In Oracle RAC 10g or later, the Oracle instances and listeners are automatically started, monitored, and recovered through the Oracle Clusterware. If PRIMECLUSTER Wizard for Oracle detects a resource failure, it will notify RMS of the “Warning” state then ask Oracle Clusterware to generate a failover or recover the failed resource (depending on the WFTime setting).
Also, the resources are started or stopped along with userApplication startup and stop.
Oracle RAC instance startup
Start an Oracle RAC instance as follows:
su - <Oracle user>
srvctl start instance -d $DB_NAME -i $ORACLE_SID
Oracle RAC instance stop
Stop an Oracle RAC instance as follows:
su - <Oracle user>
srvctl stop instance -d $DB_NAME -i $ORACLE_SID
Oracle RAC instance monitoring
The monitoring of an Oracle RAC instance is the same as standby operation and scalable operation with Oracle9i RAC. However, the monitoring interval is 60 seconds and cannot be changed.
For details, see “Oracle monitoring” of “3.1.1 Standby Operation and Scalable Operation with Oracle9i RAC”.
Listener startup
Listener startup depends on Oracle Clusterware, not PRIMECLUSTER Wizard for Oracle.
Listener shutdown
Listener stop depends on Oracle Clusterware, not PRIMECLUSTER Wizard for Oracle.
Listener monitoring
The Oracle monitoring setup is the same as standby operation and scalable operation with Oracle9i RAC.
However, the monitoring interval is fixed at 60 seconds, and tnsping monitoring is not performed.
For details, see “Oracle monitoring” of “3.1.1 Standby Operation and Scalable Operation with Oracle9i RAC”.
Oracle Clusterware startup
Start Oracle Clusterware as follows:
Start up Oracle Clusterware using “/etc/init.d/init.crs start” (Oracle RAC 10g R2/11g R1) or “crsctl start crs” (Oracle 11g R2).
Check if Oracle Clusterware is activated using “crs_stat -t” (Oracle RAC 10g R2/11g R1) or “crsctl status resource -t” (Oracle 11g R2).
Oracle Clusterware stop
Stop Oracle Clusterware as follows:
Shut down Oracle Clusterware using “/etc/init.d/init.crs stop” (Oracle RAC 10g R2/11g R1) or “crsctl stop crs” (Oracle 11g R2).
Oracle Clusterware monitoring
PRIMECLUSTER Wizard for Oracle does not monitor Oracle Clusterware. NullDetector flag is automatically enabled.
Note
If you want to startup, stop, or switch Oracle instances, execute hvswitch and hvutil on RMS userApplication. If you execute the Oracle Clusterware srvctl command or shutdown command to the Oracle instance directly, PRIMECLUSETR will consider that the instance fails, and recovery processing might be initiated for the Oracle instance. To executing the command, put userApplication into maintenance mode.
