The Cluster Server agent for Oracle continuously monitors the Oracle database and listener processes to verify they function properly.
The agent provides the following levels of application monitoring:
Primary or Basic monitoring
This mode has Process check and Health check monitoring options. With the default Process check option, the agent verifies that the Oracle and listener processes are present in the process table. Process check cannot detect whether processes are in a hung or stopped states.
The Oracle agent provides functionality to detect whether the Oracle resource was intentionally taken offline. The agent detects graceful shutdown for Oracle 10g and later. When an administrator brings down Oracle gracefully, the agent does not trigger a resource fault even though Oracle is down. The value of the type-level attribute IntentionalOffline and the value of the resource-level attribute MonitorOption must be set to 1 to enable Oracle agent's intentional offline functionality.
For example, with the intentional offline functionality, the agent faults the Oracle resource if there is an abnormal termination of the instance. The agent reports the Oracle resource as offline if you gracefully bring down Oracle using commands like shutdown, shutdown immediate, shutdown abort, or shutdown transactional.
Secondary or Detail monitoring
In this mode, the agent runs a perl script that executes commands against the database and listener to verify their status.
When the Oracle database is in READ WRITE mode, the agent performs an UPDATE query and when it is in READ ONLY mode, the agent runs SELECT query against the database.
The Oracle agent also supports IMF (Intelligent Monitoring Framework) in the process check mode of basic monitoring. IMF enables intelligent resource monitoring. The Oracle agent is IMF-aware and uses asynchronous monitoring framework (AMF) kernel driver for resource state change notifications.
See How the Oracle and Netlsnr agents support intelligent resource monitoring.
The agent detects application failure if the monitoring routine reports an improper function of the Oracle or listener processes. When this application failure occurs, the Oracle service group fails over to another node in the cluster. Thus the agent ensures high availability for the Oracle services and the database.
More Information