none
VSS Writer not called for DPM folder protection in VM RRS feed

  • Question

  • When a VM is backed up using DPM's "Child Partition Snapshot". the VM's VSS Writers are called to initiate a freeze and thaw.

    I have noticed on our customer's systems that when a folder on a VM is being backed up by DPM, this does not occur. There is no freeze or thaw, and therefore the backup is only in a crash consistent state. We are using a Child Partition Snapshot for end of day backups on each VM. We would like to use additional folder backups of critical files during the day too, but of course, these cannot be relied upon the way this is working now.

    Is there anyway to ensure that the freeze and thaw is triggered in the VM for a folder backup?

    Thanks,
    Bob.

    Tuesday, February 28, 2012 1:19 PM

All replies

  • Hi,

    If I understand your question, you have installed a DPM agent inside the Guest and are protecting a folder on a volume using DPM.  In this case, the VM is really being protected not as a hyper-v guest, but as a normal server, and the volume being protected will have the snapshot made by the guest OS, not hyper-V.    The SYSTEM writer in the guest will internally perform the freeze and thaw when the volume snapshot is created.  DPM always protects data from a snapshot, so in your case, the snapshot of the volume being protected inside the guest.


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread. Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.

    Wednesday, February 29, 2012 1:42 AM
    Moderator
  • Mike,

    That's what I thought should happen. However, our application is not receiving a freeze and thaw when DPM creates the recovery point.

    So, if I understand the process correctly, on a protected folder in a VM, on the DPM Server, DPM should be telling the VM that it should generate a snapshot, and the VM should be triggering the freeze and thaw in its applications.

    I thought so. I checked through the logs, and it looks like it is working intermittently. It looks like this is related to http://social.technet.microsoft.com/Forums/en-US/dataprotectionmanager/thread/740fd820-2f31-458f-a974-70671d969430 and http://social.technet.microsoft.com/Forums/en-US/winserverhyperv/thread/ec39f14e-b5fe-44c1-ae5c-157621eac141.

    Mike, is there any way you DPM guys could light a fire under the VSS guys to review these two topics? These problems should be a heads up to the DPM department. People are reporting their servers to run fine for a few hundred days then start failing VSS backups, whether they use DPM, WSB, or third party. It has started to break DPM for one of our customers now, and I am seeing the errors described in the second topic in other customers. Another is having their WSB backup for the C: drive fail regularly. Even if people reinstall their host server or reinstall their VMs, the problem is only solved for a few days - the same as rebooting the host server.

    Thanks,
    Bob.

    Wednesday, February 29, 2012 2:05 PM
  • Hi Bob,

    <snip> 
    However, our application is not receiving a freeze and thaw when DPM creates the recovery point.
    >snip<

    Does your application have a VSS writer - if not then it cannot be notified. If it does have a writer, then run these steps to make sure it's included when a snapshot is created for the volume being protected by DPM. 

    1) From an administrative command prompt, run diskshadow.exe
    2) add volume X:  (where is is the volume being protected by DPM)
    3) set verbose on
    4) create

    The snapshot process will start and the verbose output will show you the writers involved.

    In order to investigate VSS problems, customer need to open support incidents with us so complete and comprehensive troubleshooting and anaysis can be done.  Every customers environment and machine is different and trying to solve the worlds vss errors via forum communications is never going to work.  If during a proper investigation we find a code defect we can have it fixed.


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread. Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.

    Wednesday, February 29, 2012 8:46 PM
    Moderator
  • Mike,

    Yes, our app has its own VSS writer and normally works just fine.

    The diskshadow idea is cool. I will use it when the next outage occurs, if I can catch it, to see if I can see anything wrong with the writers.

    The support incident idea sounds expensive and probably won't fly with the customer. If we can't figure out anything else though, we may have to come back to it. In the meantime, I have a few ideas and things to try - and a neat way to verify writer problems.

    Thanks,
    Bob.

    Thursday, March 1, 2012 1:54 AM