REDP-0301 We present a quick overview of the process in this section. The first step in setting up the Linux cloning environment is to decide which directories or file systems will be split into read/write and read-only mount points. One suggested method is to have one Linux guest mount all its file systems in read/write mode so that updates can be applied. The remaining Linux guests, or clones will: Mount directories, such as /opt, /usr, /bin, /sbin, and /lib, as read-only file systems. These file systems usually contain system utilities and executables that end users should not need to modify.

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

321

Mount directories, such as /home, /proc, /dev, /mnt, and /media, as read/write file systems. The root ( /, not to be mistaken for /root) directory should also be mounted in read/write mode. This particular approach incurs additional work for the system administrator. However, the benefit of mounting parts of the operating system in the read-only mode means that mistakes will impact a much smaller area of the system. On any given clone, a user can only damage the configuration of that users own applications.

Installation methods
The installation methods listed in Table 20-1 apply to both 31-bit and 64-bit versions of Red Hat and SUSE distributions. Both distributions recommend 512 MB RAM (minimum 256 MB) for the installation program.
Table 20-1 Installation techniques Installation technique Network installation into z/VM guest using ISO images on remote server Network installation into z/VM guest from remote directory Installation using tape LPAR installation using CD Red Hat Enterprise Linux NFS protocol is supported. SUSE Linux Enterprise Server NFS protocol is supported.

NFS, FTP, and HTTP protocols are supported.

NFS protocol is supported. Yes

Requires access to the Hardware Management Console (HMC) or Support Element Workplace (SEW). CD images must be accessed through FTP. Disk must be formatted ext2 or ext3. Yes Yes Yes Yes Yes

Installation from S/390 hard drive Graphical installation Text mode installation SSH Telnet VNC

Red Hat Enterprise Linux installation overview

Red Hat Enterprise Linux 4 Installation Guide for the IBM S/390 and IBM Eserver zSeries Architectures, rhel-ig-s390-multi-en.pdf, describes the mainframe installation. We describe the high-level sequence of events in this section.

322

Solaris to Linux Migration: A Guide for System Administrators

Perform the following steps: 1. First, you must log in to z/VM as the Linux guest account using a 3270 emulator program. Enter the Conversational Monitor System (CMS) mode with the following command:
ipl cms

The Red Hat Enterprise Linux installation requires the following files: kernel.img (the Linux kernel) initrd.img (the initial RAM disk) redhat.parm redhat.conf 2. Update the redhat.parm and redhat.conf files. 3. Load (punch) the files into the z/VM guests virtual reader queue and then boot (ipl) the guest using the files in the readers queue. The installation program starts. You will be notified when you can login through ssh, telnet, or vnc to continue with the installation.

Sample Red Hat Enterprise Linux installation files

This section contains sample redhat parm a, redhat conf a, and rhel4 exec a files used to initiate Red Hat installation. Example 20-1 presents a sample redhat parm a file, and Example 20-2 presents a sample redhat conf a file.
Example 20-1 Example installation flie: redhat parm a root=/dev/ram0 ro ip=off ramdisk_size=40000 CMSDASD=191 CMSCONFFILE=redhat.conf vnc Example 20-2 Example installation file: redhat conf a HOSTNAME="Linux100.redbook.com" DASD="200-202" NETTYPE="qeth" IPADDR="192.168.1.100" SUBCHANNELS="0.0.0300,0.0.0301,0.0.0302" PORTNAME="rhLinux" NETWORK="192.168.1.0" NETMASK="255.255.255.0" BROADCAST="192.168.1.255" SEARCHDNS="redbook.com" GATEWAY="192.168.1.1" DNS="192.168.1.254"

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

323

You can manually enter the commands to start the installation, but most system administrators find it easier to create a script with the necessary commands. The z/VM editor is invoked with the xedit command. The editor is closed and the file saved with the save subcommand. For example:
xedit rhel4 exec a

Example 20-3 shows the contents of the rhel4 exec a file in REXX format. To execute this file after you have saved it, type the rhel4 command at the CMS prompt and the installation will begin.
Example 20-3 Example installation file: rhel4 exec a, in REXX format /* */ 'close rdr' 'purge rdr all' 'spool punch * rdr' 'punch kernel img a (noh' 'punch redhat parm a (noh' 'punch initrd img a (noh' 'change rdr all keep nohold' 'ipl 00c clear'

Notice that the redhat conf a (redhat.conf) file is not punched to the reader. You only need to punch the images and the parm file to begin the installation. The installation program will then read the contents of the configuration file when it is required. From this point on, you should refer to rhel-ig-s390-multi-en.pdf, Red Hat Enterprise Linux 4 Installation Guide for the IBM S/390 and IBM eServer zSeries Architectures, for details about the specific type of installation (CD, ISO image, NFS) and access mode (SSH, Telnet, VNC) that you have chosen.

SUSE Linux Enterprise Server installation overview

SUSE Linux Enterprise Server Architecture-Specific Information, sles-prep-zseries.pdf, describes the mainframe installation. We describe a high-level sequence of events in this section. Perform the following steps: 1. First, you must log in to z/VM as the Linux guest account using a 3270 emulator program. Enter the Conversational Monitor System (CMS) mode with the following command:
ipl cms

324

Solaris to Linux Migration: A Guide for System Administrators

SUSE Linux Enterprise Server installation requires the following files on the installation CD-ROM: vmrdr.ikr (the Linux kernel) initrd (the initial RAM disk) parmfile These files are usually saved on the z/VM guests minidisk with names such as: sles9.image sles9.initrd sles9.parm 2. Load (punch) the files into the z/VM guests virtual reader queue and then boot (ipl) the guest using the files in the readers queue. The installation program starts. You will be notified when you can login through ssh, telnet, or vnc to continue with the installation.

Sample SUSE Linux Enterprise Server installation files

This section contains sample sles9 parm a and sles9 exec a files used to initiate a SUSE installation. The sles9 parm file is limited to 10 lines. You place multiple parameters on a single line. Example 20-4 shows a sample of the sles9 parm a file.
Example 20-4 Example installation file: sles9 parm a ramdisk_size=65536 root=/dev/ram1 ro init=/Linuxrc TERM=dumb IP_HOST=Linux100.redbook.com IP_SEARCH=redbook.com IP_DNS=192.168.1.254 IP_ADDR=192.168.1.100 IP_GATEWAY=192.168.1.1 IP_INTERFACE=qeth IP_NETMASK=255.255.255.0 IP_BROADCAST=192.168.1.255 READ_DEVNO=0300 WRITE_DEVNO=0300 DATA_DEVNO=0302 PORTNAME=linport

You can manually enter the commands to start the installation, but most system administrators find it easier to create a script with the necessary commands. The z/VM editor is invoked with the xedit command. The editor is closed and the file saved with the save subcommand. For example:
xedit sles9 exec a

Example 20-5 on page 326 presents the contents of the sles9 exec a file in REXX format. To execute this file after you have saved it, type the sles9 command at the CMS prompt and the installation will begin.

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

325

Example 20-5 Example installation file: sles9 exec a /**/ 'close rdr' 'purge rdr all' 'spool punch * rdr' 'punch sles9 image a (noh' 'punch sles9 parm a (noh' 'punch sles9 initrd a (noh' 'change rdr all keep nohold' 'ipl 00c clear'

From this point on, refer to sles-prep-zseries.pdf, SUSE Linux Enterprise Server Architecture-Specific Information, for details about the specific type of installation (CD, ISO image, NFS) and access mode (SSH, Telnet, VNC) you have chosen.

20.4 Booting, system initialization, and shutdown

This section describes the procedures for the system boot, initialization, and shutdown processes for a z/VM Linux guest. We do not discuss native and LPAR installations.

Booting your Linux guest

Typically, Linux is booted from a storage device defined to the Linux guest user virtual machine. To boot the Linux guest, issue the following CP command, where vdev is the virtual device number containing the Linux guest operating system:
ipl vdev clear

The Linux boot process displays the boot messages on the z/VM (3270) console. To automate the process of IPLing a Linux guest during logon, include the following statement in the PROFILE EXEC A of the Linux guest user ID:
ipl vdev clear

Before including this statement in PROFILE EXEC, ensure that Linux has been successfully installed. Otherwise, you might get into a situation where z/VM will go into a continuous CP READ state. The clear operand to the ipl command is optional. If specified, the content of the virtual machine in storage is set to binary zeros before the guest operating system is loaded.

326

Solaris to Linux Migration: A Guide for System Administrators

ZIPL
The mainframe versions of Linux use the ZIPL bootloader. The configuration file is /etc/zipl.conf. If you make changes to /etc/zipl.conf you must run the zipl command to activate the changes. For example, you would run the zipl command when directed to after applying kernel updates or after you added different boot parameters such as additional disk drives.

20.5 Device management

Management of devices in native and LPAR installations requires specialized knowledge of mainframe configuration and management. This section describes the facilities provided by z/VM to Linux guests. The standard Linux commands to mount, unmount, format, and otherwise manage devices are available. The following section describes Linux reconfigurability, analogous to plug and play on the xSeries platform.

20.5.1 Linux reconfigurability

The channel device layer introduced with the Linux 2.4 kernel allows for new devices to be added to a Linux guest without an IPL. This means that you can you can use z/VM commands to add new devices, such as guest LAN NICs or real HiperSockets connections, and have these available to Linux without an IPL. In most cases, you can also use these new devices in TCP/IP immediately. Note that this applies only to devices attached through the channel subsystem. This does not apply to processors or storage.

3270 z/VM session

You can issue CP commands from the 3270 z/VM console by prefacing the command with the #CP control sequence. For example, to query the virtual machine minidisk assignments, issue the command:
#cp q v dasd

Telnet or SSH session

In order to issue CP commands from a Linux Telnet session, you need to install the cpint package. The cpint package is included in SUSE distributions. You can issue CP commands using the hcp command. For example, to display the status of virtual direct access storage devices, issue the command:
hcp q v dasd

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

327

There are two important items to note when using the hcp command: Do not enter the hcp command without any parameters; this will cause the virtual machine to enter the CP READ state. If this happens, you can log in using a 3270 emulator and enter the b command to return control of the guest to Linux. Remember to quote special characters (such as *) to prevent expansion by the bash shell. You can also download the cpint package from the following site:
http://www.linuxvm.org

20.5.2 DASD hot-plug example

DASD hot-plug is available on SLES 9 and Red Hat Enterprise Linux AS V4. For this example, we logged in to a 64-bit SLES 9 Linux guest linux041 as root: 1. First, we listed the z/VM virtual DASD and Linux DASD devices, as shown in Example 20-6.
Example 20-6 DASD device listing linux041:~ # hcp q v dasd DASD 0190 3390 51RRES R/O DASD 0191 3390 VMRU01 R/W DASD 019D 3390 51RW01 R/O DASD 019E 3390 51RW01 R/O DASD 0200 3390 LIC0F0 R/W DASD 0201 3390 LIC0E0 R/W DASD 0202 3390 LIC0EB R/W DASD 0203 3390 LIC0EC R/W DASD 0204 3390 LIC0ED R/W 107 2 146 250 3338 3338 3338 3338 3338 CYL CYL CYL CYL CYL CYL CYL CYL CYL ON ON ON ON ON ON ON ON ON DASD DASD DASD DASD DASD DASD DASD DASD DASD D08A D061 D08B D08B D0F0 D0E0 D0EB D0EC D0ED SUBCHANNEL SUBCHANNEL SUBCHANNEL SUBCHANNEL SUBCHANNEL SUBCHANNEL SUBCHANNEL SUBCHANNEL SUBCHANNEL = = = = = = = = = 0008 0002 0009 000A 000C 000D 000E 000F 0010

linux041:~ # cat /proc/dasd/devices 0.0.0200(ECKD) at ( 94: 0) is dasda 600840 blocks, 2347 MB 0.0.0201(ECKD) at ( 94: 4) is dasdb 600840 blocks, 2347 MB 0.0.0202(ECKD) at ( 94: 8) is dasdc 600840 blocks, 2347 MB 0.0.0203(ECKD) at ( 94: 12) is dasdd 600840 blocks, 2347 MB 0.0.0204(ECKD) at ( 94: 16) is dasde 600840 blocks, 2347 MB

: active at blocksize: 4096, : active at blocksize: 4096, : active at blocksize: 4096, : active at blocksize: 4096, : active at blocksize: 4096,

2. We linked to an additional disk with the following command:

linux041:~ # hcp link licdasd d158 0500 mr

328

Solaris to Linux Migration: A Guide for System Administrators

These messages appeared in /var/log/messages:

Jan 10 13:04:21 linux041 kernel: crw_info : CRW reports slct=0, oflw=0, chn=0, rsc=3, anc=1, erc=4, rsid=3 Jan 10 13:04:21 linux041 /etc/hotplug/ccw.agent[1926]: Configuring channel 0.0.0500 for event 'ccw'

3. The YaST2 GUI provides a graphical method for activating devices. We chose to use a command line utility to activate DASD 0500:
linux041:~ # dasd_configure 0.0.0500 1 0 Configuring device 0.0.0500 Setting device online

These messages appeared in /var/log/messages:

Jan 10 13:05:41 linux041 kernel: dasd(eckd): 0.0.0500: 3390/0A(CU:3990/04) Cyl:3338 Head:15 Sec:224 Jan 10 13:05:41 linux041 kernel: dasd(eckd): 0.0.0500: (4kB blks): 2403360kB at 48kB/trk linux disk layout Jan 10 13:05:41 linux041 kernel: dasdf:CMS1/ LIC158: dasdf1 Jan 10 13:05:42 linux041 /etc/hotplug/block.agent[2305]: new block device /block/dasdf Jan 10 13:05:42 linux041 /etc/hotplug/block.agent[2307]: new block device /block/dasdf/dasdf1

4. We listed the z/VM virtual DASD and Linux DASD devices again and found these additional entries:
linux041:~ # hcp q v dasd ... DASD 0500 3390 LIC158 R/W

3338 CYL ON DASD

D158 SUBCHANNEL = 0003

linux041:~ # cat /proc/dasd/devices ... 0.0.0500(ECKD) at ( 94: 20) is dasdf : active at blocksize: 4096, 600840 blocks, 2347 MB

At this point, we can access the device as /dev/dasdf. Any DASD operations, such as partitioning and formatting, are now available for use with this drive. 5. We wanted our new DASD device to be found during subsequent boots or IPLs, so we performed addition (optional) operations, as shown in Example 20-7, using mkinitrd and zipl.
Example 20-7 Using mkinitrd and zipl to make new DASD devices persistent linux041:~ # mkinitrd Root device: /dev/system/sysvol (mounted on / as ext3) Module list: jbd ext3 dm_mod dm-mod dm-snapshot dasd_eckd_mod Kernel image: Initrd image: /boot/image-2.6.5-7.191-s390x /boot/initrd-2.6.5-7.191-s390x

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

329

Shared libs: lib64/ld-2.3.3.so lib64/libacl.so.1.1.0 lib64/libattr.so.1.1.0 lib64/libblkid.so.1.0 lib64/libc.so.6 lib64/libdevmapper.so.1.01 lib64/libdl.so.2 lib64/libpthread.so.0 lib64/librt.so.1 lib64/libselinux.so.1 lib64/libuuid.so.1.2 Modules: kernel/fs/jbd/jbd.ko kernel/fs/ext3/ext3.ko kernel/drivers/md/dm-mod.ko kernel/drivers/md/dm-snapshot.ko kernel/drivers/s390/block/dasd_mod.ko kernel/drivers/s390/block/dasd_eckd_mod.ko DASDs: 0.0.0200(ECKD) 0.0.0201(ECKD) 0.0.0202(ECKD) 0.0.0203(ECKD) 0.0.0204(ECKD) 0.0.0500(ECKD) Including: udev dm/lvm2 initrd updated, zipl needs to update the IPL record before IPL!

linux041:~ # zipl -V Using config file '/etc/zipl.conf' Target device information Device..........................: 5e:00 Partition.......................: 5e:01 Device name.....................: dasda DASD device number..............: 0200 Type............................: disk partition Disk layout.....................: ECKD/compatible disk layout Geometry - heads................: 15 Geometry - sectors..............: 12 Geometry - cylinders............: 3338 Geometry - start................: 24 File system block size..........: 4096 Physical block size.............: 4096 Device size in physical blocks..: 22476 Building bootmap in '/boot/zipl' Adding IPL section 'ipl' (default) kernel image......: /boot/image at 0x10000 kernel parmline...: 'root=/dev/system/sysvol selinux=0 TERM=dumb elevator=cfq' at 0x1000 initial ramdisk...: /boot/initrd at 0x800000 Preparing boot device: dasda (0200). Syncing disks... Done.

6. After rebooting the Linux guest, our new DASD device became available at /dev/dasdf.

330

Solaris to Linux Migration: A Guide for System Administrators

20.6 Performance monitoring and tuning

The most critical phase of performance tuning is the design phase. The basic considerations are allocating the CPU, RAM, network, and DASD devices your applications and operating system will require. Allocate a RAM equivalent to your peak requirement to reduce swapping within the Linux guest. Review the documented best practices and conduct a Solution Assurance Review prior to beginning any proof of concept or proof of technology effort. The basic considerations are allocating and configuring the processor, memory, network, and DASD devices that your applications and operating system will require. For example, allocate enough memory to a Linux guest so that it does not swap.

20.6.1 Performance monitoring

Performance monitoring tools are available for both z/VM and Linux. Many third-party tools are available.

z/VM monitoring tools

The Performance Toolkit for VM enables you to monitor the resource utilization of your Linux guests, including CPU, memory, and I/O. The Distributed Data Server (DDS) acts to feed Linux monitor data to the Data Collector. DDS is distributed as the server component of the RMF PM for Linux. If you install this package, the Performance Toolkit will be able to monitor information about a Linux image, including: Linux system details Processes created per second Context switches per second CPU utilization, both user and kernel Memory utilization Network activity File system usage Apache performance data You can download the RMF PM for Linux utility from:
http://www.ibm.com/servers/eserver/zseries/zos/rmf/rmfhtmls/pmweb/pmlin.html

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

331

What's new in Performance Toolkit for VM in Version 5.1.0, gm130637.pdf, provides a high-level overview of setting up and using the Performance Toolkit for VM and RMF PM.

FCON/ESA support for Linux on z/VM

With release 3.2.03 of FCON/ESA, a Linux performance data interface has been added. Data retrieval and display of Linux internal performance data is based on the Linux Distributed Data Server (DDS) interface, originally written for use with Resource Measurement Facility (RMF) Performance Monitoring (PM). The DDS interface must be installed and active on all Linux systems that are to be monitored. Performance data retrieval is based on Extensible Markup Language (XML) requests, sent to the Linux systems via TCP/IP. Only the data actually needed for building a specific Linux performance report is retrieved. Performance data is collected only in the Linux systems, and performance history data is only available from the file systems of the Linux systems. FCON/ESA does not collect or save any Linux internal performance data.

Linux monitoring tools

Linux provides the same monitoring commands that are available on other hardware platforms. For example, the sar, vmstat, iostat, mpstat, top, and free commands are available. These commands provide a view of how a single Linux guest views its performance or behavior. These commands do not incorporate any information about any other guests in an LPAR or any z/VM resource utilization. It is difficult to determine actual or real resource utilization by a Linux guest since mainframe resources are shared and virtualized. There are many advanced z/VM and Linux monitoring tools which focus on either z/VM or Linux, but no native z/VM or Linux tools that monitor resources for both points of view. Consider, for example, a Linux guest that shows high CPU utilization with the top or sar command. This guest might actually get only a few CPU-cycles from z/VM. Accounting and monitoring for z/VM Linux guest machines, redp3818.pdf, describes one way of integrating output from the z/VM CP monitor facility and output from the Linux sar command to identify more realistically the CPU utilization of the Linux guest. More advanced performance monitoring tools are described in the IBM Redbook Linux on IBM Eserver zSeries and S/390: Performance Measurement and Tuning, SG24-6926.

332

Solaris to Linux Migration: A Guide for System Administrators

20.6.2 Performance tuning

System administrators will find this reference material extremely useful as they begin to tune mainframe distributions of Linux: Linux on zSeries: Tuning hints and tips
http://www.ibm.com/developerworks/linux/linux390/perf/index.html

Linux Performance when running under VM

http://www.vm.ibm.com/perf/tips/linuxper.html

Linux on IBM System z library

http://www.ibm.com/servers/eserver/zseries/os/linux/library/index.html

Linux on IBM Eserver zSeries and S/390: Performance Measurement and Tuning, SG24-6926 Implementing Linux on IBM Eserver zSeries and S/390: Best Practices
http://www.ibm.com/servers/eserver/zseries/library/whitepapers/pdf/gm130647.pdf

We describe high-level information from this reference material in this section.

z/VM performance tuning

z/VM offers many tuning parameters for allocating resources to virtual machines. Only a few of these parameters increase the amount of resources available to a guest. The rest of the parameters take away resources from one virtual machine and reallocate them to another one depending on its needs. Tuning does not increase the total amount of system resources. Resources are defined at the LPAR level. Tuning only allows defined resources to be used more effectively. Therefore, if your workload exceeds the capacity of your defined resources, you will not be able to achieve the performance you desire. This list contains some of the z/VM items that can be tuned to increase Linux guest performance: Tune memory requirements: Reduce storage for idle servers. Use a cloned server farm to share the kernel. Tune the Linux swap device: Disable minidisk caching or use a VDISK. Use multiple swap devices with different priorities. Tune scheduling: Assign LPAR weights.

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

333

 Configure the CP scheduler. Tune DASD: Use more, smaller disks and appropriate stripe sizes. Reduce channel subsystem contention. Configure minidisk caching (for some workloads). Tune network: Virtual LAN or direct attachment to OSA adapter.

Linux performance tuning

As with z/VM, you can achieve optimum performance only if the necessary resources are defined to the Linux guest. This list contains important Linux components for tuning on the mainframe: Tune memory requirements and processor performance: Apply the timer patch if it is not already incorporated into the kernel. Do not run unnecessary cron jobs. Do not run unnecessary processes. Tune the Linux swap device: Use multiple, smaller swap devices. Tune the file systems: Use the appropriate file system (ext2, ext3, reiserfs) for your application. Use LVM or software RAID to spread I/O across multiple drives. The preceding list is not comprehensive. For example, Linux kernel parameters can be tuned and process priorities can be set with the nice and renice commands.

20.7 Troubleshooting and diagnostics

Third-party software is available for both z/VM and Linux troubleshooting and diagnostics. This section describes standard z/VM and Linux troubleshooting and diagnostic tools and utilities.

20.7.1 z/VM troubleshooting and diagnostics

IBM introduced the Performance Toolkit for VM in the announcement of z/VM V4.4. The Performance Toolkit for VM provides enhanced capabilities for monitoring and reporting performance data. New function provided with z/VM

334

Solaris to Linux Migration: A Guide for System Administrators

V5.1 includes high-level Linux reports based on application monitor records from Linux guests. The Performance Toolkit for VM processes Linux performance data obtained from the Resource Management Facility (RMF) Linux performance gatherer, rmfpms. Performance monitor data can be viewed from Web browsers or PC-based 3270 emulator graphics. For information concerning the rmfpms application and the Performance Toolkit for VM, see:
http://www.ibm.com/servers/eserver/zseries/zos/rmf/rmfhtmls/pmweb/pmlin.html http://www.vm.ibm.com/related/perfkit/

20.7.2 Linux troubleshooting and diagnostics

Red Hat Enterprise Linux provides the sysreport package and command. SUSE Linux Enterprise Server provides the siga package and command and the dbginfo.sh command. These three commands can be used to gather system information that might be needed for troubleshooting a problem. Standard Linux troubleshooting and diagnostics techniques also work on the mainframe. Here are some examples: Set the log level for syslog and other applications. Standard network techniques such as ping, netstat, ifconfig, route, and traceroute are available. Development debugging tools such as gdb (GNU debugger), ddd (Data Display Debugger), nana, gprof (GNU profiler), and the MALLOC_CHECK_ environment variable are available. Commands such as strace, ltrace, ps, top, free, iostat, and vmstat are available. You can use the bash parameters -v and -x to debug scripts.

20.8 S/390 and zSeries-specific Linux commands

Mainframe distributions of Linux provide the commands common to all hardware platforms. These distributions also provide a unique set of commands that allow Linux to interact with the host operating system and virtualized devices.

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

335

20.8.1 Packages specifically for the mainframe

Red Hat Enterprise Linux and SUSE Linux Enterprise Server each provide packages with tools and utilities specifically designed to work with the mainframe. The packages are: Red Hat: s390-utils SUSE: cpint s390-tools s390 yast2-s390 (graphical interface for tools)

20.8.2 Commands specifically for the mainframe

These commands are provided by the packages listed in the previous section. The cpint package, while not included in Red Hat Enterprise Linux, can be downloaded from:
http://www.linuxvm.org

If the cpint package is installed, the #cp and hcp commands are available. Table 20-2 does not list the commands because the cpint package is not included in the default distribution.
Table 20-2 S/390 and zSeries-specific commands Task Send CP commands to the controlling VM from a 3270 session. The command prefix depends on your 3270 emulator settings. Modify generic attributes of channel attached devices. Read files from CMS volumes. Check a CMS formatted volume. Copy files from CMS volumes to Linux on S/390 file systems. List files on CMS formatted volumes. Report CMS formatted volume status. chccwdev cmsfscat cmsfsck cmsfscp cmsfslst cmsfsvol RHEL SLES #cp $cp

chccwdev

336

Solaris to Linux Migration: A Guide for System Administrators

Task Format disk drive. Display DASD and VTOC information. Create output similar to siga and sysreport (which are also available). Write a partition table to a DASD in the form of a VTOC (volume table of contents). Send CP commands to the controlling VM from a Telnet session. List channel subsystem devices. List channel attached direct access storage devices (DASD). List all qeth-based network devices with their corresponding settings. List channel-attached tape devices. Prepare a device for use as S/390 dump device. Get MAC and IP addresses from OSA and HiperSockets. Flush the ARP table. Configure IBM QETH function IPA, VIPA, and Proxy ARP. Create a 32-bit S/390 environment on S/390x. Create a 64-bit S/390 environment on S/390x. Display messages on a zSeries tape device. Adjust tunable parameters on DASD devices. Convert VMDUMPs into lkcd dumps. Copy dumps. Boot loader for S/390 and zSeries architectures.

RHEL dasdfmt dasdview

SLES dasdfmt dasdview dbginfo.sh

fdasd

fdasd hcp

lscss lsdasd ipconfig lstape

lscss lsdasd lsqeth lstape mkdump

qetharp qethconf

qetharp qethconf s390 s390x

tape390_display tunedasd

tape390_display tunedasd vmconvert

zgetdump zipl

zgetdump zipl

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

337

20.9 High availability

High availability (HA) techniques are available at the mainframe, z/VM, and Linux levels. All are possible, and some can be used in conjunction with others, but no single high availability solution currently exists that covers both Linux and z/VM requirements. Only Linux high availability products are available today. The basic rule of implementing a highly available environment is to identify and remove single points of failure. A general rule is to use redundant components (hardware and software). HA scenarios for mainframes and Linux on mainframes are described in various IBM Redbooks listed in 20.11, References on page 342. We describe some of the built-in HA features of the IBM mainframe and present high-level overviews of HA scenarios in this section.

20.9.1 Hardware HA
The mean time between failures (MTBF) on mainframe hardware is now measured in decades, and all mainframes are shipped with at least one hot spare processor that can be instantaneously substituted for a failing CPU. The mainframe hardware is able to detect CPU errors and transparently switch to another processor for continuous operation (CPU sparing). This function is transparent to the operating system. The zSeries memory is also self-checking and self-repairing, as is the I/O subsystem. Practically speaking, in the vast majority of cases, the only way a hardware failure can occur that will disrupt the entire environment is if the power supply to the mainframe CEC is somehow interrupted, or if all of the physical cable paths (I/O, network, or both) leading into the Linux LPAR or guest were cut. An additional hardware exposure to the mainframe environment occurs when a disruptive maintenance microcode patch has to be applied. In this case, the entire machine must be brought down. This type of maintenance is relatively rare, but does occur. In the event of a failure such as those just described, you need a backup mainframe with a similar configuration to provide HA.

20.9.2 z/VM HA
z/VM is designed using address spaces, memory protection, and fault-tolerant code. In an HA environment, z/VM can be configured to run in another LPAR: As a hot-standby, IPLed with backup guests but no work being processed. Cold-standby, defined but not IPLed.

338

Solaris to Linux Migration: A Guide for System Administrators

In a parallel processing configuration for high availability. IPLed with some guests active and processing work, and with some non-IPLed guests defined to serve as a warm standby for other operational z/VM systems.

Error recovery
The zSeries microcode and z/VM try to recover most errors, including such things as intermittent and permanent machine errors and system I/O errors, without manual intervention. When an error occurs, CP records it and sends a message to your z/VM primary system console. If the error condition does not damage system integrity and can be assigned to a single guest, z/VM makes a soft abend of the guest and operation continues. This means that in almost all cases a crash of a guest system does not affect CP operation, and other guests can operate without disruption.

Network HA
A critical component that might be overlooked when designing connectivity in a large-scale Linux installation is the z/VM TCP/IP stack itself. There can be many times when, as a result of misconfiguration or failure in a Linux system, you cannot get access through the network to a Linux guest. On these occasions, being able to obtain a connection through z/VM TCP/IP to the Linux console lets us do enough to get network connectivity again, log on, and fix the problem. Being able to get reliable access to the z/VM TCP/IP stack is, therefore, very important. The z/VM TCP/IP stack supports the VIPA function. We suggest that you look at using multiple interfaces to the z/VM stack, and use VIPA to address z/VM. The IBM Redbook Linux on IBM Eserver zSeries and S/390: ISP/ASP Solutions, SG24-6299, discusses a method of using multiple attachments in combination with a virtual IP address configured on a dummy interface. This still can form the basis of a high availability connectivity solution. As an alternative to using a dummy interface, you can configure additional addresses against the IP loopback interface (lo). This means you do not need to have another module installed. The way that the Linux kernel processes incoming IP traffic does not require the virtual IP address to be associated with a dummy or loopback interface. Your virtual address can be just as easily defined against one of your real interfaces.

20.9.3 Automating Linux guest boot and shutdown

For regular z/VM system start (IPL), or after a z/VM termination, an automated restart of z/VM can be configured. During system initialization, CP automatically

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

339

logs on (starts) user AUTOLOG1. AUTOLOG1 can automatically start other virtual machines when issuing the AUTOLOG command in the PROFILE EXEC of AUTOLOG1. If a certain start order is necessary due to application dependencies, you can implement a startup procedure for the guest systems. This startup procedure is called within the PROFILE EXEC of user AUTOLOG1. The procedure is written in the script language REXX.

20.9.4 Linux-HA
Linux high availability functions specific to the zSeries platform include the Linux Virtual Server and the Heartbeat option.

Linux Virtual Server

The Linux Virtual Server (LVS) concept ties a group of server machines together using various network and scheduling techniques so that they present a single server image to clients requesting service. A single load balancing and scheduling machine initially receives all the traffic, which it then sends to back-end servers. A sample implementation of LVS is described in the IBM Redbook Linux on IBM Eserver zSeries and S/390: Distributions, SG24-6264.

Heartbeat and network option

TCP/IP Heartbeat techniques can be used for z/VM Linux guests in same LPAR, in different LPARs in the same mainframe, or in different mainframes.

20.9.5 Backup and recovery options

There are three basic options for backing up and restoring Linux data on mainframes: The first option is to use the z/VM DASD Dump/Restore (DDR) Facility to back up and restore Linux data. This is generally the most common strategy that customers use for Linux and z/VM backups. DDR only provides volume-level backups, not file-level backups. That is, you can back up and restore entire minidisks, but not individual files. To restore individual files in this scenario, customers typically restore the minidisk that the file in question was on, and then mount that volume to the Linux guest and retrieve the lost file. The next basic option for Linux backups is to modify an existing z/OS backup strategy to include Linux volumes. The benefit of using existing z/OS backup facilities (most notably DFSMSdss) is that Linux DASD volumes can generally be treated like any other volume that is dumped and restored using dss. Although z/OS cannot access the data in the volume itself, the volume can be dumped and restored just like any other DASD volume.

340

Solaris to Linux Migration: A Guide for System Administrators

The backup strategy that is least used currently, but is growing the fastest, is to use tools that run on Linux to back up Linux volumes and files. The most attractive argument for this option is that these backup tools provide easy access to both file-level and volume-level backups.

20.10 Security
Linux running in native mode in an LPAR can only rely on the security measures provided by the Linux distribution. Linux running as a z/VM guest, however, has access to several additional security features provided by the mainframe hardware and z/VM or z/OS: Consider HiperSockets, guest LANs, and virtual switches. These networking options are implemented either in memory or directly through the host systems network adapters. Network communication between operating systems running on the mainframe never need to leave the mainframe. That means there is nothing on the physical network to be sniffed or captured by other programs. The security provided by LPAR isolation is certified by Evaluation Assurance Levels (EAL) 5 Common Criteria. Linux can access zSeries specialized cryptographic hardware. Each z/VM guest has its own user ID and password. Each guests runs in its own virtual address space and cannot access another guests memory. Linux guests can use z/OS LDAP and RACF facilities for authentication. The following publications describe the concepts and techniques for implementing these interfaces between Linux, z/VM, and zSeries hardware: Linux on IBM zSeries and S/390: Securing Linux for zSeries with a Central z/OS LDAP Server (RACF), REDP-0221 Linux on IBM Eserver zSeries and S/390: Best Security Practices, SG24-7023 For documentation concerning z/VM security and integrity resources such as the z/VM Common Criteria, refer to:
http://www.vm.ibm.com/security

This Web site also provides links to more information about zSeries, z/VM, and Linux security features. The standard Linux system administration facilities such as PAM, OpenSSL, OpenSSH, and OpenLDAP are available, too. Linux guests can secure communications using scp, sftp, and iptables. User authentication can be

Chapter 20. IBM eServer zSeries and IBM System z hardware platform specifics

341

through /etc/passwd, Kerberos and sudo. Linux security is Linux security, even on the mainframe.

