AnsweredAssumed Answered

How to ensure calls to DB don't abort a PAM process in a clustered system?

Question asked by fjkaplan on Feb 26, 2016
Latest reply on Mar 1, 2016 by Andy_Thompson

Our setup:

  • Clustered PAM (primary/secondary)
  • Clustered DBs (primary/secondary)

 

The issue:

A few weeks ago, our IT team ran some updates on our primary DB servers.  When the reboots occurred, we had a PAM process abort with the following error message:

 

"SHUTDOWN is in progress."

 

It looks as if our PAM DB operator didn't know how to handle the 'SHUTDOWN' message, and threw an exception.

 

The question:

What's the (comprehensive) solution for this issue that will:

  1. Make the PAM DB operator robust enough to handle messages from the DB server without throwing an exception, and
  2. Enable us to continue querying/writing the DB when the 'failover' to the secondary DB kicks in?

 

Any documentation/best practices for how to do this with a PAM DB operator?

 

-Fred

Outcomes