Testing for Availability - Expert Oracle RAC Performance Diagnostics and Tuning - page 81

d. Database is handed back over to the Observer and placed in a GUARD ALL status.

e. Database is opened.

f. All the required standby redo log files are added back to the configuration.

6. The target standby database becomes the new primary database. Users can start

accessing data as on any normal database. At this point it is very important to consider
the configuration differences between the original primary and the new primary. If the
configurations are not identical, the servers cannot process the same workload.

7. At this stage, depending on how long it takes to repair the old primary, there could be a

period where only one database is available. Once the old primary database is repaired,
the Observer reestablishes connection.

8. The Observer automatically reinstates the old primary database to be the new standby

database. At this stage the redo transmission starts from the new primary database to the
new standby database. The environment is back in a maximum availability mode.

In this specific scenario, because the primary database and the standby database environments are not of an
identical configuration, it is important to switch the states back to their original state. That is, the new standby should
be switched back as a primary database. This can be accomplished using the switchover operations discussed in the
“planned failover” in “Step 1—Switchover” earlier in this workshop.

RAP Phase IV—Backup and Recovery
Backup and recovery is a critical area and part of everyday operation. However, because backup and recovery
are operations for the future when a problem arises, seldom is any importance given to testing it with everyday
operations. Backup operations are performed in case problems arise in the future and recovery of the database
environment is required. Best practices warrant the use of Recovery Manager (RMAN) for backup and recovery of the
RAC database.
Part of the backup and recovery testing will also require the preparation of a backup and recovery strategy, or
rules and guidelines around how often the database should be backed up. Where will the backup files be stored: on
disk, in ASM, or directly on tape? If backup is retained on disk, how many days or weeks worth of backup files will be
retained and will have to be documented? These rules and procedures will depend on the size of the database and
mean time to repair (MTTR) policy.

Workshop
This workshop has two parts. Part 1 covers backup operations and Part 2 covers recovery operations.

Step 1
The first step in the backup testing is to define a backup strategy. The backup strategy is prepared based on the
SLA requirements of the database. The MTBF and MTTF of the database should be considered when preparing the
strategy. This strategy needs to be tested as part of RAP testing. For example, Table 3-7 gives a typical backup strategy
for a clustered database configuration.