none
error 5121 in clusterlog while backing csv volume with hardware vss writer snapshot RRS feed

  • Question

  • Hi.

     

    we are testing backing up csv volumes with dpm using hp hardware vss writers. when i backup the environment i see that the volumes are in redirect io while the snapshots are being generated and then brought online and after that the actual backup runs from the snapshots but i still get alot of error 5121 in the clusterlog should i still have these while using hardware snapshots? the disk is of course in rederict io mode during the snapshot phase so i guess it should be like that but it shouldnt be flaged as an error in that case?



    • Edited by J_Rosenberg Wednesday, May 11, 2011 7:08 AM
    Friday, May 6, 2011 11:11 AM

Answers

  • Hi,

     

    Bingo - yes that is the provider that we have seen these events from.

    Some customers have been able to resolve this problem by making the following SAN configuration change.  

    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 the option "Windows Longhorn" is replaced with "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 5121 errors are gone.

    Other customers report that the above change had no effect. 

    Let me know if that helps you, if not I would open a ticket with HP to see if they have a solution for the nag events.


    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Wednesday, May 11, 2011 3:28 PM
    Moderator

All replies

  • This is common and is supposed to be benign.  Keep your eye on any 1520 events or any crashing of VMs as a result of hardware assisted VSS.  See some of the other threads.  What storage hardware and model are you working with?

     

    Rob McShinsky (http://www.VirtuallyAware.com


    VirtuallyAware - Experiences in a Virtual World
    Friday, May 6, 2011 6:05 PM
  • J,

    Are you seeing 5121 (versus 1521 you reported) as seen in this http://technet.microsoft.com/en-us/library/ee830327(WS.10).aspx ?

    If so, they are not normal but have not seem them cause any harm.  We have only seen them occur with one VSS hardware provider to date.  Can you please run this command and provide the output.

    From an administrative command prompt run:

        vssadmin list providers


    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Friday, May 6, 2011 8:39 PM
    Moderator
  • you right my bad. its 5121 not 1521.

    here is the provider we use

     

    Provider name: 'hpEVA VSS Hardware Provider'
       Provider type: Hardware
       Provider Id: {21e5ab69-9685-4664-a5b2-4ca42bddb153}
       Version: 6.5.0.2

    Wednesday, May 11, 2011 6:30 AM
  • Hi,

     

    Bingo - yes that is the provider that we have seen these events from.

    Some customers have been able to resolve this problem by making the following SAN configuration change.  

    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 the option "Windows Longhorn" is replaced with "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 5121 errors are gone.

    Other customers report that the above change had no effect. 

    Let me know if that helps you, if not I would open a ticket with HP to see if they have a solution for the nag events.


    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Wednesday, May 11, 2011 3:28 PM
    Moderator
  • we already presented the luns as windows2008 in the commandview console.

    i will check what hp has to say about it.

     

    /JR

    Thursday, May 12, 2011 8:59 AM
  • Hi

    I have exactly the same problem. I'm using an HP EVA 6100, Hyper-V Cluster with 4 Windows 2008 R2 SP1 Hosts, Backup Software DPM2010 with HP VSS Hardware Providers Installed.

    Provider name: 'hpEVA VSS Hardware Provider'
       Provider type: Hardware
       Provider Id: {21e5ab69-9685-4664-a5b2-4ca42bddb153}
       Version: 6.5.0.2

    Host Type on the EVA is Windows 2008 selected for the 4 Servers. (MPIO Version is 4.01, VSS 6.05, Command View 9.3)

    For the testbackup I have installed on each Hyper-V Host 1 Virtual Maschine. The Hosts are using one CSV, where all the virtual machines are stored on.

    When I backup only one virtual machine I receive no error. But when multiple machines are backuped the same time I reiceive the following events:
    Event ID 5121, