Monday, April 7, 2014

OM12x Resource Pools Greyed Out? Try This Approach

Issue
Bumped into an issue where the OM12 Resource Pools were greyed out and no matter what I tried (dumping the cache of the Health Service, rebooting ALL SCOM 2012x Management Servers) it didn’t change. Even after an hour or so the situation remained the same:
image

This is bad. Everything was in place and configured as intended, including the AV exclusions. So it was time for another approach in order to get this train moving again.

Solution
In this case I was dealing with three OM12x Management Servers. The OpsMgr event log on all these servers was filled with events about the Pool Manager, not being able to force an election about who owns a certain Resource Pool, hence this situation.

Time to force it a bit Smile.

This is what I did:

  1. Stopping ALL 3 OM12x services on one OM12x Management Server (in this case the first server installed in the new MG);
  2. Keeping an eye on this server so no hidden process/work flow would start them again;
  3. Restarting the Health Services on the two remaining OM12x Management Servers.

Within a few minutes I started to see movement again. Gradually the MG started to show statuses. Okay, some were in a critical status, but that’s not important at that moment. It’s a live status which is good!

After 15 minutes all MG Functions and MG Infrastructure components in the SCOM Console were showing a live status and were stable. Time for the next move:

  1. Restarting all OM12x services on the first OM12x Management Server;
  2. Now the status for some MG Functions and/or MG Infrastructure Components started to change to a critical state;
  3. Within two minutes all was GREEN!!!
  4. And after an hour it was still green and stable.

Recap
The Resource Pools provide high availability and load balancing out of the box. Which is really good. However, Resource Pools can be a challenge as well, especially when an owner of a Resource Pool can’t be selected. In order to solve this, stop ALL three OM12x services on the Management Server which was first installed for the MG and go from there. Many times this will solve the issue.

BIG word of thanks
This method is the result of team work. Thanks Bob Cornelissen for tuning in and sharing your thoughts and experiences. Together we came to this approach and it worked! Awesome!

No comments: