5.2 Monitoring the Oracle Real Application Clusters (Oracle RAC) Environment with Oracle Cluster Health Advisor

Oracle Cluster Health Advisor is automatically provisioned on each node by default when Oracle Grid Infrastructure is installed for Oracle Real Application Clusters (Oracle RAC) or Oracle RAC One Node database.

Oracle Cluster Health Advisor does not require any additional configuration. The credentials of OCHAD daemon user in the Grid Infrastructure Management Repository (GIMR), are securely and randomly generated and stored in the Oracle Grid Infrastructure Credential Store.

When Oracle Cluster Health Advisor detects an Oracle Real Application Clusters (Oracle RAC) or Oracle RAC One Node database instance as running, Oracle Cluster Health Advisor autonomously starts monitoring the cluster nodes. You must use CHACTL, while logged in as the Grid user, to explicitly turn on monitoring of the database.

To monitor the Oracle Real Application Clusters (Oracle RAC) environment:

  1. Run the following command to monitor a database:
    $ chactl monitor database –db db_unique_name
    

    Oracle Cluster Health Advisor monitors all instances of the Oracle Real Application Clusters (Oracle RAC) or Oracle RAC One Node database using the default model. Oracle Cluster Health Advisor cannot monitor single-instance Oracle databases, even if the single-instance Oracle databases share the same cluster as Oracle Real Application Clusters (Oracle RAC) databases.

    Oracle Cluster Health Advisor preserves database monitoring status across cluster restarts as Oracle Cluster Health Advisor stores the status information in the GIMR. Each database instance is monitored independently both across Oracle Real Application Clusters (Oracle RAC) database nodes and when more than one databases run on a single node.

  2. Run the following command to stop monitoring a database:
    $ chactl unmonitor database –db db_unique_name
    

    Oracle Cluster Health Advisor stops monitoring all instances of the specified database. However, Oracle Cluster Health Advisor does not delete any data or problems until it is aged out beyond the retention period.

  3. Run the following command to check monitoring status of all cluster nodes and databases:
    $ chactl status
    

    Use the –verbose option to see additional details, such as the models used for the nodes and each database.