Re: Cluster Disk corruption

Tech-Archive recommends: Fix windows errors by optimizing your registry

That's pretty much the only way...kill clussvc.exe on the node (stopping the cluster service) will force the group to failover to a different node.

FYI, while the disk is in an online pending status, it would be doing the chkdsk for you. You can look in the c:\Windows\cluster folder for chkdsk logs for specific disk signatures.

Regards,
John

"Martin" <smisik@xxxxxxxxx> wrote in message news:OUjHFXe2GHA.3372@xxxxxxxxxxxxxxxxxxxxxxx
Hi All,

recently I had an issue when during failover on our 2003 A/A file cluster on of the disk has entered pending online status but didn't come online nor it didn't fail. In event log I saw messages about corruption of that disk. I did change the properties of that disk by running "cluster res "Affected_Disk" /priv Skipchkdks=1" but then I was looking for some way how to force this disk go offline or move it to the other node. Only way I could think off at that moment was to move all groups to other node expect of one with corrupted disk and then I stopped cluster service on the node owning corrupted disk resource and respective group which forced that group to failover. After this the disk and the group came online, then I've run chkdsk to correct any errors on the disk.

I'm wondering if you guys know any other way how to force resource group to go offline or failover under such circumstances. Thank you.

--
Martin
smisik@xxxxxxxxx

Relevant Pages

  • Re: Is there a best practice for moving all cluster groups from on
    ... I agree with Mike, stopping cluster service or shutdowning the node is the ... worst way of doing a failover, as it forces it instead of doing it "quietly". ... it is gracefully shutdown before failing over. ...
    (microsoft.public.sqlserver.clustering)
  • Re: Standby server and external storage array
    ... what mechanism uses cluster service to recover database/log files after a ... failover / failback? ... > after a hard shutdown is a desperation move, not something I would bet my ...
    (microsoft.public.sqlserver.clustering)
  • Re: Is this supposed to be?
    ... want resources to failover to Node 2, you should stop the cluster service on ... you would then re-start the cluster service on Node 2. ... If I do not want resources to ... then it shows its owner is node2. ...
    (microsoft.public.windows.server.clustering)
  • RE: Cluster does not failover when unpluging the public network cable on the active node
    ... I would assume that the cluster service is connecting over the private ... The major thing that will cause a failover to initiate is the ... cluster service cannot connect to SQL Server. ...
    (microsoft.public.sqlserver.clustering)
  • Re: cluster event 1153 - res group failover
    ... flag set until the resource finally fails, this will cause "Cluster Service" ... after failover of a resource group. ...
    (microsoft.public.windows.server.clustering)