Cannot delete a cluster resource

Question

_{Thursday, April 1, 2010 2:05 PM}

Hi!

I have a 2-node Failover Cluster on Win Server 2008 SP2 for hosting Hyper-v virtual machines.

Most of virtual machine resource groups has a Generic Service type of resource added. It runs IBM Tivoli TSM backup scheduler service. One day both cluster nodes have been started accidentally with not all disk LUNs being available. As a result one of those TSM resource failed to bring online and stucked in Failed state. The service behind runs fine and does its own job. The rest of the virtual machine cluster app seems to run OK.

As I can not correct the situation with the stucked resource I decided to delete it, as there are no dependents from it. Although I can not do it, because the Cluster Manager says it can not be deleted, because the resource is online! Actually the resource is in Failed state. Of course the action "Take offline" is not active on GUI, and the cluster.exe "offline" command does not succeed too. I tried to delete the resource by command "cluster.exe res <resname> /delete". It returns the same error, the resource is online and can not be deleted.

Any advise please!

Thanks,

Miklos

All replies (3)

_{Thursday, April 1, 2010 6:01 PM ✅Answered | 1 vote}

Here's a few things I would try:

- Check the services.msc on both nodes for the actual IBM Tivoli TSM backup service. If it's running, stop the service. If it's set to automatically start, set it to manual or disabled. If the service was started on either of the nodes, try deleting the resource after manually stopping the service

- If that didn't work, again ensure the service is set to start manual or disabled in services.msc. Then stop the cluster service on all nodes. With cluster down, open the registry editor on each node and find the resource in the HKLM\Cluster\Resources\GUID> keys. Once you've identified this resource, change the "PersistentState" value to 0. Make this change on all cluster nodes. Now try restarting cluster service and see if you can delete the resource.

- If you still can't delete the resource, you can try deleting the HKLM\Cluster\Resources\GUID> key on all nodes. Also, you would need to remove <GUID> value from the HKLM\Groups\GUID>\Contains key. Once you do this on all nodes, restart the cluster and the resource should be gone.

Hope this helps. 

Visit my blog about multi-site clustering - http://msmvps.com/blogs/jtoner

_{Thursday, April 1, 2010 8:00 PM}

Thank you for your quick reply.

As I see the 2nd and 3rd solutions require downtime in services. I will try them in the next scheduled maintenance time frame. I'll be back with the results.

I'm not sure the virtual machines can survive (stay online) while the cluster service is restarted.

_{Monday, May 3, 2010 12:14 PM}

Thank you, it helped.

The second recommendation worked. After the cluster service on the first (one of them) node was restarted the resource got in the right state and could be deleted.