Skip navigation

HACMP Takeover - Monitorization of Filesystems and alarms

score 0
You have not voted. New

We have the following scenario:

 

- Two AIX machines in Active-Active HACMP. Let's call this machines A and B. In both machines there is a Patrol Agent running.

- In certain situations, when there is a problem some filesystems migrate from machine A to B.

- When the problem is solved, those filesystems return to origin machine A.

 

The issue we have is that when those filesystems return to machine A, a lot of alarm events are generated related to the filesystems that are unmounted in machine B.

 

When this happens we have to delete the variable /PUK/FILESYSTEM/moniList manually issuing a query from the Presentation Server and then restart the Agent.

 

This has the drawback of being a manual solution and is not operative in many situations (for example when there is a takeover at 03:00 AM)

 

Suggestion: implement some logic at KM level so we can specify which resources are in HA between two machines. Unless the filesystem is not mounted in any machine, an alarm sould not be generated.

Comments

Vote history