none
DPM 2010 backup Hyper-V csv event id 5121

    Question

  • Hello,

    we are evaluating DPM 2010.

    When we backup our Hyper-V R2 cluster, we get a lot Eventi ID 5121 entries in our event log.
    Looks like http://technet.microsoft.com/en-us/library/ee830327(WS.10).aspx
    I think, this will happen when DPM is backing up a hyper-v-server in the cluster.
    We are using serialization, at the moment no hardware vss provider.

    Is the event 5121 "normal"?
    Or is there a problem with the cluster/csv or backup?

    Anybody also having this event entry?
    The backups are working correctly, in the last 3 months we had 3 times hyper-v-cluster problems.
    Unplanned reboots of cluster-nodes.
    So we are inverstigating the eventlogs and we have found this error frequently.

    When it is informational, why is it an error?

    Thank you

     

    Thursday, October 21, 2010 9:02 PM

Answers

  • Are you leveraging the built-in System provider or have you installed and configured a Hardware VSS provider? If you're using the System provider chances are this is occuring due to us having to place the VM in redirected IO mode for the duration of the backup.

    Understanding Protection for CSV
    http://technet.microsoft.com/en-us/library/ff634189.aspx 
    Cheers, Tyler F [MSFT] - This posting is provided "AS IS" with no warranties, and confers no rights.
    Friday, October 22, 2010 5:37 PM
    Moderator

All replies

  • Are you leveraging the built-in System provider or have you installed and configured a Hardware VSS provider? If you're using the System provider chances are this is occuring due to us having to place the VM in redirected IO mode for the duration of the backup.

    Understanding Protection for CSV
    http://technet.microsoft.com/en-us/library/ff634189.aspx 
    Cheers, Tyler F [MSFT] - This posting is provided "AS IS" with no warranties, and confers no rights.
    Friday, October 22, 2010 5:37 PM
    Moderator
  • Hello,

    thank you for your answer. The event 5121 seems to be a problem with the storage and SCSI-3-Reservations. We have to change SAN settings.

    Regards

    Marc 

    Monday, October 25, 2010 6:12 AM
  • Hi Marc,

    Can you supply the forum with more details about what exactly you needed to change on the SAN to elimnate that error ?  Can you also supply details about the  SAN  make / model involved.


    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Tuesday, February 15, 2011 4:24 PM
    Moderator
  • Hello,

    we are using HP EVA 4400. You can configure fibrechannel hosts(with multiple WWNs) to the eva and then present vdisks/luns to this hosts.
    When you present a lun/vdisk to a host you can define the "host type". There is a host type "windows" and a host type "windows Longhorn", in newer firmware releases "windows 2008". There are differences between "windows" and "windows 2008" host type.

    After presenting the lun for the host as Windows 2008 and rebooting the cluster nodes the errors are gone. The csv is now online after backup(redirected) mode.

    Regards

     

    Tuesday, February 15, 2011 7:18 PM
  • Thanks Marc - that's very useful information that may help others also experiencing those 5121 errors to resolve them.
    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Tuesday, February 15, 2011 7:44 PM
    Moderator
  • We're seeing this issue as well.

     

    IBMVSS Hardware provider.

    CSVs are SVC IBM SAN storage on the back end.

    Backups are successful, but what is happening is the storage LUN gets put in redirected mode which causes the owner of the LUN for that specific VM being backed up to go into redirected access (Even though we're using Hardware VSS provider)

    On a backup of 18 VMs that are spanned across a 2 node cluster, it's causing the owner of the actual LUNs to flip/flop between the two hosts.

    I'm not showing any actual service outtage caused because of this, and I have a call into M$ to see if I can find out a way to calm them down before we roll this into our production clusters.

    Just figured I'd give my two cents on our setup. I'm also going to go through and make sure we're all up to date on firmware/drivers etc.

     

    I'll update when I hear from Microsoft today.

    ~John

    Thursday, August 25, 2011 4:24 PM