20.11 References
For more information, refer to the following publications: Linux on IBM zSeries and S/390: Server Consolidation with Linux for zSeries, REDP-0222 z/VM and Linux on zSeries: From LPAR to Virtual Servers in Two Days, SG24-6695 Linux for S/390, SG24-4987 Linux on IBM Eserver zSeries and S/390: Large Scale Linux Deployment, SG24-6824 Linux on IBM zSeries and S/390: High Availability for z/VM and Linux, REDP-0220 Linux on IBM Eserver zSeries and S/390: Building SuSE SLES8 Systems under z/VM, REDP-3687 Printing with Linux on zSeries Using CUPS and Samba, REDP-3864 Linux with zSeries and ESS: Essentials, SG24-7025 IBM System z9 and Eserver zSeries Connectivity Handbook, SG24-5444 Linux on IBM Eserver zSeries and S/390: Best Security Practices, SG24-7023 Accounting and Monitoring for z/VM Linux guest machines, REDP-3818 Linux on IBM Eserver zSeries and S/390: Performance Measurement and Tuning, SG24-6926 Linux on IBM zSeries and S/390: Securing Linux for zSeries with a Central z/OS LDAP Server (RACF), REDP-0221 Linux on IBM Eserver zSeries and S/390: Best Security Practices, SG24-7023 Linux for IBM System z9 and zSeries, SG24-6694 Networking Overview for Linux on zSeries, REDP-3901 IBM Eserver zSeries 900 Technical Guide, SG24-5975 IBM Eserver zSeries 890 Technical Introduction, SG24-6310 Technical Introduction: IBM Eserver zSeries 800, SG24-6515 IBM Eserver zSeries 990 Technical Guide, SG24-6947 IBM System z9 109 Technical Guide, SG24-7124 What's new in Performance Toolkit for VM in Version 5.1.0
http://www.vm.ibm.com/library/gm130637.pdf

342

Solaris to Linux Migration: A Guide for System Administrators

Part 4

Part

Appendixes
Appendix A, Tasks reference on page 345 provides quick reference tables that present methods for accomplishing equivalent tasks between Solaris and Linux. Appendix B, Commands and configuration files reference on page 365 provides quick reference tables for the comparison of configuration files, comparable commands, and common commands. Appendix C, UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter) on page 371 provides sample content from a related publication: Mendoza, et al., UNIX to Linux Porting: A Comprehensive Reference, Prentice Hall, 2006, ISBN 0131871099. Appendix D, Example: System information gathering script on page 397 provides a description of the getsysinfo sample diagnostic scripts, used to gather detailed system configuration information. Appendix E, Additional material on page 401 provides information about how to access the additional material provided with this IBM Redbook, available on the Internet.

Copyright IBM Corp. 2006. All rights reserved.

343

344

Solaris to Linux Migration: A Guide for System Administrators

Appendix A.

Tasks reference
This appendix contrasts common tasks between the Solaris and Linux operating systems. Tasks are grouped according to the following major categories. Each major category is contained within a table. Tables can also include file location information or pertinent information that is related to the category they contain. This reference provides information about Linux and Solaris in the following categories: Packaging Installing and upgrading tasks Booting and shutting down User management tasks Device management and configuration Network management and configuration Managing system resources Printer management and configuration Disk and file system management Logical volume management Network troubleshooting

Copyright IBM Corp. 2006. All rights reserved.

345

Packaging
Table A-1 provides information that shows differences in Linux and Solaris packaging-related tasks.
Table A-1 Packaging comparison Units Smallest installable unit. Single installable image; must be distributed and installed as a unit. Logical grouping of packages. Logical grouping of packages and software clusters. Solaris Package Patch Package Software cluster Patch cluster Software configuration clusters, for example: Core: Required operating system files End-User System Support: Core plus window environment Developer System Support: End-User plus the development environment Entire Distribution: Developer System plus enhanced features Entire Distribution Plus OEM: Entire Distribution plus third-party hardware drivers (on SPARC only) Linux Package Package Package Workstation install (stand-alone client) Developer install (above, including developer tools, compiler) Server install (server services and software) Full Install (all software/services) Custom (choose packages manually)

Installing and upgrading tasks

Table A-2 provides information that shows differences in Linux and Solaris installation and upgrade tasks.
Table A-2 Installing and upgrading tasks Task Install packages Solaris pkgadd Red Hat SUSE rpm -i

346

Solaris to Linux Migration: A Guide for System Administrators

Task Display installed packages Remove software package Upgrade/install a package Verify correct installation Install a patch List content of installed package Check which file belongs a package Check package information Remove a patch Display installed patches Install OS on another disk (alternate disk installation) Create an installation server for network installation Create a boot server for network installation Set up a client for network installation

Solaris pkginfo or pkgparam pkgrm pkgadd pkgchk patchadd Look in /var/sadm/install/ contents Look in /var/sadm/install/ contents Look in /var/sadm/pkg/ PKGNAME/pkginfo patchrm showrev -p Live Upgrade

Red Hat

SUSE rpm -qa rpm -e rpm -U rpm -V

rpm -F (upgrades only if already present) rpm -ql

rpm -qf

rpm -qi N/A N/A Install different OS on different disk/partition, dual boot using GRUB anaconda (see http://www.tldp.org/H OWTO/Network-InstallHOWTO.html) grub (see http://www.tldp.org/H OWTO/Clone-HOWTO/) anaconda (see http://www.tldp.org/H OWTO/Network-InstallHOWTO.html) autoyast (see http://www.tldp.org/H OWTO/Network-InstallHOWTO.html) grub (see http://www.tldp.org/H OWTO/Clone-HOWTO/) autoyast (see http://www.tldp.org/H OWTO/Network-InstallHOWTO.html)

setup_install_server install_dir_path

setup_install_server -b bootdirpath add_install_client

Appendix A. Tasks reference

347

Booting and shutting down

Table A-3 provides information that shows differences in the process and locations of items that are involved in booting and shutting down a system in Linux and Solaris.
Table A-3 Boot and shutdown tasks Task/location Boot process Solaris Phases: 1. OpenPROM: Display system information, run POST, load bootblk, and locate ufsboot. 2. Boot programs: bootblk loads and executes the ufsboot. 3. Kernel initialization: ufsboot loads and executes the core kernel, initializes core kernel data structures, loads other kernel modules based on the /etc/system file, and starts /sbin/init program. 4. init: Starts other processes based on the /etc/inittab file. 5. /sbin/rcX runs the corresponding /etc/rcX.d scripts to start up other components. Linux Phases: 1. BIOS: Checks the system and peripheral devices. Locates and runs the Master Boot Record (MBR). 2. MBR loads GRUB (GRand Unified Bootloader). 3. GRUB boots the kernel using information in /boot/grub/menu.lst. 4. Kernel starts initializing devices and loads additional device drivers from initrd image, and then starts init process from /sbin/init. 5. init: Starts other processes based on the /etc/inittab file. 6. /etc/rc.d/rc in RHEL or /etc/init.d/rc in SLES runs the corresponding /etc/rcX.d scripts to start up other components. /boot partition or directory, contains the kernels and initrd images.

Default kernel location

/kernel - Contains common components. /platform/platform-name/kernel - Contains the platform-specific kernel components. /platform/hardware-class-name/ke rnel - Contains kernel components specific to this hardware class. /usr/kernel - Contains kernel components common to all platform within a particular CPU set.

348

Solaris to Linux Migration: A Guide for System Administrators

Task/location Default kernel modules location

Solaris In each of the subdirectories mentioned in the proceeding cell, they might contain the following subdirectories: drv - Contains loadable device drivers. exec - Contains modules that run programs stored in different file formats. fs - Contains file system modules. misc - Contains miscellaneous system-related modules. sched - Contains operating system schedulers. strmod - Contains loadable System V STREAMS modules. sys - Contains loadable system calls. cpu - Contains CPU type-specific modules. tod - Contains time-of-day hardware modules. sparcv9 - Contains 64-bit versions of specific modules.

Linux Use the output of the following command: /lib/modules/uname -r/

System run levels Run level S or s Single user level (boot -s). Only some file systems are mounted as opposed to run level 1. Shuts down the operating system. Does not attempt to turn off the power. Returns to OpenPROM ok prompt. Administrative single user mode. Can access all available file systems. User logins disabled. Multiuser without NFS resources shared. Minimal single user mode on SUSE. In Red Hat, it is same as initlevel 1. Power-down state, shuts down the operating system and tries to turn power off if supported by the system. Equivalent to Solaris run level 5. Single user mode.

Run level 0

Run level 1

Run level 2

Multiuser without NFS resources shared. Text only login.

Appendix A. Tasks reference

349

Task/location Run level 3

Solaris Multiuser with NFS resources shared (default run level for Solaris). The login is text or graphical depending on the console capabilities. N/A (alternate multiuser). Power-down state, shuts down the operating system and tries to turn power off if supported by the system. Reboot. N/A who -r telinit run-level-number init run-level-number /sbin/rc run-level-number

Linux Multiuser with NFS resources shared. Text only login.

Run level 4 Run level 5

N/A (alternate multiuser). Full multiuser, NFS resources shared, and graphic login with X display manager. Reboot. Same as in Solaris. Ignore the /etc/inittab file and launch the default shell. who -r runlevel telinit run-level-number init run-level-number In RHEL: /etc/rc.d/rc run-level-number In SLES: /etc/init.d/rc run-level-number grub cat /boot/grub/menu.lst BIOS

Run level 6 Emergency run level Determine a system's run level Change a system's run level Startup script

Use new kernel Display boot loader information Display or alter the list of boot devices Boot methods Default boot mode

N/A N/A boot

boot or if auto-boot is set to true, it is automatic. Note: The commands listed in this column is assumed that you are at the OpenPROM prompt.

Automatic or press Enter at grub prompt. Note: The commands listed in this column assume that you are at the GRUB prompt. Add -s, single, S, or 1 at the end of boot line. Not needed.

Single user mode Reconfiguration mode

boot -s boot -r

350

Solaris to Linux Migration: A Guide for System Administrators

Task/location Verbose boot mode

Solaris boot -v

Linux In RHEL, remove rhgb quiet from the boot command line; in SLES, press Esc when booting. Edit grub boot line to specify other kernel and options. Boot from a rescue CD-ROM or from network. Possible but vendor dependent.

Interactive boot mode Recovery boot mode Interactive start of services while booting Emergency boot mode Shutdown and reboot Shutdown

boot -a boot cdrom boot net N/A

N/A reboot shutdown -i 6 halt poweroff shutdown

Add emergency or -b option at the end of boot line. reboot shutdown -r now halt poweroff shutdown

User management tasks

Table A-4 provides information that shows differences in Linux and Solaris user management-related tasks.
Table A-4 User management tasks Task/locations Run multiple tasks in a GUI environment Add a user Remove a user Change a user List users Password files Group files Solaris admintool useradd userdel usermod listusers /etc/passwd and /etc/shadow /etc/group Linux In RHEL: system-config-users In SLES: yast2 useradd userdel usermod N/A /etc/passwd and /etc/shadow /etc/group

Appendix A. Tasks reference

351

Task/locations Process resource limits for users System-wide environment file Configuration information for user authentication Profile templates

Solaris N/A N/A /etc/pam.conf /etc/skel/

Linux /etc/security/limits.conf /etc/profile /etc/login.defs /etc/skel/

Device management and configuration

Table A-5 provides information that shows differences in Linux and Solaris device management and configuration tasks.
Table A-5 Solaris versus Linux device management commands Device management task Run multiple tasks in a GUI environment Solaris admintool Red Hat system-config-* (keyboard, mouse, printer, network, soundcard, and so on) /dev/MAKEDEV kudzu mknod kudzu SUSE yast2

Configure a device Define a device Remove a device List devices Install a device driver kernel module List installed device driver kernel modules Remove a device driver kernel module

devfsadm devfsadm devfsadm -C or -c sysdef modload modinfo modunload

/dev/MAKEDEV hwinfo mknod hwinfo /dev/MAKEDEV ls /sys/devices cat /proc/devices insmod lsmod rmmod

352

Solaris to Linux Migration: A Guide for System Administrators

Network management and configuration

Table A-6 provides information that shows differences in Linux and Solaris network management and configuration tasks.
Table A-6 Network management and configuration tasks Task Run multiple tasks in a GUI environment Configure TCP/IP Solaris N/A Editing the following files: /etc/hostname.* /etc/inet/* /etc/defaultrouter /etc/defaultdomain /etc/nodename /etc/netmasks ifconfig netstat -i ifconfig netstat vi /etc/resolv.conf vi /etc/nsswitch.conf ndd /dev/ip \? ndd /dev/tcp \? ndd ndd N/A netstat -r ping traceroute Linux system-config-network in RHEL, yast2 in SLES RHEL: /etc/sysconfig/network and /etc/sysconfig/networking/* SLES: /etc/sysconfig/network/*

Display interface settings Display interface status and statistics Configure interface Check various network statistics Change resolver Configure name services Display kernel network parameters Configure kernel network parameters Check for network link Rename a network interface Check routing table Testing for connectivity Check IP path

ip ifconfig netstat -i ifconfig ip ifconfig netstat vi /etc/resolv.conf vi /etc/nsswitch.conf sysctl -a | grep ^net sysctl -w variable=value ethtool mii-tool ip netstat -r route ping traceroute

Appendix A. Tasks reference

353

Task Capture network packets

Solaris snoop

Linux tcpdump tethereal ethereal

NFS management and configuration

Table A-7 provides information that shows differences in Linux and Solaris NFS management and configuration tasks.
Table A-7 Solaris NFS and Linux NFS differences Task/configuration file Manually start the NFS server Solaris /etc/init.d/nfs.server start mount -F nfs server://resource /mount/point /etc/default/nfs /etc/default/nfslogd shareall share /etc/dfs/dfstab Linux RHEL: /etc/init.d/nfs start SLES: /etc/init.d/nfsserver start mount server://resource /mount/point /etc/sysconfig/nfs /etc/sysconfig/nfs exportfs -a exportfs -o /etc/exports

Mount a resource on a NFS client NFS server general config file NFS logging options config file Share all exported file systems Share a new file system Config file of shared file systems

Managing system resources

Table A-8 provides information that shows differences in querying Linux and Solaris systems for general information. For more detailed descriptions of the commands and their usage, refer to Chapter 9, Managing system resources on page 139.
Table A-8 System information tasks Task List system information List processor information Solaris uname psrinfo Linux uname cat /proc/cpuinfo

354

Solaris to Linux Migration: A Guide for System Administrators

Task List system memory size List disk space usage List file space usage List physical disks Show the systems host name List systems network configuration List active processes List system configuration List system diagnostic information

Solaris prtconf | grep Memory df du format hostname ifconfig prstat prtconf prtdiag

Linux cat /proc/meminfo | grep MemTotal df du fdisk -l hostname ifconfig ps -e Red Hat: lshal SUSE: hwinfo Red Hat: lspci and /proc/* files SUSE: hwinfo

Managing system services

Table A-9 provides information that shows differences in Linux and Solaris service management and configuration tasks.
Table A-9 Summary of system service tasks Task Add a service that does not already exist in chkconfig Show existing service links Enable service to start at boot time Solaris Copy start script into the appropriate run level directory, prefixed with S. ls run level directory. Copy start script into the appropriate run level directory, prefixed with S. Red Hat chkconfig --add service name SUSE chkconfig -a or --add service name

chkconfig --list | grep service name chkconfig service name on

chkconfig service name chkconfig -s or --set service name on

Appendix A. Tasks reference

355

Task Disable service from running at boot time

Solaris Rename the service start script to a name that does not start with S (typically lowercase s).

Red Hat chkconfig service name off

SUSE chkconfig -s or --set service name off

Managing scheduling and cron

Table A-10 provides information that shows differences in Linux and Solaris scheduling and cron management and configuration tasks.
Table A-10 Scheduling and cron service tasks Task Run a command at a predefined time Run a command or set of commands periodically Run a command or set of commands periodically, running when possible if system is off at the scheduled time Solaris at cron and crontab N/A Linux at cron and crontab anacron and anacrontab

Managing quota
Table A-11 provides information that shows differences in Linux and Solaris quota management and configuration tasks.
Table A-11 Quota tasks Task Display disk usage and limits Edit users quotas Check consistency between a file system and the disk quota file Solaris quota edquota quotacheck Linux quota edquota quotacheck

356

Solaris to Linux Migration: A Guide for System Administrators

Managing process accounting

Table A-12 provides information that shows differences in Linux and Solaris process accounting management and configuration tasks.
Table A-12 Process accounting tasks Task Turn on process accounting List users connection time List commands run previously List users connection time and idle time Solaris accton N/A lastcomm who Linux accton ac lastcomm who

Printer management and configuration

Table A-13 provides information about tasks that are involved in printer management and configuration in Linux and Solaris. For detailed instructions about configuring Linux printers, see Chapter 10, Printing services on page 151.
Table A-13 Printer management and configuration Task Run multiple tasks in a GUI environment Add a printer Start a print queue Stop a print queue Display print queue status Cancel a print job Add a print queue Change a print queue Remove a print queue Display settings of a print queue Solaris admintool lpadmin enable disable lpstat cancel lpadmin lpadmin lpadmin lpadmin Linux RHEL: system-config-printer SLES: yast2 lpadmin lpc lpc lpc or lpq lprm lpadmin lpadmin lpadmin lpadmin

Appendix A. Tasks reference

357

Disk and file system management

Table A-14 and Table A-15 provide information that shows differences in Linux and Solaris disk and file system management and configuration tasks.
Table A-14 Disk management tasks Task Create device files for new device To create partitions on the disk. Display partition table Maximum usable partitions Solaris devfsadm format or fmthard prtvtoc 7 Linux udev fdisk or parted fdisk or parted 15 for SCSI 63 for IDE

Table A-15 File system management tasks Task Run multiple tasks in a GUI environment Format a disk Check a file system Mount a file system Display available file system space Partition a disk List a volume's table of contents Format a file system Unmount a file system Back up file systems, files, directories Restore file systems, files, directories Change a file system Display file system information Display file system mount table Solaris N/A format fsck mount df format prtvtoc newfs or mkfs umount ufsdump ufsrestore tunefs cat /etc/vfstab /etc/mtab Linux N/A fdisk fsck mount df fdisk fdisk mkfs umount dump restore RHEL and SLES: resize2ext SLES: resize_reiserfs cat /etc/fstab /etc/mtab

358

Solaris to Linux Migration: A Guide for System Administrators

Swap management
Table A-16 provides information that shows differences in Linux and Solaris swap management and configuration tasks.
Table A-16 Swap tasks Task Create a swap space Enable a swap space Enable all swap spaces Disable a swap space Disable all swap spaces Delete a swap space Display the swap space usage Display the swap space status Set swap priority level Solaris swap -a swap -a N/A N/A N/A swap -d swap -s swap -l N/A Linux mkswap swapon swapon -a swapoff swapoff -a N/A swapon -s N/A swapon -p

