Update to NMS RAC Failover for NMS Services (Doc ID 2717973.1)

Last updated on NOVEMBER 10, 2022

Applies to:

Goal

Historically Oracle NMS Services (since the early 2000’s) has relied on the Oracle Transparent Application Failover (TAF) feature of the Oracle RDBMS to support failover from one Oracle RAC node to another during a failure. Oracle NMS no longer recommends using TAF described in KM document 1467906.1 "How to Utilize a RAC Database With NMS". Since the original Oracle TAF implementation, the Oracle RDBMS product has added functionality to the Oracle Call Interface (OCI) client libraries that can be leveraged for a similar purpose.  The NMS services will now support receiving Fast Application Notification (FAN) events.

The updated version addresses issues where during failover of a RAC node, the NMS system would in some situations freeze up soon afterward.  The system required a restart due to this situation.

Oracle NMS can be configured to run against an Oracle RAC RDBMS. There are two primary purposes for this type of configuration:

1. Allow the Oracle RAC RDBMS to be patched live (one RAC node at a time) while the RDBMS remains active on at least one alternate RAC node.
2. General RDBMS availability (not scalability). If a RAC node fails the Oracle NMS should automatically fail over to a surviving RAC node.

To effectively achieve the benefits noted above, the Oracle NMS and Oracle RAC RDBMS must be properly configured. The purpose of this knowledge article is to document what configuration options have been validated (internally within Oracle) to generally work. Note there may be other configuration options that can also be successfully deployed but have yet to be fully validated.

Solution

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.