1

I'm in the middle of server maintenence, I upgraded 1 node's bios and windows updates and now I'm attempting to failover our database cluster to the passive, updated node. It fails over fine, but two of the 8 databases end up "suspect".

Could this be because of the windows updates or bios update?

Edit: I am using the cluster move group command from the cluster administrator.

When I move it back the suspect goes away and everything functions fine. The issue is only on the one node.

1

Sounds like you have a database connected to a drive that isn't setup on the cluster, or the dependencies aren't setup correctly. This will cause the SQL Server to start before the disks are ready which will cause the database to go suspect.

Check the ERRORLOG and see what the reason for the database going suspect is.

0

Anything is possible but I'd be more inclined to check the hard drive (or RAID) for problems first.

Ask yourself the following:

  1. What else has recently changed (that could be a more likely cause) in your environment?

  2. Take a look at your SQL Server logs to see if you can determine WHEN the error occurred. Does this correspond with your BIOS update and Windows Update theory?

  3. Talk to your team members to ask them what changes have been made and when.

  4. See if the issue has occurred on any other servers in your environment.

1
  • Suspect only happens when the second machine owns the cluster group. If I move it back there is no longer any problem. Jul 9 '09 at 19:11
0

How are you failing over? I hope you're using the CLUSTER GROUP /MOVE command, or at least through the MMC - MOVE GROUP.

I hate seeing the term "failover" when talking about clusters.

1
  • Sorry, yes you're right, I used cluster group/move. My bad on the terminology :) Jul 9 '09 at 19:10
0

I suspect (pun intended) databases that are having problems have shared storage drives that are not seen correctly by windows at the OS level.

Check that when you fail over that you can navigate to the shared drives & find the mdf & ldf files for these databases, if not then you should remove the node from the cluster & re-initalize the shared drives on the server.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.