Logical volume management

Table A-17 and Table A-18 on page 361 provide information that shows differences in Linux and Solaris logical volume management and configuration tasks.
Table A-17 Logical volume management Volume management task Initialize a disk for VM Solaris Create partition area/size: metainit metainit volumename raidtype devices... N/A Linux Create partition area/size: pvcreate mkraid, mdadm lvcreate LVname vgcreate VGname devicename mkraid, mdadm lvchange -a y LVname vgchange -a y VGname raidstart

Create a volume or volume group (VG) Enable the volume or volume group

Appendix A. Tasks reference

359

Volume management task Disable the volume or volume group Delete the volume or volume group Add a device to the volume or volume group Delete a device from the volume or volume group Create a soft partition or logical volume (no RAID) Create a soft partition or logical volume (RAID 0) Create a soft partition or logical volume on a specific device

Solaris N/A

Linux lvchange -a n LVname vgchange -a n VGname raidstop, mdadm lvremove LVname vgremove VGname lvextend LVname vgextend VGname newdevname lvreduce LVname vgreduce VGname devicename raidreconf lvcreate -Lsize -nLVname VGname lvcreate -iNumOfStripes -IStripeSize -nLVname VGname mdadm, mkraid Same as above, but add the devicename to the end of the command line. lvremove /dev/VGname/LVname raidreconf lvextend -Lsize /dev/VGname/LVname raidreconf Red Hat - ext2/ext3: resize2fs SUSE - ext2: resize2fs Red Hat - reiser: resize_reiserfs Note: View the man pages for unmount requirement, and check if online version is available. ext2: resize2fs lvreduce reiser: resize_reiserfs lvreduce Note: View the man pages for unmount requirements.

metaclear metattach or metainit metadetach

metainit -p metainit with RAID 0 on devices first, and then metainit -p to create the soft partition volume. Same as above, but the second metainit -p will have the devicename at the end of the command line. metaclear metattach volume [size or devicename] growfs

Delete a soft partition or logical volume Extend a volume or logical volume Extend a file system after volume has been grown

Reduce a volume or logical volume

metadetach Volname devicename

360

Solaris to Linux Migration: A Guide for System Administrators

Table A-18 Volume management commands Volume management task Set up or display metadb Display metadevice status Initialize raw devices to metadevices Attach metadevices Detach metadevices Clear and remove metadevices Replace a metadevice Rename a volume Check metadevice ID configuration Manage hot spares Set submirrors offline Set submirrors online Change volume parameters Back up volume group metadata Recover soft partition configuration information Set up root file system for mirroring Administer disk sets for sharing Resynchronize volume during reboot Expand a file system size Start up the kernel metadevice modules Set the default number of available volumes Solaris metadb metastat metainit metattach metadetach metaclear metareplace metarename metadevadm metahs metaoffline metaonline metaparam N/A metarecover metaroot metaset metasync growfs /etc/system /kernel/drv/md.conf Linux vgdisplay, lvdisplay, lsraid vgdisplay, lvdisplay cat /proc/mdstat pvcreate vgchange, lvchange vgchange, lvchange vgremove, lvremove raidreconf raidreconf mdadm mdadm, raidhotadd, raidhotremove mdadm mdadm vgchange, lvchange vgcfgbackup vgcfgrestore N/A N/A N/A vgextend, lvextend, resize2fs, resize_reiserfs, raidreconf N/A /etc/sysconfig/lvm /etc/raidtab

Appendix A. Tasks reference

361

General troubleshooting
Table A-19 provides general troubleshooting information highlighting differences between methods and configuration file locations between Solaris and Linux (RHEL, SLES).
Table A-19 General troubleshooting Task/location Change a host name Solaris Minimum change required for the following files: /etc/nodename /etc/hosts /etc/hostname.* /etc/net/*/hosts /etc/services Red Hat hostname (while running) or edit /etc/sysconfig/network (for permanent change after network restart) SUSE /etc/HOSTNAME and /etc/hosts

List of well-known networking services and port numbers List of well-known protocols Display NFS and RPC statistics Display a snapshot of virtual memory Display virtual memory statistics Display I/O statistics Report system activity Display simple and complex lock contention information Report CPU usage Display system error log Display paging/swapping space Disk crash dump utilities

/etc/services

/etc/protocols nfsstat N/A vmstat iostat sar N/A

/etc/protocols nfsstat cat /proc/meminfo vmstat iostat sar View /var/lock directory

top dmesg swap -l dumpadm savecore

top dmesg cat /proc/swaps /etc/init.d/diskdump savecore lkcd

362

Solaris to Linux Migration: A Guide for System Administrators

Task/location Disk crash config files Network crash dump utilities Network crash config files Crash analyzing tools Crash dump default store directory Specify users who have access to cron (every user has access to cron if the access file does not exist.) Specify users who have no access to cron Specify remote users and hosts that can execute commands on the local host Default superuser log Configure syslogd logging Display physical RAM Start or stop scripts directory Devices directory

Solaris /etc/dumpadm.conf N/A

Red Hat /etc/sysconfig/ diskdump /etc/init.d/netdump /etc/init.d/netdump -server /etc/sysconfig/netdump /etc/netdump.conf crash /var/crash

SUSE /etc/sysconfig/dump lkcd

N/A mdb /var/crash /etc/cron.d/cron. allow

/etc/sysconfig/dump lcrash /var/log/dump

/etc/cron.d/cron.allow

/etc/cron.d/cron.deny /etc/hosts.equiv

/etc/cron.d/cron.deny /etc/hosts.equiv

/var/adm/sulog /etc/syslog.conf prtconf /etc/init.d /dev and /devices

/var/log/messages /etc/syslog.conf cat /proc/meminfo /etc/init.d and /etc/rc.d /dev /etc/init.d

Network troubleshooting
Table A-20 on page 364 provides network troubleshooting information highlighting differences in methods and configuration file locations between Solaris and Linux (RHEL, SLES).

Appendix A. Tasks reference

363

Table A-20 Troubleshooting network problems Tasks Display interface settings Display interface status and statistics Configure interface Check various network statistics Check DNS resolver Check name services config Display kernel network parameters Configure kernel network parameters Check for network link Query DNS Check routing table Check ARP entries Testing for connectivity Check IP path Capturing network packets Solaris ifconfig netstat -i ifconfig netstat more /etc/resolv.conf more /etc/nsswitch.conf ndd /dev/ip (\?)parameter ndd /dev/tcp (\?)parameter ndd -set driver parameter ndd driver link_status nslookup dig netstat -r arp -a ping traceroute snoop Linux ip ifconfig netstat -i ifconfig ip ifconfig netstat more /etc/resolv.conf more /etc/nsswitch.conf sysctl -a | grep ^net sysctl -w variable=value ethtool mii-tool nslookup dig netstat -r route arp ping traceroute tcpdump tethereal ethereal

364

Solaris to Linux Migration: A Guide for System Administrators

Appendix B.

Commands and configuration files reference

This appendix contains the following tables: Table B-1 on page 366: Configuration file comparisons between Solaris and Linux Table B-2 on page 367: Command equivalencies between the two operating systems Table B-3 on page 369: Commands common to both systems

Copyright IBM Corp. 2006. All rights reserved.

365

Configuration and other files

Table B-1 provides a quick reference to compare names and locations of configuration files between Solaris and Linux.
Table B-1 Configuration and other files Solaris /etc/auto_master, /etc/auto_home /etc/default/login /etc/default/nfs /etc/default/nfslogd /etc/defaultdomain, /var/yp/binding/'domainname'/yp servers /etc/dfs/dfstab /etc/ftpusers (Solaris 8) /etc/ftpd/ftpusers (Solaris 9) /etc/inet/ (directory) Linux /etc/auto.master, /etc/auto.home /etc/securetty /etc/sysconfig/nfs /etc/sysconfig/nfs /etc/yp.conf Comments

/etc/exports /etc/vsftpd.ftpusers(vsftpd) /etc/ftpusers (wu-ftpd) /etc/hosts, /etc/HOSTNAME, /etc/sysconfig/network (directory) /etc/services /etc/xinetd.conf, /etc/xinetd.d (directory) /etc/krb5.conf N/A /etc/logindevperm /etc/mime.types /etc/mtab /etc/nscd.conf /etc/pam.d (directory) /etc/cups/printers.conf Different format (see Chapter 10, Printing services on page 151). Cache names different. Different format. See Chapter 12, Monitoring and performance on page 173. Different formats. Different format.

/etc/inet/services /etc/inetd.conf /etc/krb5/krb5.conf /etc/logadm.conf /etc/logindevperm /etc/mime.types /etc/mnttab /etc/nscd.conf /etc/pam.conf /etc/printers.conf

366

Solaris to Linux Migration: A Guide for System Administrators

Solaris /etc/rmmount.conf /etc/security/policy.conf /etc/syslog.conf /etc/system /etc/vfstab /etc/vold.conf /usr/share/lib/zoneinfo/* /var/adm/messages, /var/log/syslog

Linux N/A N/A /etc/syslog.conf /etc/sysctl.conf /etc/fstab N/A /usr/share/zoneinfo/* /var/log/messages, /var/log/syslog, /usr/adm/messages, /var/log/maillog /etc/crontab, /var/spool/cron

Comments See Chapter 6, Device management on page 85. See Chapter 11, Users and groups on page 159. Slightly different format.

See Chapter 6, Device management on page 85.

/var/spool/cron/crontabs/root

Comparable commands
Table B-2 provides a quick reference for comparable commands between Solaris and Linux.
Table B-2 Comparable commands Solaris /sbin/swapadd adb admintool apptrace arch -k cc devfsadm Linux /sbin/swapon gdb yast2 (SUSE) /usr/bin/system-* (Red Hat) strace and ltrace uname -m gcc modprobe, kerneld, insmod, hotplug, cardctl Usage Enable swap devices Debugger Admin GUI Application API tracing List machine type C compiler Add device without reboot

Appendix B. Commands and configuration files reference

367

Solaris df -k getfacl, setfacl gzcat file | tar -xvf lpsched modinfo mountall mvdir nawk patchadd patchrm, pkgrm pkgadd pkgchk pkginfo, pkgparam priocntl -e, nice priocntl -s, renice prstat prtdiag prtvtoc ps -ef snoop tip truss, sotruss umountall useradd who -r

Linux df getfacl, setfacl tar -xzvf file lpd lsmod modinfo mount -a mv gawk rpm -U rpm -e rpm -i rpm -V rpm -qa nice renice top dmesg fdisk -l ps [-]aux tcpdump, ethereal minicom strace, ltrace umount -a adduser /sbin/runlevel

Usage File system allocation in units of kilobytes Get, set ACLs Unbundle compressed TAR file Printer daemon List loaded kernel modules Mount all entries in fstab Move a directory Awk scripting language Update package Remove package Add package Verify package Query packages Start a process with a given priority Change the priority of a running process Report process statistics Error reporting tool Read disk label List all system processes Network packets sniffer Serial port access program Tracing Unmount entries in mtab Add a user account Show run level

368

Solaris to Linux Migration: A Guide for System Administrators

Solaris xterm/dtterm

Linux konsole/gnome-terminal

Usage GUI terminal program

Common Solaris and Linux commands

Table B-3 provides a quick reference for common commands between Solaris and Linux. Note: Although options and syntax might differ in some cases, the purpose of these utilities is the same.
Table B-3 Common commands and utilities apropos awk bzip2 chgrp col csh dd disable eject exec false fmt gencat gettext gzip hostname jobs lastcomm ldd arp bash cancel chmod comm csplit df dmesg enable exit fdformat format getconf gprof head iconv join ldapadd lesskey as batch captoinfo chown continue ctags diff domainname env expand fdisk fsck getent grep help init kdestroy ldapdelete let at break case chroot cp cut dig echo eqn export fgrep ftp getfacl groups history install kill ldapmodify link atq bunzip2 cat cksum cpio date dirname ed eval expr file function getopt gunzip host ipcrm killall ldapmodrdn ln atrm bzcat cd clear crontab dc dirs egrep ex factor finger fuser getopts gzexe hosted ipcs kpasswd ldapsearch locale

Appendix B. Commands and configuration files reference

369

localedef lp mail mknod mt nice paste ping pushd read rm script sftp split strace sync time trap typeset unexpand users whatis xargs ypwhich

logger lpq man mktemp mv nl patch pkill pwd readonly rmdir sdiff sh ssh strings syslogd timeout troff ul uniq vi whereis yacc zcat

login lpr mesg modinfo neqn nroff pathchk popd quota red rpcinfo sed shift ssh-add sty tar times true ulimit units view which ypcat zipinfo

logname lprm mkdir more netstat nslookup pax printenv ranlib renice rsh select sleep ssh-agent su tbl touch tset umask unlink vmstat whoami ypinit

logout lpstat mkfifo mount newgrp nsupdate perl profiles rcp return rup set sort sshd sum tee tput tsort umount unset wall whois ypmatch

look ls mkisofs mpstat nfsstat passwd pgrep ps rdist rlogin scp setfacl source ssh-keygen suspend tic tr tty uname uptime wc write yppasswd

370

Solaris to Linux Migration: A Guide for System Administrators

Appendix C.

UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)
This appendix contains the Table of Contents and Chapter below, excerpted from UNIX to Linux Porting: A Comprehensive Reference (ISBN 0131871099), reprinted courtesy of Pearson Education , Inc., Pearson Education, Inc., 2006. We provide this content to provide you with some technical perspective on the subject of application porting. Porting of applications and programs from the Solaris operating system platform to Linux is a broad technical topic that is outside the scope of this book. Nonetheless, completing required porting tasks will be a key part of any successful migration project. We are pleased to provide you with this introductory content about the subject of porting. The sample content is presented in this appendix in two parts: Table of contents Chapter 1 Porting Project Considerations

Copyright IBM Corp. 2006. All rights reserved.

371

Note: If you are interested in obtaining a copy of UNIX to Linux Porting: A Comprehensive Reference, you can receive a 35% discount on the publisher price by using the following URL and coupon code to order:
http://www.prenhallprofessional.com/mendoza

Coupon Code: MENDOZA-IBM

Table of contents
UNIX to Linux Porting: A Comprehensive Reference (ISBN 0131871099)

Authors:
Alfredo Mendoza Chakarat Skawratananond Artis Walker

Preface Chapter 1: Porting Project Considerations Chapter 2: Scoping Chapter 3: Analysis Chapter 4: Porting Solaris applications Chapter 5: Porting AIX applications Chapter 6: Porting HP-UX applications Chapter 7: Testing and Debugging Appendix A: Solaris to Linux Reference Tables Appendix B: AIX to Linux Reference Tables Appendix C: HP-UX to Linux Reference Tables Appendix D: Linux on POWER Appendix E: gprof helper

372

Solaris to Linux Migration: A Guide for System Administrators

Chapter 1 Porting Project Considerations

Application porting refers to the process of taking a software application that runs on one operating system and hardware architecture, recompiling (making changes as necessary), and enabling it to run on another operating system and hardware architecture. In some cases, porting software from one platform to another may be as straightforward as recompiling and running verification tests on the ported application. In other cases, however, it is not as straightforward. Some large applications that were written for older versions of operating systems and compiled with native compilers may not adhere to present language standards and will be more difficult to port, requiring the full attention of project management. This chapter supplements currently available project management materials and books about application porting projects. Topics such as how to use formalized requirements processes, how to better communicate between software developers, and how to practice extreme project management are topics that modern-day project managers are already well informed of. However, because software development is not exactly the same as porting and migrating software, a gap exists in current publications, a gap that this chapter addresses. The focus here is on software porting and migration process details and technical risks. All information in this chapter has been collected and gathered through countless porting engagements. This chapter presents best practices and methodologies to achieve high-quality ported applications. This chapter will appeal specifically to those involved in software porting and migration projects, especially project managers and application architects.

Software Application Business Process

Before embarking on a porting project, it is important to note which business processes within the application life cycle will be affected. Business processes currently in place must be modified to accommodate the ported application in terms of having a new platform to support, new test environment, new tools, new documents, and (most important) new customers to support and establish relations with. The three main areas of the application life cycle to consider that may affect the business process are development and testing, customer support, and software application distribution: Development and testing. Within the development and test organization, personnel must be trained to develop and test on Linux in terms of application programming interface (API) differences, development tools, debugging tools,

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

performance tools, third-party applications needed by the application to be ported, and performance tools. Customer support. Within the customer support organization, support personnel must be trained on Linux system administration, third-party applications needed by the ported application, installation and administration procedures, package maintenance options in the Linux environment, debugging tools and procedures, and anything else that needs to be learned to provide customer support for the ported application. Software application distribution. Within the software application distribution organization, sales and consultants must be trained in overall Linux features and capabilities.1 Personnel in the software distribution channels must be trained as trainers of the software application in a Linux environment. From a customer's perspective, they will also look for integration skills to integrate Linux within their IT infrastructure. Porting an application to Linux means modifying business organization processes that can be affected by a newly ported application. The three main areas should be considered and included in the overall project plan before starting the porting process.

The Porting Process

Developers involved in porting projects tend to follow similar steps when porting any software application. These steps include scoping, analyzing, porting, and testing. Each step in the process builds a foundation for the next step in the process. Each step, when done properly, makes the next step of the process easier to accomplish. Figure C-1shows the high-level steps taken during a porting project.

Figure C-1 High-level steps taken during a porting project

Different Linux distributions exist that may require separate training sessions for each.

374

Solaris to Linux Migration: A Guide for System Administrators

Scoping
Scoping is the step in which the project manager asks the porting expert2 and a domain expert3 to get together to identify the products, development, and test environment the application to be ported relies on. The key areas to identify during the scoping process include product dependencies, development environment components, build environment components, and test environment components: Product/software dependencies. Identifying which products the application to be ported relies on means determining which versions of database, middleware, and third-party libraries it uses. By knowing the products and versions, the porting expert can assess whether those products and versions are available for the Linux platform. Development environment components. Assessing the development environment includes identifying what programming language the application is written in. Applications written in a more recent programming language such as Java tend to port more easily, whereas applications written in C or C++ usually require more analysis and porting effort. Build environment components. Assessing the build environment includes making sure the build tools are available on Linux. Platform-dependent compiler and link flags used on the source platform must be investigated to determine whether equivalent flags exist on Linux. Some build environments may have dependencies on the source platform that it will require a little more effort to port to Linux. Test environment components. Identifying the test environment for the ported application leads to questions pertaining to the ownership of testing after the application is ported. Usually porting engineers will do unit testing for the part of the application they port and then hand it off to a testing group for more verification and systems tests. But who is the testing group? In most cases, the scoping step leads to identifying associated risks that will be assumed by the project as a whole when it begins. Some risks that can be identified in the scoping step are included in the following scenarios: Versions of the database, middleware, or third-party libraries are not available for the Linux platform. The application has some assembler routines that need to be translated to assembly language instructions on the Linux platform.

2 A software developer who is experienced in porting applications and has knowledge of source and target platforms as well as other third-party products the application uses. We also refer to this person as a porting engineer in this book. 3 A person who is knowledgeable about the application to be ported. This could be the application architect or the chief developer of the application.

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

The application was written using a set of APIs or a programming model unique to the source platform. This also includes assumptions about such things as word size or endian-ness. The application was written to draft-standard C++ semantics that relied on the source platform's native compiler framework. The test environment requires complicated client/server framework. The development environment requires third-party tools that need to be compiled or ported to the Linux platform. The distribution or installation method of the application requires facilities or tools unique to the source platform. Scoping is an involved step in the porting process that takes into account every new piece of information that can be learned from asking the right questionsquestions about documentation, packaging, and performance tuning, to name a few. These and others are mentioned in the sample porting questionnaire at the end of this chapter.

Analysis
There are two views in the analysis step of the porting process: a project management point of view, and a porting point of view. From a project management point of view, analysis is the step that assesses the various porting issues and risks identified in the preceding step and what impact they bring to the porting project as a whole. The analysis step involves the formulation of the project plan, which includes identifying scope and objectives, creating work schedules, procuring resources needed, and assigning roles within the project. Identification of scope and objectives defines the boundaries and responsibilities of the project manager and the members of the project team. Boundaries are clearly identified sets of work that need to be done for the project. For example, a simple statement such as Module A in application XYZ needs to be ported and tested on platform B can be a good start to defining the boundaries of the work. After boundaries have been identified, tasks for the porting work can be defined, leading to a work breakdown schedule.4 The work breakdown schedule helps define which tasks need to be done and whether those tasks can be done in sequence or in parallel. In addition, the work breakdown schedule identifies the resources needed. An overall schedule for the entire porting project results from identifying tasks and resources needed. From a porting point of view, analysis is a step in the porting process during which a porting engineer examines the application architecture in more detail.
4

A project management term used to denote work as an activity arranged in a hierarchical order that has tangible results otherwise referred to as a deliverable.

376

Solaris to Linux Migration: A Guide for System Administrators

The porting engineer begins to identify the APIs and system calls used in the application. In addition, the application is assessed as to its use of dynamic linking and loading, networking, threads, and more. This analysis feeds information back to the project manager that can be used to formulate more detailed tasks and accurate schedules.

Porting
Porting is the step in the process during which the porting engineers perform their assigned tasks. Depending on the application and the work breakdown schedule that resulted from the preceding step, porting engineers may only be able to work their assigned tasks in a serial fashion. Serial tasks are a product of tightly coupled applications. This means that some modules of the application are highly dependent on other parts of the application and that these modules can be worked on only when the part it depends on is ported. A prime example of this is the build environment. If the build environment was designed to build the whole application as a monolithic procedure, chances are all modules depend on common configuration files that need to be modified before any porting work can be performed. If the porting tasks are independent of each other, however, the work effort can be parallelized. Loosely coupled modules can be worked on separately and simultaneously by different porting engineers. A key example of this are shared libraries that can be built independently of each other, do not have common components, and exist for other modules to build on. Identifying which tasks can be done in parallel is important and should be done as part of the analysis step of the process. The porting engineers' task of compiling the code on the Linux platform includes identifying and removing architectural dependencies and nonstandard practices if possible. Identifying and removing architectural dependencies means heeding compiler errors and warnings produced at compile time and correcting them as needed. Removing architectural dependencies and nonstandard practices involves examining and changing the code to use more portable structures or coding standards. Experienced and quality-conscious porting engineers usually do the latter as they correct compiler errors and warnings. The porting effort includes porting the build environment to the Linux platform, too. This task should be identified during the scoping step. Although some build environments are portable, some are not. Making sure the build environment does not create any potential problems is a task easily overlooked and needs to be scoped and analyzed with extreme caution. After the application has been portedmeaning it compiles on the Linux platform without errorsthe porting engineer is expected to run unit tests on the application. The unit tests can be as simple as executing the application to

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

determine whether it produces any runtime problems. If runtime problems are detected, they are debugged and fixed before the application is handed to the testing group; the test group then runs more thorough tests on the ported software.

Testing
During testing, the assigned testers run the ported application against a set of test cases, which vary from simple execution of the application to stress-type tests that ensure the application is robust enough when run on the Linux platform. Stress testing the application on the target platform is where most porting problems related to architectural dependencies and bad coding practices are uncovered. Most applicationsespecially multithreaded applicationswill behave differently during stress testing on a different platform, in part because of different operating system implementations, especially in the threading area. If problems are found during testing, porting engineers are called in to debug and fix them. Some application porting may also include porting a test harness to test the application. Porting this test harness is a task that is also identified during scoping and analysis. Most of the time, the testers need to be trained on the application before they can test it. Learning about the application is a task that is independent of the porting tasks and can be performed in parallel. If bugs are found, they are fixed and the source recompiled; the testing process repeats until the application passes all testing requirements.

Support
When the port is complete, the development phase ends, and the support phase begins. Some of the porting engineers are retained to help answer any porting-specific questions the customer may have. In addition, the developers are expected to train the customer on configuring and running the application on the Linux platform. The typical support phase lasts from between 60 to 90 days after the portin which time, the porting engineers train and answer technical support and sales personnel questions regarding the newly ported application on the Linux environment. After the technical support and sales personnel have been trained in the newly ported product, the porting engineer's task is done.

Defining Project Scope and Objectives

Project scope is defined as the specific endpoints or boundaries of the project as well as the responsibilities of the project manager and team members.5 A clearly
5

Radical Project Management, Rob Thomsett, 2002, Prentice Hall

378

Solaris to Linux Migration: A Guide for System Administrators

defined project scope ensures all stakeholders (individuals and organizations involved in or that might be affected by project activitiesproject team, application architect, testing, customer or sponsor, contracts, finance, procurement, contracts and legal) share a common view of what is included as the objectives of the project. Clear project objectives/requirements lead to a better understanding of the scope of the project. During the analysis step in the porting process, customer objectives and requirements are gathered, and these objectives transform themselves to work breakdown structures and eventually project deliverables. Project objectives and requirements provide a starting point for defining the scope of the project. After all project objectives have been laid out, the project scope becomes more defined. One way to define the scope of a project is to list objectives that will and will not be included in the project. A requirements list from the customer is a good start. After the requirements have been gathered, the project manager and the technical leader review the list in detail. If there are questions about the requirements list, they should appear in the not to be included list of objectives, at least initially. The list is then reviewed by the customer, who may make corrections to the list. The list is revised until all parties agree that the requirements list presents the true objectives of the project. Again, it should be sufficiently detailed so that it lists requirements that will and will not be included in the project. Yes, it is really that important to include details that are beyond the scope of the project. Objectives that do not provide enough definition to the scope of the project may turn out to be a point of contention as the project draws closer to its release date. An example of this might be listing how pieces of the porting work will be delivered to the porting teamwill it be in phases or will it be in a single delivery? Will each delivery be considered a phase, or will there be smaller pieces of work within a phase? How many phases will make up the entire project? The list not only defines the project itself, it also sets the expectations for all stakeholders in the project. Here are the basic rules to follow when creating the project objectives list: 1. Define project scopes with as much detail as possible. 2. Confirm that all stakeholders agree to project scopes. 3. List unresolved items on the not-included/not-in-scope list until they are resolved. Table C-1 on page 380 shows an example of a simple in-scope and not-in-scope table for a sample porting project.

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

Table C-1 In-Scope and Not-in-Scope Objectives Project: Customer X Application Port to Linux In Scope Port Directory, Mail, and Order Entry modules to Linux to be divided up in three phases. Each module will be one phase. Functional verification tests on these modules will be conducted by porting team. Definition of functional verification tests and acceptance criteria will be defined in another section. Provide debugging assistance to customer developers during system verification tests. Not in Scope System tests of in-scope modules are not included in the porting project. Customer will be responsible for systems test. Packaging of application is not in scope. Fixing bugs found to be present in the original source code are not in scope. External customer documentation of ported modules is not in scope. Customer to produce external customer documentations.

Part of defining the objectives is listing the acceptance or success criteria. A consensus regarding the porting project success criteria must be reached by all stakeholders. Project success may mean passing a percentage of system tests on the Linux platform or passing a level of performance criteria set out by the systems performance groupthat is, if the project objectives are to run a specific set of tests or performance analysis, respectively. Regardless of how the project success is defined, all stakeholders must understand and agree on the criteria before the porting effort starts, if possible. Any changes to the criteria during the course of the porting cycle must be communicated to all stakeholders and approved before replacing the existing criteria.

Estimating
Many porting projects go over budget and miss schedules because risks were not managed properly. Risks play a major part in estimating the schedule as well as resources needed for porting projects. In an application porting project, such risks will come from different aspects that relate to application porting, including the following: Skill levels and porting experience Compiler used Programming language used Third-party and middleware product availability Build environment and tools Platform-dependent constructs

380

Solaris to Linux Migration: A Guide for System Administrators

Platform- and hardware-dependent code Test environment setup needed User interface requirements Depending on the application to be ported, each of these aspects will present varying levels of complexity and risks to the project. Assessing complexity and risk level help you determine whether they are manageable.

Skill Levels and Porting Experience

The most glaring difference between porting applications and software development is the skill set of the programmers. Although software application developers tend to be more specialized in their knowledge domain, software developers who do porting and migration need broader and more generalized skill sets. An application developer can be someone who is an expert in Java and works on a Windows development environment. Another may be a programmer specializing in database programming on a Sun Solaris operating system environment. On the other hand, engineers who port code are expected to be experts in two or more operating system platforms, programming languages, compilers, debuggers, databases, middleware, and the latest Web-based technologies. They are expected to know how to install and configure third-party database applications and middleware. Whereas application developers tend to be specialists, porting engineers need to be generalists. Application developers may work on an application for about 18 months (the typical development cycle), whereas porting engineers work on a 3to 6-month project cycle and are ready to start on a new port as soon as the last one is finished. Finding skilled porting engineers who fit the exact requirements of a porting project may be difficult at times. This is most true when the porting efforts require porting from legacy technology to newer ones.

Compiler
The compiler and compiler framework used by the source platform makes a big difference when porting applications to the Linux environment. If the compiler used in both the source and target platforms are the same, the task becomes easier. An example in this case is when the GNU compiler is used in both the source and the target platform. Besides the --g and --c flags, different brand of compilers use flags differently. Compiler differences become more difficult if the programming language used is C++. Another thing to consider is the version of the compilers. Source code compiled a few years back with older versions of compilers may have syntax that does not conform to present syntax checking by compilers (because of standards conformity). This makes it more difficult to port on different compilers because these compilers may or may not support backward compatibility for standards.

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

Even if the compilers support older standards, different compilers may implement support for these standards differently.

Third-Party and Middleware Product Availability

The complexity of the port increases when the application to be ported uses third-party and middleware products. Complexity increases even more when the versions of those products are not available for the target platform. In rare cases, some third-party products may not even be available on the target platform at all. In the past few years, middleware vendors such as IBM, Oracle, and Sybase and many more have ported their middleware products to Linux. They have made efforts to make their products available for companies that are ready and willing to make Linux their enterprise platform. This is partly why we are seeing more and more companies willing to port their applications to Linux.

Build Environment and Tools

The simpler the build environment, the less time the porting team takes to understand how to build the source code. More complicated build environments include multipass builds where objects are built with different compiler flags to get around dependencies between modules. After a first-pass build to build some modules, a second pass compiles yet more modules, but this time builds on top of the previously compiled modules (and so on until the build completes). Sometimes the build scripts call other scripts that automatically generate files on-the-fly based on nonstandard configuration files. Most of these files may be interpreted as ordinary files to be ported, but in fact, it is really the tools and the configuration files that need to be ported to work on the target module. These types of tools are mostly missed during assessment and analysis, which can result in less-than-perfect schedules. One build environment that came into fashion sometime in the early 1990s was the use of imake. Imake is a makefile generator intended to ease build environment portability issues. Instead of writing makefiles, one writes imakefiles. Imakefiles are files that contain machine-independent descriptions of the application build environment. Imake creates makefiles by reading the imakefile and combines it with machine-dependent configuration files on the target platform. The whole concept of architecting a build environment around the imake facility to make it portable is a noble idea. Of the most recent 50 porting projects done by our porting group, only a few were built around the imake facility. Unfortunately, all of them needed as many modifications to their own environment and imake files, negating the ease that it is intended to provide in the first place. And because imake was rarely used, anyone who needed to do a port needed to relearn the utility every time. Nowadays, makefiles are written for the open source make facility called GNU Make (or gmake, for short). Most makefile syntax used in different UNIX

382

Solaris to Linux Migration: A Guide for System Administrators

platforms is syntax that gmake can understand or has an easy equivalent to. Architecting applications around the gmake6 facility is the most accepted and widely used method in build environments today. Source code control is another aspect that must be considered. In the Linux environment, CVS is the most frequently used source code control environment. Other source code control environments are Subversion7 and Arch.8 A porting project needs to have source code control if there is a possibility of several porting engineers working on the same modules at the same time.

Platform-Dependent Constructs
When porting from UNIX platforms based on RISC architecture to x86-based platforms, a chance exists that application code needs to be assessed for byte-endian dependencies. For example, the application implements functions that use byte swapping for calculations or data manipulation purposes. Porting the code that uses byte-swapping logic to an Intel-based machine, which is little-endian architecture, requires that code be modified to adhere to little-endian semantics. In cases where byte-swapping logic is not changed between platforms, debugging becomes a chore because it is difficult to track where data corruption takes place when it happens several instructions before the application fails. Make sure to identify platform-dependent constructs in the scoping and analysis steps.

Platform- and Hardware-Dependent Code

Applications that require kernel extensions and device drivers to operate are the hardest to port. Kernel APIs for all platforms do not follow any standards. In this case, API calls, number of arguments, and even how to load these extensions into the kernel for a specific platform will be different. In most cases, new code that needs to run on Linux must be developed. One thing is for certain: Kernel code will usually, if not always, be written in C or platform-dependent assembler language.

Test Environment and Setup

When the port is done and the project scope includes system tests and verification, equipment required to test the code may contribute to the complexity of the port. If the application needs to be tested on specialized devices that need to be ordered or leased, this can add complexity to the project. If the application needs to be tested in a complicated clustered and networked environment, setting up the environment and scheduling the resource adds time to the project.
6

Some may favor Autoconf even though it may add some level of complexity compared to using just plain old Makefiles. 7 http://subversion.tigris.org 8 http://www.gnu.org/software/gnu-arch/

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

Testing may also include performance testing. Normally, performance tests require the maximum configuration setup one can afford to be able to load up the application to see how it scales in big installations. Porting of the application test harness needs to be included in the porting schedule. A test harnesses that includes both software tests and specialized hardware configurations need to be assessed as to how much time it will take to port and configure.

User Interface Requirements

Porting user interfaces can range from being easy to complex. User interfaces based on Java technology are easy to port, whereas user interfaces based on X11 technology are more complex. In this type of application, the code may be written in C or C++. Table C-2 is a summary of the various aspects involved. Each aspect has been ranked as easy, medium, or high complexity, depending on the technology used for each aspect.
Table C-2 Software Porting Risks and Complexities Aspects Compiler/programming Language Easy Java Shell scripts Use of nonportable syntax Original source code written to other standards of compiler Medium Complexity Use of C, COBOL, Fortran language High Complexity C++ using compiler-specific framework that supports varying levels of the C++ standard Source code written in languages other than Java or C Use of nonstandard compilers such as ASN or customer-built interpreters Use of assembler language Use of RPG language Use of third-party products or middleware None Supported and available on target platform Not supported on target platform Use of third-party tools written in C++

384

Solaris to Linux Migration: A Guide for System Administrators

Aspects Build environment and tools

Easy Simple make files

Medium Complexity Combination of make files and build scripts

High Complexity Use of noncommon build tools such as imake, multiple pass builds, on-the-fly-generated code modules Extensive use of kernel extensions and device driver code

Platform / hardware-dependent code

No platform- or hardware-dependent code

Platform/hardwaredependent code comes from third-party products already ported to target platform Client/server setup

Test environment and setup

Stand-alone

Networked, high availability, clustered Needs external peripherals to test such as printers, Fibre Channel-attached disks

User interface

Java based

Standard-based interface such as X11, CDE

Nonportable user interface such as custom interfaces using proprietary code

Experience shows that real efforts needed to port the application are uncovered during the first two or three weeks of the porting project. This is the time when the code is delivered to the project team and porting engineers get a first glimpse of the software application. In other cases, this will be the first time porting engineers will work on the application in a Linux environment. They begin to dissect the application components and start preliminary build environment setups. Within the first two to three weeks, the porting engineers will have configured the build environment and started compiling some modules using GNU-based tools and compilers. After two to three weeks, it is good to revisit the estimated time schedules to determine whether they need to be adjusted based on preliminary real experiences of porting the application into the Linux platform.

