2010 2nd International Conference on Industrial and Information Systems

A Brief Analysis of the Disaster Recovery Backup Technology in Oracle Database

DataGuard

Xiu- ju LIU

Heze University Computer and Information Engineering

department

Heze City of Shandong Province, China,

Email:[email protected]

Abstract: This paper has discussed the data to accommodate replaced the local system and to ensure business
the disaster backup the concept as well as the DataGuard continuity. Application disaster recovery data disaster
outline. From DataGuard key technologies , principle of work recovery at a higher level than that in other places to
and so on, it analyzed the Oracle DataGuard disaster recovery
establish a complete system with substantial local data
technology with high reliability and high security.
backup applications, the disaster occurs, the remote

key words: Database; DataGuard; Backup Oracle

application system will quickly take over or assume the
business of local applications to run.
I. INTRODUCTION
III. DATAGUARD OUTLINE
As we all know, the data is very valuable corporate
asset, the data is very important for enterprises, especially DataGuard the most original version is implemented
highly dependent on the data of enterprises is very in Oracle 7, then known as the Standby Database, but this
important. The data backup is more important than all, version has an automatic transmission log file can not be
and we need to know that almost any computer data achieved, but the need to hand-write a script or program
backup system is absolutely essential component of to achieve the transfer archived logs have not been widely
unexpected power failure, system or server crashes, user used. After Oracle 8i improved, DataGuard technology
errors, disk damage and even catastrophic loss of data really a lot of use, is the Oracle 9i database, and it is now
center may caused the damage or loss of database files, through the Oracle 9i R2, Oracle 109 version of the
and these documents often contain valuable data, not by DataGuard technology to improve and strengthen already
any loss. "9.11" incident of data resulting from the World close to perfect, not only to achieve real-time backup of
Trade Center disaster revealed the deeper truth. Before physical standby database query , but also to support the
the disaster, about 250 companies working in the World physical standby database to read and write operations,
Trade Center, one year after the accident, return to the but also to achieve high-speed backup and recovery and
World Trade Center into a 150 companies. 200 companies other advanced features.
since the destruction of important information systems, In many disaster recovery solutions, DataGuard
critical data loss and permanent closure, disappeared. But Technology is a powerhouse among. DataGuard is built
also in the World Trade Center's financial giant Morgan on the Oracle database, a functional module is the
Stanley in a few hours after the incident, the Department management, monitoring, automation software
announced the world's business as usual in the second day infrastructure that create, maintain and monitor one or
of work, reason is that the company set up remote data more standby databases to protect enterprise data
backup and disaster recovery system, enables the structure without failure, disaster, error, and the collapse
company important data can be saved. From this we can of impact. DataGuard standby database to maintain the
see a modem enterprise for the Information Age, when production database with a copy of the same issues, these

disaster comes, doing disaster backup is essential. standby databases may be located thousands of miles
away from the production data center for remote disaster
II. DATA DISASTER RECOVERY recovery site, it may be in the same city, same school and

Disaster Recovery in different places through the even the same building. When the production database or

establishment and maintenance of a backup storage unexpected interruption due to planned interruptions

system uses geographic separation to ensure system and become unavailable, Data-Guard can switch any standby

data on the resilience of catastrophic events. Disaster database to the production role, thus interrupt-related

disaster recovery system according to the degree of downtime and minimize and prevent any data loss.

resistance can be divided into data disaster recovery and

IV. DATAGUARD KEY TECHNOLOGIES
disaster recovery applications. [1] data off-site disaster
recovery is a data system that can be critical applications If you want to use DataGuard Oracle database
on the local system of real-time data replication, when technology to achieve a reliable disaster recovery, Oracle
there is a disaster, the remote system can be quickly database on the need for some of the important concepts

978-1-4244-8217-7110/$26.00 ©201O IEEE lIS 2010

234
related to technology and a more clear awareness. model must be configured at least two standby databases,
thus providing double fault protection. When the last
A. redo mechanism (Redo)
participating standby database is unavailable, the master
Oracle database redo mechanism is to prevent data database processing will stop. This ensures that all the
errors and to achieve disaster recovery foundation. Redo standby database when it lost contact, it will not lose
mechanism implementation is divided into non-archive services; but if the network is not smooth, it will cause
mode and archive mode. Non-archive mode, the database LGWR can not transmit data, will cause serious
only from non-normal shutdown, power failure and other performance problems.
errors common stock to recover data; in archive mode, (2)Maximize Performance (maximum performance):
the data inventory allows time-based recovery, recovery, is the default protection mode, compared with the highest
based on changes in all cases to restore lost data to availability, provides slightly less protection of the
achieve the eating "regret it later" effect. primary database data, but provides higher performance.

B. The online redo log (Online Redo Log) In this mode, when the master database processing
transactions, redo data from the log write process and the
Both in non-archive mode or in archive mode, online
asynchronous transfer to a standby database. In any case,
redo logs are to be there. It records all the way circulation
are first to complete the write operation on the primary
database changes. It is at least two online redo log file
database, submitted to the primary database backup
group composition. DataGuard standby database in there
operation does not wait for the database to confirm
another type of online redo logs, known as the Standby
reception. If the goal of any standby database becomes
Redo Log, is mainly used for data synchronization.
unavailable, then the process will continue in the main
C. archived redo log (Ar chive Redo Log) database, which only a small impact on the performance
or no effect. When the primary database availability and
Only systems running in archive mode, archived
performance than the risk of losing a small amount of
redo log files will be generated. When online redo log
data is more important, you should use the highest
switch to the next online redo log group, the system
performance mode
would generate a redo log an offline copy of the file, this
file is an archived redo log. Archived redo log is to (3) Maximize Availabilit (maximum availability):
achieve "regret it later," one of the necessary conditions, also known as no data loss mode, which allows data

