none
DPM Cluster Shared Volume Back up Problem RRS feed

  • Question

  • Hello,

    We have IBM BladeCenter S Chassis and we are using Microsoft Windows Server Datacenter Edition 2008 R2 with Hyper - V. We have 2 IBM HS22 Blade Servers. one of them cluster node 1 and other HS22 is cluster node 2. There is a cluster shared volume on IBM BladeCenter S chassis Storage Module. Both of HS22 server access this shared disk volume on BladeCenter internal storage. And we have one more IBM x3400 M3 Server, this server using for DPM. There is no Virtualization on this x3400 M3 Server. We want to back up all virtual machines (Hyper - V Guests) with DPM. But we couldn't back up. When we start the back up operation on DPM, a problem occured on Cluster Shared volume and all virtual machines shut down. Before this operation our cluster shared volume running on online mode. But now cluster shared volume running on redirected access mode.

    So we want to back up all virtual machines. Could you help us about this problem? 

    Wednesday, October 19, 2011 1:06 PM

Answers

  • If the above suggested answer does not help in resolving the thread please re-open it.
    --------------------------------------------------------------------------------
    Regards, Michael V [MSFT] - This posting is provided "AS IS" with no warranties, and confers no rights.

     

    Wednesday, January 25, 2012 10:52 PM
    Moderator

All replies

  • Hi,

    When a backup of a Cluster Shared Volume on a Windows Server R2 Failover Cluster occurs, the CSV is placed into a BackupInProgress state. You will see this noted on the CSV in Failover Cluster administrator. While in this state, the CSV cannot be moved or brought offline. This prevents the backup job from failing. It is possible that a Backup Job can fail in such a way that the BackupInProgress state will not be cleared, leaving the CSV unmanageable. In this case, the only native way of resetting the state is to simulate a failure of the CSV which will require all of the VM's on the CSV to be taken offline.

    Do you have a VSS hardware snapshot provider installed on the cluster nodes ?  

    From administrative command prompt run:   VSSADMIN LIST PROVIDERS

    Are there any events from that provider or from volsnap when the error occured.


    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Thursday, October 20, 2011 6:34 PM
    Moderator
  • Thank you for your answer.

     

    We don't have VSS hardware snapshot provider. Our storage is internal on the Blade Server. I searched in the IBM web site for VSS but I couldn't find it. Could you help me for this patch?

     

    Thanks,

    Monday, November 21, 2011 12:22 PM
  • Hi,

    When a backup of a Cluster Shared Volume on a Windows Server R2 Failover Cluster occurs, the CSV is placed into a BackupInProgress state. You will see this noted on the CSV in Failover Cluster administrator. While in this state, the CSV cannot be moved or brought offline. This prevents the backup job from failing. It is possible that a Backup Job can fail in such a way that the BackupInProgress state will not be cleared, leaving the CSV unmanageable. In this case, the only native way of resetting the state is to simulate a failure of the CSV which will require all of the VM's on the CSV to be taken offline.

    Do you have a VSS hardware snapshot provider installed on the cluster nodes ?  

    From administrative command prompt run:   VSSADMIN LIST PROVIDERS

    Are there any events from that provider or from volsnap when the error occured.


    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.

    Hi Mike,

    In some environments if these VM's are production bringing them offline may be unacceptable. Is there another way to correct this without bringing them offline such as performing a live migration to another node?


    My Blog | www.buchatech.com | www.dpm2010.com
    Tuesday, November 22, 2011 6:28 AM
    Moderator
  • Serkan,

    The list of tested hardware providers can be found here: http://blogs.technet.com/b/dpm/archive/2010/02/05/tested-hardware-vss-provider-table.aspx

     


    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Tuesday, November 22, 2011 2:57 PM
    Moderator
  • Buchatech,

    I believe the Windows cluster support group has a utility called clearstate.exe that will reset the flag, post a question in the windows cluster forum. 


    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Tuesday, November 22, 2011 3:00 PM
    Moderator
  • Buchatech,

    I believe the Windows cluster support group has a utility called clearstate.exe that will reset the flag, post a question in the windows cluster forum. 


    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.

    Thanks Mike. I will look into that. There has to be an alternitive to taking production machines down.

     

    Thanks


    My Blog | www.buchatech.com | www.dpm2010.com
    Tuesday, November 22, 2011 6:23 PM
    Moderator
  • If the above suggested answer does not help in resolving the thread please re-open it.
    --------------------------------------------------------------------------------
    Regards, Michael V [MSFT] - This posting is provided "AS IS" with no warranties, and confers no rights.

     

    Wednesday, January 25, 2012 10:52 PM
    Moderator