Creating a Porting Project Schedule

When creating the porting schedule, consider all risks, including technical and business-related issues. In terms of technical issues, common things such as

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

resource availability, hardware availability, third-party support, and Linux experience need to be considered. On the business side, issues such as reorganizations, relocations (moving offices), customer release dates, and business objective changes may affect the porting schedule as a whole. These technical and business issues will create several dependencies external to the porting project that may not be controlled; therefore, it is advisable to consider each external issue to mitigate risks accordingly. Creating schedules for porting projects is just like creating schedules for software development projects. Figure C-2 is an adaptation of a diagram from Radical Project Management by Rob Thomsett (Prentice Hall, 2002, p. 191). The figure shows that the average estimation error made before the project scopes and objectives are cleared up is +400 percent and 400 percent. As the objectives, scope, risks, and other technical and business issues are clarified, the estimated porting schedule comes closer to the real schedule.

+400%

Scoping

Analysis

Porting

Testing

Estimation Point #3 Estimation Point #2 -400%

Estimation Point #1
Figure C-2 Average estimation error made before scopes and objectives are cleared

Each transition to the next step in the porting process allows the project team to reassess the project estimates in terms of the schedule and resources. In essence, these transitions can and should serve not only as milestones for the project but also as checkpoints to reexamine the original estimates.

386

Solaris to Linux Migration: A Guide for System Administrators

Porting Process from a Business Perspective

The porting process is not just about porting the software application; the resulting ported application will need the necessary business and support structures to make it a successful business initiative. Although porting efforts are taking place, the stakeholders need to prepare other organizations that will provide support for the application. Organizations such as customer support and software distribution will require documentation and training on the application running on the Linux environment. Getting the necessary Linux training for personnel in the customer support and software distribution should be high on the priority list of objectives. As with any new product, new releases tend to generate the most questions from users and system administrators alike. Expect to answer system administration questions about Linux, too. Our experience shows that newly ported applications on a new operating environment will generate system administration questions and the use of the new operating system in three out of five support line calls. As system administration and installation questions begin to subside, more technical questions about the application will arise. Support organizations need access to live test machines to duplicate customer problems. These machines may differ from development or porting machines depending on the application and technical problems that will require instances of the application to be running for debugging purposes. From an overall project perspective, the need to train support personnel and the availability of physical resources to support the ported application will come at later stages of the porting project. As the technical aspects of the porting project near completion, the business aspects of the project pick up.

Annotated Sample Technical Questionnaire

This questionnaire, based on which other scoping questions can be formed, serves as a guide to the porting technical leader. The customer in this case is an internal or external organization that needs to port their application(s) to the Linux platform.

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

Platform-Specific Information
1. What is your current development platform for the application? The question is about the development platform used to develop the application. It does not assume that the development platform is the same platform the application is deployed on. This is asked in the next question. 2. What platform does the application currently run on? Porting engineers need to know the source or reference platform the application to be ported is using. 3. Has this application been deployed on any other platform than the development one? If yes, which version of the platform is it running on? Asking this question gives you a sense of the portability of the application if it has been ported to other platforms. A word of caution: Although an application may have been ported to other platforms, it may have been done on older versions of the platform. 4. Describe any hardware (that is, graphics adapters, or cards) used by your application and whether the required drivers are available on the Linux platform. Make sure any platform dependencies are available on Linux.

Application-Specific Information
1. Please describe your application and its architecture in detail. This is where the customer describes the application architecture. Have them include an architectural diagram if possible. All components in the application need to be described. This should also give you what type of application framework, if any, the application runs on. Most Java applications run on product-specific frameworks such as WebSphere or Weblogic. If they are written in C++, they may run on a CORBA framework, which means you may have to deal with a specific CORBA framework that may or may not exist on Linux. 2. What are the different components of your software? Please provide name and version numbers for each component. This will give you a sense of the breakdown of their applications. Being able to break down the application into different discrete components can mean that the porting work can be broken down into smaller independent tasks.

388

Solaris to Linux Migration: A Guide for System Administrators

3. Which of these pieces will need to be ported or not ported? Please include the version number? The customer needs to tell you what is in scope and what is not in scope. 4. What percentage of the application(s) to be ported is written in the following programming languages? a. Java b. C# c. C d. C++ e. Assembler f. Visual Basic (Microsoft) g. Shells (ksh, csh, perl, awk, others) Ascertains the complexity of the application by asking what languages they are using and what percentage of those are in use. 5. Please provide a rough estimate of the number of lines of code in the software, listed by language types. This is another way of asking Question 4. Asking questions in different light often brings up data points that contradict each other. This opens the door for open discussions that can result in better project scoping. 6. For Java apps: Does the application use the JNI9 to link native libraries? Please describe. Ascertains the complexity of the application to be ported. Most of the time, Java applications that are not 100 percent pure Java need platform-dependent routines that can only be handled in native language such as C. Be aware of such platform-dependent code because they take more time to port. 7. Does the application use kernel modules? If yes, please describe. Ascertain complexity of the application to be ported. Kernel modules and routines used by the application are nonportable. These will take more time to convert to equivalent routines in Linux. 8. Is it a 2D/3D graphics application? Please describe. Ascertains complexity of the application to be ported. Make sure compatible graphics toolkits and development tools are available in Linux, whether they are supplied in Linux by default or by other third-party vendors.
9

Java native interface

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

9. Does the application use UNIX pipes, message queues, shared memory, signals, or semaphores? Please describe. Most of these have standard UNIX interfaces that can be easily ported to Linux. Although the interfaces may be standard, implementation underneath the covers will differ. The catch is to make sure that the intended behavior remains the same when ported to Linux. 10.Is the application, or any of its components, multithreaded? If yes, which threads library is currently being used? Does the application rely on any proprietary threading priorities on your development platform? Depending on the source platform, multithreading interfaces can vary from standard to nonstandard. Linux supports several threading libraries, but the one that is standard in present and future Linux distributions is the Native Posix Threads Library (NPTL) implementation. NPTL is discussed in other sections of this book. The point is the closer the source implementation is to NPTL, the easier it is to port. 11.Does the application perform operations that assume specific byte storage order? Is this likely to be an issue during the port? This question relates to the endianess of the application. Most Linux ports will target the Intel platform, which is small endian, whereas most source platforms will be big endian. Nonportable code that assumes endian-specific characteristics will break when not ported correctly. Worse yet, the ported code will not exhibit errors during the porting phase. The problem usually crops up during system testing where it is harder to find. 12.Which compiler(s) and version is used in the development platform? a. GNU b. Java (what version) c. Platform (HP, AIX, Sun, Microsoft, NCR, AS/400, S390, True64) compiler? Which platform? d. Others (please specify) Ascertains complexity of the application to be ported. If the source application uses the GNU gcc or g++ compiler, it becomes easier to port to Linux because the native compiler for Linux is GNU gcc or g++. Applications that are developed on other platforms and are compiled in their native compilers tend to use native compiler semantics, which must be converted to GNU compiler semantics. C++ applications become harder to port than C applications when the application starts to use C++ features such as templates. Because some C++ standards are implemented differently by different compiler vendors, porting this type of code will take more time than simpler C++ code.

390

Solaris to Linux Migration: A Guide for System Administrators

13.In addition to the development environment, are there any dependencies on debugging tools such as memory leak debuggers, performance analysis tools, exception handling, and so forth? This goes back to scoping and dependencies. Third-party tools that may or may not exist in Linux need to be assessed. Who needs to provide for the license? Who is responsible for obtaining the tools? What will be the support structure if third-party support is needed? 14.Is it a socket-based application? If yes, does it use RPC? Please describe. Although Linux supports standard-based socket and RPC semantics, the intent is ascertain portability. Asking this question may bring to light nonportable architecture the customer may have implemented in the application. This question can also lead to questions on what setup is needed at the testing phase. 15.Does the application use any third-party software components (database tools, application server, or other middleware)? If yes, which ones? Every third-party software component adds complexity to the port. If any third-party software components are used, ask what version of the component is used and whether it is available on Linux. Third-party components may require extra time to learn and even configure or build if necessary 16.How is the application delivered and installed? Does it use standard packaging? Will the installation scripts need to be ported, too? A Linux standard packaging mechanism is RPM. RPM is discussed in other parts of the book. Ascertain whether the customer will need the packaging part of the application ported, too. 17.Is the application or any of its components currently in 64 bit? Will any component need to be migrated to 64 bit? With the advent of 64-bit platforms and operating systems, this question pertains to the level at which the application needs to run or be ported. Most 32-bit applications will port to a 64-bit environment without problems through the use of modern compilers. Today's compilers have become efficient at flagging syntax and semantic errors in the application at compile time, allowing the porting engineer to make necessary changes. The one consideration here is that it will require additional time for porting and debugging.

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

Database Information
1. What databases are currently supported? Please include version numbers. Almost all enterprise applications today require a database back end. Making sure the database for the application is available on Linux is important. Differences in database products and versions can add substantial porting effort to the project. 2. What database is the ported application expected to run with? In addition to Question 1 in this section, what database does the customer expect the ported application to run with on the Linux platform? 3. Does the application use any nonrelational or proprietary databases? Any proprietary database needs to be ported to Linux. Make sure the code to run the database is available and is part of the scoped porting work. 4. How does the application communicate with the database? a. Programming language(s) (for example, Java, C/C++, other) b. Database interface(s) (for example, ODBC, OCI, JDBC, etc.) Ascertains that programming languages and interfaces are available on Linux, whether or not supplied by third-party vendors. 5. Does the application require the use of extended data types (XML, audio, binary, video, and so on)? This information can be used to assess the skills needed by the porting group for porting the application.

Porting Project Time Schedule Information

1. What is the desired General Availability date for the application on the target platform? This question is asking whether any business objectives need to be considered when creating the porting schedule. 2. Has the porting of the application already started on the target platform? Sometimes there may have been efforts in the past to port the application to Linux. This may be helpful in assessing complexities and problems encountered on the port. 3. What is the estimated port complexity level (low, medium, or high)? Take the answer to this question with a grain of salt. There may be other factors present today that may have not been fully realized in previous porting efforts.

392

Solaris to Linux Migration: A Guide for System Administrators

4. What factors were considered in this complexity ranking? Any information from previous porting efforts needs to be assessed and compared to future porting efforts on the Linux platform. 5. If the application has been ported to another platform, how long did that port take? How many resources were dedicated to it? What problems were encountered? This question attempts to compare previous porting efforts to the Linux port. This is only useful if the porting engineer technical lead has previous porting experience on other platforms as well as Linux. 6. What is your rough estimate of the project porting time and resources required? The application or parts of it may have already been ported to other platforms, and knowing the time it took to port to those other platforms may be of some use. Experience and lessons learned from previous ports may come in handy. Knowing some of the lessons learned may help in avoiding problems when porting to Linux.

Testing-Specific Information
1. Please describe the acceptance testing setup. 2. What kind of networking and database setup would be required for unit testing? 3. How much testing will be required after porting (number of days, number of resources)? 4. Do you have established test scripts and application performance measurements? 5. Will benchmarks need to be run for comparison testing? 6. Is performance data available on current platforms? 7. When were the performance tests last executed? All testing-specific questions pertain to application software testing on the Linux platform. Asking these questions may bring out other issues related to porting test scripts and the software application test harness, which will add risks and time to the whole project. Pay close attention to the response to Question 1 of this section. Question 1 relates to the type of acceptance criteria that needs to be agreed on before porting starts. The acceptance criteria can be something like passing all tests suites from tests scripts or test programs that tests the newly ported application. When the acceptance criteria are met, the port is considered complete and formal QA tests can then be performed on the application.

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

Porting Project Execution Information

1. Please select one or more options, according to how you would like to proceed with this project. a. Technical assistance will be provided to the porting engineers as necessary. b. Customer will be responsible for acquiring third-party licenses and technical support. c. Other (please describe). Add other items in this part of the questionnaire that you need the primary customer to consider. Some issues may relate to staff training or testing the application. 2. What kind of hardware will be needed for the project? Consider asking this question to ascertain whether existing hardware may be used or whether extra hardware will be neededfor porting, testing, training, and support purposes if necessary. Although this questionnaire is comprehensive, it should not be the only basis for scoping. Scoping should also include actual examination of application source code when pertinent. Software application documentation needs to be examined to learn more about the application from a customer usage and point of view.

Summary
Project managers and porting technical leads need to consider multiple aspects of a porting project before embarking on the actual port itself. Careful investigation of these aspects has led us to many successful porting projects in the past. Because every porting project is different, these aspects may take on different shapes and forms. Even so, the underlying concepts and descriptions will remain the same. A porting project can be summarized as follows: Scoping. Ask questions, inspect the code if available, ask more questions and investigate the risks, and develop the schedule. Analysis. Ask more questions, investigate the risks, and create the schedule based on technical and business issues. Porting. Set up the build environment, compile, and unit test. Testing. Set up the test environment, test, and analyze performance. Go back to porting process if bugs are found.

394

Solaris to Linux Migration: A Guide for System Administrators

Training. Train support and sales personnel on the application and operating environment. Because complexities and risks associated with different technical aspects of the porting project may affect the overall schedule and budget, the transition point between each step of the porting process allows the project team to reassess its original estimates. Reassessing the original estimates helps set the proper expectations of all stakeholders early on before the real work of porting begins. As in all projects, the success of the project is not only defined by the successful delivery of the final product; it is also defined by how well the project was managed to the satisfaction of all parties involved. In the next chapters, we go through each step of the porting process. We present information about Linux that porting engineers can use as they go through the scoping, the analysis, the porting, and the testing phases of a porting project.

Appendix C. UNIX to Linux Porting: A Comprehensive Reference (table of contents and sample chapter)

396

Solaris to Linux Migration: A Guide for System Administrators

Appendix D.

Example: System information gathering script

This appendix describes the use of two versions of a sample system diagnostic script. The scripts are designed to gather detailed system configuration information. One version is based on the ksh shell and works for Solaris systems. The other version is based on the bash shell and works on Linux systems. Therefore, when considering the two scripts together, you have a demonstration of how to port tools such as this (ksh to bash) from Solaris to Linux. Both scripts are lengthy, so we do not present the source code in this appendix. We describe what they do and how to use them. You can download the script sources as part of the additional materials provided for this Redbook. See Appendix E, Additional material on page 401 for information about how to access this additional material. The scripts are designed to gather the following groups of information: Basic system information: Host name OS release Kernel information Platform

Copyright IBM Corp. 2006. All rights reserved.

397

 Number of CPUs Memory size And so on Kernel configuration information (for Linux only): Current kernel runtime settings Boot command line used Device and driver files Memory information and statistics Kernel modules And so on Operating system configuration information: Boot loader type and configuration (Linux only) Password and group files System-wide profile setting PAM configuration Inittab file RC files Default configuration files for various software components Syslog configuration Print queues Cron settings Packages and patches And so on Disk configuration information: Disk usage information Mount point ownership and permission settings File system table file Swap devices Physical disk information Disk partitions If SVM or LVM is used, display configuration information

398

Solaris to Linux Migration: A Guide for System Administrators

Network configuration information: NIC settings Host files Routers Name resolver configuration Services NFS exports Remote mounts Automounter configuration And so on System configuration information: For Solaris: prtdiag, prtconf, eeprom For Red Hat: hwconf file For Novell SUSE: hwinfo output Other software configurations: SSHD server configuration file SSH client configuration file sudo configuration Sample listing of processes that are currently running Not all available software components are included in the sample scripts. However, these samples should cover most of the common system components about which you might want to gather information. If you have other software components that you use and want to monitor using a tool such as this, you can adapt these samples as necessary. These scripts were originally designed for two purposes: As a snapshot tool for documenting the server configuration As a source of information for emergency purposes (system recovery) The scripts are designed to provide output in two types of file formats: Plain text format HTML format Key command line parameters for the script are: h, H, html, or HTML: Output in HTML format only

Appendix D. Example: System information gathering script

399

t, T, text, or TEXT: Output in text format only b, B, both, or BOTH: Output in HTML and text format -h or -?: Display usage information The default output format is text. The output location is the present working directory. These scripts need to be run with root authority in order to be able to access certain portions of the operating system information. The output file naming convention works as follows: <hostname>.sysinfo.YYYYMMDD.<output-type> An example set of output files is: fred.sysinfo.20051116.html fred.sysinfo.20051116.txt bart.testlab.com.sysinfo.20051116.html bart.testlab.com.sysinfo.20051116.txt After these output files have been generated, you might want to copy them to another system for safe keeping in the event that the system that generated the files needs to be recovered (using information saved in these files as a guide).

400

Solaris to Linux Migration: A Guide for System Administrators

Appendix E.

Additional material
This redbook refers to additional material that can be downloaded from the Internet as described here.

Copyright IBM Corp. 2006. All rights reserved.

401

Locating the Web material

The Web material associated with this redbook is available in softcopy on the Internet from the IBM Redbooks Web server. Point your Web browser to:
ftp://www.redbooks.ibm.com/redbooks/SG247186

Alternatively, you can go to the IBM Redbooks Web site at:

ibm.com/redbooks

Select Additional materials and open the directory that corresponds with the redbook form number, SG247186.

Using the Web material

The additional Web material that accompanies this redbook includes the following file: File name SG247186-getsysinfo.zip Description ZIP file containing getsysinfo sample system diagnostic scripts.

How to use the Web material

Create a subdirectory (folder) on your workstation, and unzip the contents of the Web material ZIP file into this folder. We provide detailed information about how to use the sample getsysinfo scripts in Appendix D, Example: System information gathering script on page 397.

402

Solaris to Linux Migration: A Guide for System Administrators

Related publications
The publications listed in this section are considered particularly suitable for a more detailed discussion of the topics covered in this redbook.