if you want to completely restore the database, but also a differences, allows asynchronous transmission. Normally

full backup of the database, Oracle 1Og's FlashBack run in maximum protection mode, the primary database

technology to achieve the purpose of data rollback. and standby database disconnect or connect the network
abnormal, it will automatically switch to maximum
D. database backup of the database used performance mode, the primary database operations can
Oracle database backup into: logical backup, which still continue. In the case of poor network performance
means using IMP tool to export a backup, but will lose has a greater impact.
some of that back up the physical structure, such as If the realization of DataGuard technology disaster
indexes, views, etc.; the other is the physical backup, it is recovery and backup, the above technology is essential,
used, such as RMAN, OS Copy production data such as business-related managers must master.
inventory tools lossless backup, the establishment of such
v. WORK PRINCIPLES OF DATAGUARD
a physical backup is to be used in Data-Guard technology.
DataGuard to achieve disaster recovery and backup, A. DataGuard process structure
through the physical standby database and logical standby
In the main database, Oracle DataGuard using the
database to achieve, if the production database fails, it can
log writer (LGWR) or archive process (ARCH) to collect
be physical or logical standby database to switch to the
transaction redo data and its transfer to the standby
production database; If, to the production database,
database; use for archive log process (FAL) provides a
hardware upgrades, on the need to switch the current
client-server mechanism for the primary database and
production database physical or logical standby database.
standby database communications between the
E. DataGuard protection mode interruption, he will send archive logs to the standby
database to automatically fill and re-sync interval. [3]
Oracle DataGuard to provide three high-level modes
The standby database, Oracle DataGuard Remote
of data protection to Pingheng cost, availability,
file server (RFS) process receives redo records from the
performance and transaction protection. [2]
primary database; use and management of the recovery
(1) Maximize Protection (maximum protection):
process (MRP) will be applied to physical standby redo
refers to the master database provides the highest level of
information in the database; using logical standby process
data protection, ensuring a comprehensive zero-data loss
will be through the SQL conversion redo information
disaster recovery solutions. When the maximum
applied to the database.
protection mode, the redo log records written by the
synchronization process to transfer from the primary B. The working principle of the physical standby
database to the standby database, and transaction data databases
until confirmed at least one backup server disk is
If the production database is running in the
available, submit only the main database affairs. This
Maximize Protection mode, then the LGWR process on it

235
very frequently not only to the changes in the redo logs standby database can not be achieved without loss of data
written to online redo log at the same time, it will be as a backup, but it can achieve asynchronous search
through the physical standby database, the RFS process function, so it can share some of the production database
writes standby redo log. Then, if the production database query task.
redo log switch occurs logs, archived redo logs generated, Logical standby database Maximize Performance
it also triggers the physical standby database, standby model and the physical standby database model similar to
redo logs on the log switch occurs and archived redo logs Maximize Performance.
generated, if the physical standby time database recovery
VI. CONCLUSION
mode is set to manage, then the physical standby database
using the MRP process will simultaneously generate Oracle DataGuard technology has the advantage of
archive logs to restore them to the database in order to disaster recovery, very clear, it has good reliability, high
complete the database on the health of non-real-time security, very little impact on performance. When the
backup of data loss. However, in this mode, each data database machine fails, only a few minutes to run the
change is always submitted to two databases at the same database as usual, the data loss rate is almost zero. The
time, the physical standby database, so if there is any practicality of this disaster recovery method development
failure, the production has a huge database, such as to determine its future prospects are very broad.
physical standby database crashes, the database will be
generated At the same time hang.
If the production database is running in the REFERENCES
Maximize Performance mode, DataGuard's default, the [1] Qing-Ming Wu.On the Oracle database disaster recovery in data
transmission method.Modern computer,pp. 59 - 60,February 2003.
production database will occur in the online redo log
archive log and generate a log switch at the same time, [2] Yan Gong. Build system with Oracle 9i dataguard disaster
recovery . Computer Applications and Software, vol. 23,pp.66-68
the data passed by the ARCH process to a physical October 2006.
standby database, and through its on RFS process directly [3] Jian Huang.Based on Oracle Data Guard disaster recovery strategy
generates archived redo logs, eventually, MRP process design and implementation . Technology Square, November 2006.
will archive the log back to the physical standby database, [4] Zhen Chen. Oracle database disaster recovery. Network manager
thus completing the data backup. Can be seen, this model, in the world , September2007.
the physical standby database is not synchronized with
the production database, even in normal circumstances,
there will still be a gap between the current online redo
log, so this way the loss of data backup will occur . [4]
If the production database is running in the
Maximize Availability mode, then the database under
normal circumstances, work and Maximize Protection
mode is the same, but if the physical standby database
fails, then the production database will automatically
switch to Maximize Performance Mode.
Suggest that if only one physical standby database,
run by the third best; if there are more servers, it can also
add a third by the first operating mode, so that the
maximum availability of the data protection .

C. The working principle of logical standby databases

DataGuard standby database if the deployment is
logical, then the production database can only work
Maximize Performance and Maximize Availability in two
modes, and the logical standby database does not support
this standby redo log structure, therefore, will appear this
way the loss of data .
When the production database running in Maximize
Availability mode, the LGWR process on the logical
standby database connected to the RFS process, and ways
to synchronize the data in the redo log buffer is written to
the logical standby database archived redo log. When the
production database online redo log switch occurs, you
will also trigger the logical standby database archive log
switch, then LSP process (using Logmnr technology) to
resolve these archive logs to restore to the SQL statement
and the logical standby database, the complete the backup
job. Although the logical standby database as the physical

236