Friday 17 April 2009

How to upgrade or apply a hotfix to a Clustered RMS

We have had a some difficulties over the months trying to apply hotfixes to any clustered Root Management Server. The hotfix installers try and stop the 3 OpsMgr services, apply the hotfix and them restart the services. These steps result in some unexpected behavior on a Clustered RMS.

What I have seen is that after the hotfix installer stops the RMS services the Cluster Service will try and restart them again.

This results in either the services running on the node you are trying to patch or the RMS resource group failing over. If the RMS fails over then, when the hotfix tries to restart the services after it has completed, it either times out or you have the RMS services running on both Cluster nodes. It it restarts the service on the node you are trying to patch the install will fail as the files are in use.

Here are some steps that can avoid these issues:

1. In Cluster Administrator ensure the RMS resource group is on the node you are patching. Then Pause the Node you are not patching. Simply right Click on the node and select Pause. This stops the resource group failing over to the other node.

2. Set the 3 RMS services not to restart automatically. Selecting each service, right Click and select Properties. On the Advanced tab set it to "Do not restart"

3. Apply the hotfix to the current node

4. When the hotfix has completed reset the services to restart automatically and resume the other node.

Repeat this to apply the patch to the second node.

This process also worked for me in upgrading a clustered lab environment to the Release Candidate of R2


Kevin Holman said...

Also see another perspective at

Derek Harkin said...

Thanks Kevin. I tried that way also and did not like that it started the services on the passive node.

If the msi offered a command line switch option to skip service restarts this would be a good solution.