Cluster Taken Offline Accidentally

  • Hello,

     

    I have a two node cluster, and both nodes are up (this is a testing environment) - both Virt. SQL servers are up as well. The cluster resources are in their own group (iem cluster IP, cluster name, Q drive, etc). The question is...

     

    How do you restore this Cluster Resource Group if it has accidentally been taken offline and neither node can connect to it, since the IP, name, a Q drive have been taken offline?

  • Can you check to see if the cluster service is running on the physical nodes of the cluster?  If it's not, try starting it.

     

  • Yes, cluster server is running on both nodes. The issue is - since the IP address, clus name and Q drive were taken offline, how do you then re-connect to the cluster group via Cluster Administrator to turn them back online, when the Cluster Administrator can't connect to it?

  • Have a look at this KB article.  You might be able to change the persistent state and restart the cluster service.

    http://support.microsoft.com/kb/259243/en-us

     

     

  • Hmmm - gave it shot - didn't seem to help. I think the issue is that the IP for the Cluster Name is down (it cannot even be pinged) so no IP, no connection. This is a "watch-out" for issue with Cluster Administrator - one wrong slip of the mouse, and the whole cluster goes down. Is there any way of bringing the IP online outside of cluster administrator?

  • You could try connecting to 127.0.0.1 from one of the cluster nodes using the cluster administrator. That works for me, although my cluster group is up.

    What is interesting is that the cluster hive in the registry  and the "persistent state" keys tell the cluster service what state the individual groups and resources need to be in when the cluster comes online.  If these are set to "1" and the cluster is restarted, and they don't come online, then there is perhaps something else wrong.

    Another thing you could try is to restart both physical nodes at the same time.  This would force the cluster service to read its startup state from "somewhere".  I'd try the localhost connection first, though.  I would also suggest making sure you have a good backup

     

     

     

     

  • Just found a quick way per a friend...

     

    1. Connect to one of the nodes using Remote Desktop.

    2. Start Cluster Administrator.

    3. Use a period (.) for the cluster name. The Cluster Service will start on the node automatically.

    4. Bring the Cluster group back online.

Viewing 7 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic. Login to reply