HP Serviceguard Oracle DataGuard Toolkit User Guide, April 2011

then run the Recovery package, which will bring up the database as a primary database by doing
a role failover. Note that in this case, role management is handled by ODG toolkit. The start up
time of recovery package is usually double the start up time for an ordinary database package.
This is because, the recovery package runs role transition commands which involve shutting down
the database after the role is changed and restarting the database. Hence “start_script_timeout”
parameter in the package configuration should be appropriately specified.
Restoring the cluster in a Continentalclusters to its original state is a manual process. The following
steps must be performed to restore the clusters back to their original state:
1. Halt the Recovery package
2. Start the Primary package—this will bring up the database on Primary Cluster as a primary
database
3. Start the Data Receiver package—this will bring up the database on the Recovery Cluster as
a standby database
RAC environment
Figure 7 RAC environment
In an RAC environment, all the three packages (Primary, Data receiver and recovery packages)
are configured as Multi-node packages (MNPs).
When the Primary cluster goes down, the recovery package on the recovery cluster must be brought
up by manually running the “cmrecovercl” command. This command will first halt the Data Receiver
package, which will halt the standby database. It will then run the Recovery package, which will
bring up the database as a primary database by doing a role failover. Note that in an RAC
environment, the start up time of recovery package is more than that in a single-instance environment
since the role transition is carried out in a slightly different manner. In an RAC environment, during
role failover only one of the standby database instances (say, Instance A) should be running while
the others should be shutdown. These other instances wait for this instance to perform role failover,
shutdown and then restart. Once these operations are performed on Instance A, the other instances
will be started. Hence “start_script_timeout” parameter in the package configuration should be
appropriately specified.
Restoring the cluster in a Continentalclusters to its original state is a manual process. The following
steps must be performed to restore the clusters back to their original state:
Supported configurations 13