IBM Redbooks
For information about ordering these publications, see How to get IBM Redbooks on page 411. Note that some of the documents referenced here may be available in softcopy only. Accounting and Monitoring for z/VM Linux guest machines, REDP-3818 Advanced POWER Virtualization on IBM System p5, SG24-7940 IBM Eserver BladeCenter, Linux, and Open Source: Blueprint for e-business on demand, SG24-7034 IBM Eserver xSeries and BladeCenter Server Management, SG24-6495 IBM Eserver zSeries 890 Technical Introduction, SG24-6310 IBM Eserver zSeries 900 Technical Guide, SG24-5975 IBM Eserver zSeries 990 Technical Guide, SG24-6947 IBM System z9 109 Technical Guide, SG24-7124 IBM System z9 and Eserver zSeries Connectivity Handbook, SG24-5444 Linux Client Migration Cookbook A Practical Planning and Implementation Guide for Migrating to Desktop Linux, SG24-6380 Linux for IBM System z9 and zSeries, SG24-6694 Linux for S/390, SG24-4987 Linux Handbook A Guide to IBM Linux Solutions and Resources, SG24-7000 Linux on IBM Eserver i5 Implementation Guide, SG24-6388 Linux on IBM Eserver zSeries and S/390: Best Security Practices, SG24-7023 Linux on IBM Eserver zSeries and S/390: Building SuSE SLES8 Systems under z/VM, REDP-3687 Linux on IBM Eserver zSeries and S/390: Cloning Linux Images in z/VM, REDP-0301 Linux on IBM Eserver zSeries and S/390: Distributions, SG24-6264

Copyright IBM Corp. 2006. All rights reserved.

403

Linux on IBM Eserver zSeries and S/390: ISP/ASP Solutions, SG24-6299 Linux on IBM Eserver zSeries and S/390: Large Scale Linux Deployment, SG24-6824 Linux on IBM Eserver zSeries and S/390: Performance Measurement and Tuning, SG24-6926 Linux on IBM Eserver zSeries and S/390: System Management, SG24-6820 Linux on IBM zSeries and S/390: High Availability for z/VM and Linux, REDP-0220 Linux on IBM zSeries and S/390: Securing Linux for zSeries with a Central z/OS LDAP Server (RACF), REDP-0221 Linux on IBM zSeries and S/390: Server Consolidation with Linux for zSeries, REDP-0222 Linux System Administration and Backup Tools for IBM Eserver xSeries and Netfinity, SG24-6228 Linux with zSeries and ESS: Essentials, SG24-7025 Networking Overview for Linux on zSeries, REDP-3901 Printing with Linux on IBM Eserver zSeries Using CUPS and Samba, REDP-3864 Technical Introduction: IBM Eserver zSeries 800, SG24-6515 The IBM Eserver BladeCenter JS20, SG24-6342 Tuning Red Hat Enterprise Linux on IBM Eserver xSeries Servers, REDP-3861 Tuning SUSE LINUX Enterprise Server on IBM Eserver xSeries Servers, REDP-3862 Up and Running with DB2 for Linux, SG24-6899 z/VM and Linux on zSeries: From LPAR to Virtual Servers in Two Days, SG24-6695

Other publications
These publications are also relevant as further information sources: Mendoza, et al., UNIX to Linux Porting: A Comprehensive Reference, Prentice Hall, 2006, ISBN 0131871099 Johnson, S., Performance Tuning for Linux Servers, IBM Press, 2005, ISBN 013144753X

404

Solaris to Linux Migration: A Guide for System Administrators

Kuo, P. and N. Rahimi, SUSE LINUX Enterprise Server 9 Administrator's Handbook, Novell Press, 2005, ISBN 067232735X Linux Standard Base Team, Building Applications with the Linux Standard Base, IBM Press, 2004, ISBN 0131456954 Nemeth, E. and G. Snyder, Linux Administration Handbook (2nd Edition), Prentice-Hall PTR, 2006, ISBN 0131480049 Petersen, R., Red Hat Enterprise Linux & Fedora Core 4: The Complete Reference, McGraw-Hill Osborne Media, 3 edition, 2005, ISBN 0072261544 Siever, E., et al., Linux in a Nutshell, O'Reilly Media, Inc., 5 Edition, 2005, ISBN 0596009305

Online resources
These Web sites and URLs are also relevant as further information sources: IBM developerWorks open source section
http://www.ibm.com/developerworks/opensource

IBM Tivoli Workload Scheduler

http://www.ibm.com/software/tivoli/sw-bycategory/

What's new in Performance Toolkit for VM in Version 5.1.0

http://www.vm.ibm.com/library/gm130637.pdf

IBM Eserver BladeCenter Management Module Users Guide

http://www.ibm.com/pc/support/site.wss/MIGR-45153.html

Cisco Systems Intelligent Gigabit Ethernet Switch Module for the IBM Eserver BladeCenter Installation Guide
http://www.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-57858

IBM Eserver BladeCenter Management Module Command-Line Interface Reference Guide

http://www.ibm.com/support/docview.wss?uid=psg1MIGR-54667

RMF PM for Linux

http://www.ibm.com/servers/eserver/zseries/zos/rmf/rmfhtmls/pmweb/pmlin.html

z/VM Security and Integrity Resources

http://www.vm.ibm.com/security

Linux virtualization on POWER5: A hands-on setup guide

http://www.ibm.com/developerworks/edu/l-dw-linux-pow-virutal.html

Related publications

405

IBM Eserver xSeries server platforms

http://www.ibm.com/servers/eserver/xseries

IBM BladeCenter server platform

http://www.ibm.com/servers/eserver/bladecenter

IBM LPAR Validation Tool

http://www.ibm.com/servers/eserver/iseries/lpar/

IBM Systems Hardware Information Center

http://publib.boulder.ibm.com/infocenter/eserver/v1r3s/index.jsp

Management Processor Command Line Interface (MPCLI)

http://www.ibm.com/pc/support/site.wss/MIGR-54216.html

IBM Director
http://www.ibm.com/pc/support/site.wss/MIGR-57057.html

z/VM publications
http://www.vm.ibm.com/pubs/

IBM Eserver zSeries and IBM System z

http://www.ibm.com/servers/eserver/zseries/

Linux Documentation Project, a collection of Linux documentation and How-To guides about almost every topic related to installing, configuring, or using Linux
http://www.tldp.org

Red Hat Enterprise Linux documentation

http://www.redhat.com/docs/

SUSE Linux Enterprise Server documentation

http://www.novell.com/documentation/sles9/

SUSE Linux Enterprise Server Support

http://www.novell.com/support/products/linuxenterpriseserver/

Red Hat Package Manager (RPM)

http://www.rpm.org

SUSE Linux portal

https://portal.suse.com

Red Hat Network (RHN)

https://rhn.redhat.com

406

Solaris to Linux Migration: A Guide for System Administrators

Common UNIX Printing System (CUPS) information

http://www.cups.org

Printer compatibility and other printing information

http://www.linuxprinting.org

Linux Assigned Names And Numbers Authority (LANANA), which maintains the Linux Device List
http://www.lanana.org

Web-Based Enterprise Management (WBEM) and the Common Information Model (CIM)
http://www.dmtf.org

Simple Network Management Protocol (SNMP)

http://www.net-snmp.org

Linux Assigned Names And Numbers Authority (LANANA), which maintains the Linux Device List
http://www.lanana.org

GNU General Public License

http://www.gnu.org/copyleft/gpl.html

Linux Standard Base

http://www.linuxbase.org

Free Software Foundation

http://www.fsf.org

Open Source Initiative

http://www.opensource.org

Knoppix
http://www.knoppix.com

Red Hat SELinux Guide

http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/selinux-guide/

NSA SELinux information page

http://www.nsa.gov/selinux

GRand Unified Bootloader

http://www.gnu.org/software/grub

SYSLINUX
http://syslinux.zytor.com

Related publications

407

SUSE Linux Portal: How to Setup/Use Multipathing on SLES

http://portal.suse.com/sdb/en/2005/04/sles_multipathing.html

Red Hat device-mapper-multipath enhancement update

https://rhn.redhat.com/errata/RHEA-2005-280.html

Filesystem Hierarchy Standard

http://www.pathname.com/fhs/

Linux LVM How-To

http://www.tldp.org/HOWTO/html_single/LVM-HOWTO/

VERITAS Software (now Symantec)

http://www.veritas.com

Information about the cdrecord command

http://freshmeat.net/projects/cdrecord/

Mochel, P., The sysfs file system paper, presented in Volume 1 of the 2005 Ottawa Linux Symposium proceedings
http://www.linuxsymposium.org/2005/

Quagga Routing Suite

http://www.quagga.net

Distributed Management Task Force (DMTF)

http://www.dmtf.org

Internet Engineering Task Force (IETF)

http://www.ietf.org

Web-Based Enterprise Management (WBEM)

http://www.dmtf.org/standards/wbem

OpenPegasus CIM Server

http://www.openpegasus.org

OpenWBEM
http://www.openwbem.org

Net-SNMP package
http://www.net-snmp.org/

SNMP tutorials
http://www.simpleweb.org/

Common UNIX Printing System (CUPS)

http://www.cups.org/

408

Solaris to Linux Migration: A Guide for System Administrators

LinuxPrinting.org Web site

http://www.linuxprinting.org

OpenLDAP
http://www.OpenLDAP.org/

Rsync snapshot article

http://www.mikerubel.org

Advanced Maryland Automated Network Disk Archiver (AMANDA)

http://www.amanda.org/

Sun Security Resources

http://sunsolve.sun.com/pub-cgi/show.pl?target=security/sec

Red Hat Support Resources

https://www.redhat.com/apps/support/

SUSE Online Security Support

http://www.novell.com/linux/security/securitysupport.html

Yet Another Solaris Security Package (YASSP)

http://www.yassp.org/

Bastille Linux
http://www.bastille-linux.org

Security Enhanced Linux (SE-Linux)

http://www.nsa.gov/selinux/

Very Secure FTP Daemon, or vsftpd

http://www.vsftpdrocks.org

vsftpd FTP server

http://vsftpd.beasts.org

UNIX IP Stack Tuning Guide

http://www.cymru.com/Documents/ip-stack-tuning.html

OpenSSH
http://www.openssh.com

Using DVDs with Linux

http://dvd.sourceforge.net

Linux hotplugging
http://linux-hotplug.sourceforge.net

Related publications

409

Submount: Linux Removable Media Handling System

http://submount.sourceforge.net

K3b application
http://k3b.sourceforge.net

Linux Channel Bonding

http://sourceforge.net/projects/bonding

Linux Class-based Kernel Resource Management

http://ckrm.sourceforge.net/

Standards Based Linux Instrumentation for Manageability (SBLIM) project

http://sblim.sourceforge.net

OProfile profiling suite

http://oprofile.sourceforge.net

High Availability Linux project

http://linux-ha.org

Open Cluster Framework (OCF)

http://opencf.org

The Bash shell

http://www.gnu.org/software/bash/

Linux Kernel Crash Dump utility

http://lkcd.sourceforge.net

Memtest86 - A Stand-alone Memory Diagnostic

http://www.memtest86.com

Latest BIOS and firmware code

http://www.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-4JTS2T

Driver matrixes
http://www.ibm.com/pc/support/site.wss/DRVR-MATRIX.html

YaBoot
http://penguinppc.org/bootloaders/yaboot/

Linux for S/390 project

http://www.Linuxvm.org/

410

Solaris to Linux Migration: A Guide for System Administrators

How to get IBM Redbooks

You can search for, view, or download Redbooks, Redpapers, Hints and Tips, draft publications and Additional materials, as well as order hardcopy Redbooks or CD-ROMs, at this Web site:
ibm.com/redbooks

Help from IBM

IBM Support and downloads
ibm.com/support

IBM Global Services

ibm.com/services

Related publications

411

412

Solaris to Linux Migration: A Guide for System Administrators

Index
Symbols
#cp 327, 336 $cp 336 /bin/ash.static 212 /dev/hvcs0 299 /etc/auto.master 63 /etc/auto_master 63 /etc/ftpd/ftpusers 192 /etc/ftpusers 192 /etc/group 160 /etc/inetd.conf 189 /etc/inittab 133 /etc/issue 202 /etc/ldap.conf 164 /etc/motd 202 /etc/nsswitch.conf 164 /etc/printcap 155 /etc/racoon/racoon.conf 199 /etc/securetty 31, 196 /etc/skel/* 163 /etc/sysctl.conf 192 /etc/xinetd.d/ 190 /home 163 /lib/libipsec.so 198 /media/cdrom 94 /media/floppy 94 /proc 57, 89 /proc/mdstat 227 /sbin/init 133 /sbin/racoon 198 /sbin/setkey 198 /sys 90 adding a device 92 Advanced Intelligent Tape 99 Advanced Maryland Automated Network Disk Archiver 185 advanced system installation 29 Advanced System Management (ASM) 243 Advanced System Management Interface 259 agetty 104 AIT 99 AMANDA 185 Apache Web Server 5 arp 231 ASET 189 ASMI 259 AutoFS 62 AutoFSCK 62 automated restart 339 automated software management 81 automount 6263 automountd 6263 autoyast 37

B
backup 182 badblocks 225 Baseboard Management Controller 239 bash 212 bash parameters 335 basic system installation 24 Bastille Linux 188 best practice 309, 311, 321, 331, 333 bgpd 115 BIOS 127, 238 blade 239 BladeCenter 237, 244, 247 BladeCenter boot sequence 272 BladeCenter LAN Switch I/O Modules 268 BladeCenter Management Module 239, 244, 264 configuring 265 BladeCenter naming 271 block device 89 BMC 240 bonding 117

Numerics
8 mm tape 99

A
accept (print jobs) 154 access control list 60 access control list (ACL) 194 acl 60 addbadsec 55

Copyright IBM Corp. 2006. All rights reserved.

413

boot command 277, 326 boot configuration files 132 boot loader 128 boot loaders 126 boot source 127 cdrom / dvd 127 floppy and various USB devices 127 hard disk 127 network 127 boot troubleshooting 216 booting 126 default boot 126 emergency boot 127 interactive boot 126 interactive start of services 127 reconfiguration boot 126 recovery boot 127 single user 126 verbose boot 126 Bootparams 136 br0 298 brctl 298 bridge-down.sh 297 bridge-up.sh 297 bridge-utils 296297 bridging 117 BSM 189 bunzip2 182 business intelligence 16 bzcat 182 bzip2 182 bzip2recover 182 bzless 183

C
CacheFS 5657 cancel (print jobs) 154 Capacity on Demand 249 CD 96 cda 98 cdda2wav 98 cdparanoia 98 cdrecord 97 CD-ROM 94 bootable 18 cdrw 97 central processing units (CPUs) 312 central processor (CP) 312

central storage 313 cfgadm 93 chacl 194 channel 313, 319 channel path 314 channel-path identifier (CHPID) 314 character device 88 chccwdev 336 chgrp 223 chmod 223 chown 223 CHPID 314, 318 chroot 230 CIM object manager 147 CIM server 147 Cisco Systems Intelligent Gigabit Ethernet Switch Module 271, 274 CKRM 141 Class-based Kernel Resource Management 141 client partition profile 291 configuring Linux 302 cluster 206 cmsfscat 336 cmsfsck 336 cmsfscp 336 cmsfslst 336 cmsfsvol 336 commercial distributions 6 Common UNIX Printing System (CUPS) 152 Companion CD 183 compiling patches 82 compress 182 compression tools 182 console command 277 console for the JS20 blade 273 Control Program (CP) 320 control unit 313 core files 216 cpint 336 cpio 182 crash 219 crash dumps 218 cryptographic hardware 341 csh 212 cupsd 155

D
DASD 310, 334

414

Solaris to Linux Migration: A Guide for System Administrators

DASD Dump/Restore (DDR) Facility 340 dasdfmt 337 dasdview 337 DAT 99 database serving 15 dbginfo.sh 335, 337 dd 182 Debian 6 debugging tools 335 Dedicated Service Tools 251 dependencies 79 dependency 80 developerWorks 83 devfsadm 52 device access 89 device configuration information 89 device drivers 92 install 92 DFSMSdss 340 DHCP 119, 136 Linux 119 /etc/dhcpd.conf 119 /etc/init.d/dhcpd 119 /etc/sysconfig/dhcpd 119 /var/lib/dhcpd.leases 119 Solaris 119 /etc/inet/dhcpsvc.conf 119 dhcpconfig 119 dhcpmgr 119 dhtadm 119 in.dhcpd 119 pntadm 119 dhcpconfig 119 dhcpmgr 119 dhtadm 119 diagela 303 dig 231 Digital Audio Tape 99 Digital Linear Tape 99 disable (printers) 154 disabling rc scripts 134 disk drives 315 disk partition 59 disk partitioning 53 disk recognition 52 disk slice 59 Disk-Based File System Management 55 diskdump 219 diskdumpctl 219

diskdumpfmt 219 diskettes 94 disks 52 disks and disk partitioning 52 displaying tape drive status 101 Distributed Data Server (DDS) 331332 Distributed Management Task Force 147 distribution 4 distributions 5 DLT 99 dmesg 91, 221, 232 DMTF 147 DNS 120 DST 251 dump 183, 185 dumpadm 218 DVD 94, 96 dvd+rw-tools 97 dvdrecord 98 dynamic logical partitioning 253

E
e2fsck 5556 eject 96 enable (printers) 154 Enterprise Resource Planning (ERP) 15 env -T command 277 Epson 157 ESC/P printers 157 ESCON 314 ethereal 231 ethtool 231 expanded storage 313 ext2 54, 62 ext3 5556, 59, 62 extended user attributes 60

F
FCON/ESA 332 FCP 315 fdasd 337 fdformat 96 fdisk 5355 Fibre Channel Protocol (FCP) 315 FICON 314 file system dump and restore 183 file system journaling 59 file system snaphot 185

