AppManager for Oracle (UNIX) supports monitoring Oracle in both Real Application Clusters (RAC) and non-RAC clustered environments. Each environment uses a different method for determining whether a node is active or inactive.
An Oracle RAC consists of a single database shared across multiple Oracle instances. Each instance runs on a separate node. The Oracle RAC software keeps track of both active and inactive instances.
It is recommended that you use the new Oracle UNIX RAC module for clustered monitoring.
To use this module, you must first discover the Oracle UNIX RAC resources. For additional information about Discovering Oracle RAC UNIX, see Discovering OracleRACUnix Resources.
While you are still able to monitor clusters using the Oracle UNIX Knowledge Scripts, it is recommended that for cluster monitoring, you use the Oracle UNIX RAC Knowledge Scripts.
NOTE:This way of discovering clustered environment makes the clustered environment appear as a standalone environment in the treeview
To enable monitoring in a RAC environment, use the SetMonitoringOptions Knowledge Script to provide the following information:
Service name for the RAC cluster
Directory containing the tnsnames.ora file (only if this file is in the non-default location). The default location is $ORACLE_HOME/network/admin.
AppManager for Oracle (UNIX) uses this information to prevent the generation of spurious alerts by jobs running on inactive nodes of a cluster.
The ClusterInstanceDown Knowledge Script lets you know when an instance is inactive. Use this Knowledge Script to raise an event when such a condition occurs on a node in the cluster. This Knowledge Script is primarily useful on nodes of the cluster that should be active.
For more information about Discovering Oracle UNIX Knowledge Scripts, see Discovering Oracle Resources