Index

415

file systems 178 file/print infrastructure 14 Filesystem Hierarchy Standard 60 find 222 firewall 203 chains 203 policies 203 firewalls 202 firmware 238 floppy diskette 96 fmthard 53 format 53 free 332, 335 free command 59 Free Software Foundation 5 fsck 5556 fsck -f 225 fsck -y 225 fsck.ext3 5556 fsck.reiserfs 56 fssnap 184185 FTP 191 fuser 232

H
halt 135 Hardware Management Console 249 hcp 327329, 337 Heartbeat 207, 340 Helsinki University of Technology 197 high availability 206, 338 HiperSockets 316, 341 HMC 249 hosted installation 248 hot-plug 52, 93, 328 hotplug command 90, 93 hot-plugging 93 HP printers 157 HSFS 55 hsfs 56 hvcs 299 hwbrowser 92, 233 hwconf 91 hwinfo 92 Hypervisor Virtual Console 299

I
IA32 238 IBM Director 240 IBM Machine Reported Product Data database 303 IBM VIO Server 285 IBM virtual SCSI server module 300 ibmvscsis 300 IDE 52 IEEE POSIX 10 IEEE POSIX 1003.1-2001 7 ifconfig 111, 335 ifenslave 117 IFL 310, 312 IKE 198 imap4 122 inetd 117 /etc/default/inetd 118 /etc/inetd.conf 118 /etc/services 118 ENABLE_CONNECTION_LOGGING 118 ENABLE_TCPWRAPPERS 118 info 8 init 135 initrd 128 inittab 102 insmod 86

G
gdb 233 General Public License (GPL) 45 getfacl 194 gfloppy 96 gfxmenu 30 GID 165 GNOME desktop 5 GNU 45 group administration 160 growfs 6667 growisofs 97 GRUB 126 config file 128 guest LAN 317 guest operating system 320 gunzip 182 gzcat 182 gzexe 182 gzgrep 182 gzip 182 gznew 182

416

Solaris to Linux Migration: A Guide for System Administrators

install automation 29 Linux client from Solaris server 47 PXE 47 Red Hat Kickstart 32 remote display 31 serial console 29 software bundles 26 Solaris client from Linux server 50 Solaris Jumpstart 32 SUSE AutoYaST 37 vnc 31 installation automation 29 installation methods 320 Integrated Virtualization Manager 249 Internet Engineering Task Force 147 Internet Printing Protocol (IPP) 152 inter-user communication vehicle 317 iostat 232, 332, 335 ip 231 IP masquerading 202 IP redirects 193 IP routing 114 Linux 115 /etc/sysctl.conf 115 bgpd 115 net.ipv4.ip_forward 115 ospf6d 115 ospfd 115 quagga 115 radvd 115 ripd 115 ripngd 115 Solaris 114 /etc/defaultrouter 115 /etc/gateways 115 /etc/notrouter 115 /usr/sbin/in.rdisc 115 ndpd 115 RDISC 114 RIP 114 ripngd 115 IP stateful firewalling 123 ipcs 232 iPlanet Directory Server 163 IPSEC 198 IPSec 198199 IPSec and IKE 116 iptables 203, 341

IPv4 108 Linux 109 /etc/bootparams 109 /etc/hosts 109 /etc/nsswitch.conf 109 /etc/protocols 110 /etc/resolv.conf 109 /etc/services 110 ifcfg-interface-id-mac-address 109 ifcfg-interface-name 109 ifconfig 111 network profiles 110 sysctl 110 Solaris 108 /etc/bootparams 108 /etc/defaultdomain 108 /etc/defaultrouter 108 /etc/dhcp.interface 108 /etc/ethers 108 /etc/hostname.interface 108 /etc/inet/hosts 108 /etc/inet/protocols 108 /etc/inet/services 108 /etc/netmasks 108 /etc/nodename 108 /etc/nsswitch.conf 108 /etc/resolv.conf 108 dhcpagent 108 ifconfig 111 ndd 110 IPv6 Linux 113 ifcfg-interface-id-mac-address 114 ifcfg-interface-name 113 Solaris 113 /etc/hostname6.interface 113 /etc/inet/ipnodes 113 iso9660 56 ITEF 147 IUCV 317

J
Jaz drive 94 JetDirect 153 journal mode 60 journaling 59 JS20 247 jumpstart 32

Index

417

K
K3b 97 KDE desktop 5 Kerberos 199 kernel 45, 128 kernel.core_pattern 217 kernel.core_uses_pid 217 kickstart 32 kill 232 Knoppix 18 ksh 212 ksymoops 233 kudzu 105

L
LANANA 88 lcrash 220 LDAP 120, 163 ldapadd 164 ldapcompare 164 ldapdelete 164 ldapmodify 164 ldapmodrdn 164 ldappasswd 164 ldapsearch 164 ldapwhoami 164 ldd 232 Lexmark 157 librtas 303 LILO 126 Linus Torvalds 45 Linux Assigned Names And Numbers Authority 88 Linux Device List 88 Linux on POWER installation 278 Novell SUSE 280 Red Hat 279 Linux Standard Base (LSB) 9 Linux Virtual Server (LVS) 340 Linux-HA 206 live CD 18 LKCD 219 LOADLIN 126 locally attached printers 153 locate 222 logadm 221 logging 193 logical disk devices Linux 87

Solaris 87 logical partition 54 logical partitioning 248 logical volume 68 logical volume groups 177 logical volume management 66 logical volumes 177 logs 220 LOM 30 loopback 57, 339 lp 154 lpadmin 154 LPAR 248, 309, 311, 318319, 333, 341 lpc 154 LPD 152 lpfilter 154 lpforms 154 lpget 154 lpinfo 155 lpmove 154 lpoptions 155 lppasswd 155 lpq 154 lpr 154 lprm 154 lpsched 154 lpset 154 lpshut 154 lpstat 154 lpsystem 154 lpusers 154 lscss 337 lsdasd 337 lsmod 86 lsof 232 lspci 91, 233 lsqeth 337 lsraid 67 lstape 337 lsusb 91 lsvpd 303 ltrace 232, 335 LUN number 52 lvchange 65, 67 lvcreate 6566 lvdisplay 67, 72 lvextend 67 LVM 54 LVM2 66, 68

418

Solaris to Linux Migration: A Guide for System Administrators

lvm2 66 lvmdiskscan 68 lvreduce 6566 lvremove 65 lvscan 68

M
mail server 15 mail services 122 major numbers 87 MAKEDEV 86 man 8 Mandrake 6 Master Boot Record (MBR) 54 MBR 127 mdadm 6567 mdb 217218 memtest86 233 messaging 15 metaclear 65, 67 metadb 67 metadetach 6567 metadevadm 67 metahs 67 metainit 65, 67 metaoffline 67 metaparam 67 metarecover 67 metarename 67 metareplace 67 metaroot 67 metaset 67 metastat 67 metasync 67 metattach 65, 67 mgetty 104 MIB 149 microcode 303, 312, 338339 Micro-Partitioning 252 migration planning 11 mii-tool 231 mingetty 104 MINIX 4 minor number 87 mk2fs 55 mkdump 337 mke2fs 56 mkfs 5556

mkfs.ext3 5556 mkfs.reiserfs 5556 mkinitrd 329 mkisofs 97 mkraid 65 mkreiserfs 56 mkswap 5859 modem 103 modinfo 86, 93 modload 86 modprobe 93 module stacking 196 modunload 86 monitoring 175 monolithic installation 248 mount 57 mpstat 332 msdos 56 mt 101 mtools 96 mtx 101 multichip module (MCM) 312 multipath devices 53 mzip 96

N
NAT masquerading 202 National Security Agency (NSA) 189 Nautilus File Manage 95 ndd 192 ndpd 115 netdump 219 netdump-server 219 Netfilter 202 net-snmp 150 netstat 231, 335 NetWare 153 network boot services 136 network booting 135 Network File System (NFS) 58, 62 Network File Systems 58 network printing 153 network troubleshooting 230 network trunking 116 newfs 55 newsyslog 221 NFS 57, 121, 137 NIS 121

Index

419

NIS (and NIS+) 164 NIS and NIS+ 121 NIS+ 121 non-commercial distributions 6 Novell 6 nslookup 231 NSS (name switch service) 164 NTP 120

O
OpenFirmware 283 OpenLDAP 163 OpenPegasus 148 OpenPROM 126 OpenSSH 198 OpenWBEM 149 operating system 4 opreport 233 OSA-2 317 OSA-Express 317 ospf6d 115 ospfd 115

P
pack 182 package 76 package distribution 80 package management 76 Packages 76 paging 313 PAM module types 195 pam_login.so 197 pam_nologin.so 197 pam_securetty.so 196 parted 5355 partition 53 partitions 52 passwd 230 patch 78 Patch activation 82 patch cluster 80 Patch Manager 81 patchadd 79 patches 78 patching 79 PatchPro Expert 81 PatchPro Interactive 81 patchrm 79

pax 182 pcat 182 PCFS 55 pcfs 56 PCL printers 157 pcsf 56 pdksh 213 performance 178 Performance Toolkit for VM 331, 335 pgrep 232 ping 231, 335 pkgadd 76 pkgchk 76 pkginfo 76 pkgparam 76 pkgrm 76 pkill 232 Pluggable Authentication Modules (PAM) 194 pmadm 105 pntadm 119 pop3 122 port initialization 104 portmap 226 POSIX threads 7 POST 128 postfix 122 PostScript printing 157 power command 277 poweroff 135 ppc64-utils 303 PPD (PostScript Printer Description) 157 PReP boot partition 279, 281 primary partition 54 print server 319 printer 319 printer types 156 printing subsystem 152 process accounting 146 processes 179 processing units (PU) 312 proof of concept 331 proof of technology 331 prtvtoc 53 ps 231, 335 Public Domain Korn shell 213 pvcreate 65, 67 pvdisplay 68, 71 pvs 69 pvscan 69

420

Solaris to Linux Migration: A Guide for System Administrators

PXE 47

Q
QDIO 310 qetharp 337 qethconf 337 quorum 68 quota 145

R
racoon keying daemon 198 radvd 115 RAID 54, 66, 68 raidhotadd 67, 227 raidhotremove 67 raidreconf 6667 raidstart 65 raidstop 65 raidtools 66 ramfs 58 RARP 136 raw device 89 raw disk slices 52 RDISC 114 reboot 135 Red Hat 6, 198 Red Hat Network (RHN) 8081 Redbooks Web site 411 Contact us xix ReiserFS 60 Reiserfs 54 reiserfs 5556, 59 reiserfsck 56 reject (print jobs) 154 remote display installation 31 Remote Supervisor Adapter II 239 remote system management 147 removable media 94 access 95 formatting 96 supported types 94 reset command 277 resize_reiserfs 6667 resize2fs 6667 resources 139 restore 183 REXX 340 RHEL install boot options 280

RIP 114 ripd 115 ripngd 115 rksh 212 rmformat 96 rmmod 86 Road Warrior server 199 IPSEC 199 root password recovery 229 route 231, 335 rpcbind 226 rpm 76 RSA II 241 rsh 212 run control files 133 run levels 131 0 132 1 132 2 132 3 132 4 132 5 132 6 132 emergency 132 single user 132

S
S/390 336337 s390-tools 336 s390-utils 336 s390x 337 SAN 52 sar 232, 332 savecore 218219 scheduling 144 scp 341 SCSI 52, 316 sd.conf 52 SDS 65 Secure Shell (SSH) 197 securing inetd/xinetd 189 Security Enhanced Linux (SE-Linux) 189 security/directory infrastructure 15 sendmail 122 serial console 104 serial interface as console device 30 serial interface as input/output device for GRUB 30 serial interface as input/output device for install 29

Index

421

Serial over LAN 273 configuring 274 using 276 serial port devices 102 Serial Ports Tool 102 ServeRAID 209 setfacl 194 setserial command 103 setup 8 sftp 341 sh 212 shared processing 252, 284 showrev 79 shutdown 135 shutting down 135 siga 233, 335 Simple Network Management Protocol 147 Single UNIX Specification (SUS) 9 SLES 199 SLES install boot options 282 AddSwap 283 Gateway 282 HostIP 282 Insmod 283 Install 282 Netdevice 282 Proxy 282 ProxyPort 283 SSHPassword 283 Textmode 283 UseSSH 283 VNC 283 VNCPassword 283 slice 53 slices 52 Small Computer System Interface 316 SMB 185 SMS 261 snapshots 184 SNMP 147 snoop 231 soft partitions 68 software bundles 26 software RAID 176 SoL 273 Solaris Volume Manager 65 Solaris Volume Manager (SVM) 65 Solstice DiskSuite (SDS) 65

Solstice DiskSuite / Solaris Volume Manager to Linux LVM 65 Solution Assurance Review 309 sound-juicer 98 splashimage 30 squid 122 SSH 197 SSH1 197 SSH2 198 SST 251 startup scripts 142 stat 223 statistics 173 statserial 233 strace 232233, 335 subchannel 314 subfs 95 submount file system 95 SunScreen 202 SunSolve 80 Super Digital Linear Tape 99 Super DLT 99 superuser access 196 support 10 SUSE 6 SUSE Linux portal 80 SVM 59, 65, 68 swap 5859 swapoff 5859 swapon 5859 SYN attacks 192 sysctl 192 sysfs 90 syslinux 47 syslogd 221 sysreport 233, 335 system information 140 System Management Services 261 System Service Tools 251 system services 141 system-config-netboot 136

T
tables 203 tape device name format 99 tape drives 98, 315 tape390_display 337 tapeinfo command 101

422

Solaris to Linux Migration: A Guide for System Administrators

tar 182183 tar and cpio 182 Tatu Ylnen 197 TCP wrappers 123, 190 tcpdump 231 tcsh 212 telinit 105, 135 TFTP 136 TMPFS 57 tmpfs 58 top 231232, 332, 335 traceroute 335 Travan 98 trunking 117 truss 232 ttymon 104 ttyS0 104 tune2fs 62 tunedasd 337

U
udev 52, 87 udevd 52 UDF 55 udf 56 UFS 56, 59 ufs 56 UFS snapshot 184 ufsdump 183 UID 165 ulimit 232 uname 92 unpack 182 up2date 8081 updatedb 222 user administration 160 user_xattr 60 useradd 160 userdel 160 usermod 160 usfrestore 183

Very Secure FTP (vsftp) 191 vfat 56 vgcfgbackup 67 vgcfgrestore 67 vgchange 65, 67 vgck 69 vgcreate 65 vgdisplay 67, 69 vgexport 69 vgextend 65, 67 vgimport 69 vgreduce 65 vgremove 65 vgscan 69 VIPA 339 virtual channel-to-channel 316 virtual consoles 284 virtual Ethernet 284 virtual file system 57 virtual file systems 5657 virtual I/O 253 Virtual I/O Server 250 virtual I/O server partition profile 285 configuring Linux 295 virtual machine 320 virtual memory 175 virtual network bridging 296 Virtual Partition Manager 251 virtual SCSI 284 virtual terminals 102 virtualization 284 vmconvert 337 vmstat 232, 332, 335 vold 95 volume group 68 volumes Solaris Volume Manager 65 VPN 116 vsftpd 191 VSWITCH 317

W V
VCTC 316 VDISK 313, 333 VERITAS 73 VERITAS VxVM 73 VERITAS VxVM and VxFS 73 warning banners 202 WBEM 147 Web and application serving 15 Web proxy and cache servers 122 Web Proxy Server 122 Web-Based Enterprise Management 147

Index

423

WholeDisk slice 53 writeback 60 writeback mode 60 wu-ftp 191

X
X Window System 7 xedit 325 xinet /etc/xinetd.conf 118 /etc/xinetd.d/ 118 xinetd 117118 xmcd 98 xSeries server 237

Y
YaBoot 283 yaboot.conf 283 YASSP (Yet Another Solaris Security Package) 188 YaST 8 YaST online update 81 YaST2 8 yast2 105 yast2-s390 336 you 8081

Z
zgetdump 337 zip 182 zip drive 94 zipcloak 182 ZIPL 327 zipl 330, 337 zipnote 182 zipsplit 182 zsh 212

424

Solaris to Linux Migration: A Guide for System Administrators

Solaris to Linux Migration: A Guide for System Administrators

(0.5 spine) 0.475<->0.875 250 <-> 459 pages

Back cover

Solaris to Linux Migration:

A Guide for System Administrators
A comprehensive reference for a quick transition Presents a task-based grouping of differences between the operating system environments Additional content about how to optimize Linux on IBM Eserver platforms
The goal of this IBM Redbook is to provide a technical reference for IT systems administrators in organizations that are considering a migration from Solaris to Linux-based systems. We present a systems administrator view of the technical differences and methods necessary to complete a successful migration to Linux-based systems, including coverage of how those differences translate to two major Linux distributions: Red Hat Enterprise Linux and SUSE Linux Enterprise Server. The book is designed primarily to be a reference work for the experienced Solaris 8 or 9 system administrator who will need to begin working with Linux. It should serve as a guide for system administrators that need a concise technical reference for facilitating the transition to Linux. The book also provides details about how to leverage the additional industry-leading technologies in IBM Eserver xSeries servers, IBM POWER technology-based systems (iSeries/pSeries), and IBM Eserver zSeries systems that make them very powerful and flexible platforms for hosting Linux-based solutions.

INTERNATIONAL TECHNICAL SUPPORT ORGANIZATION

BUILDING TECHNICAL INFORMATION BASED ON PRACTICAL EXPERIENCE IBM Redbooks are developed by the IBM International Technical Support Organization. Experts from IBM, Customers and Partners from around the world create timely technical information based on realistic scenarios. Specific recommendations are provided to help you implement IT solutions more effectively in your environment.

For more information: ibm.com/redbooks

SG24-7186-00 ISBN 0738496081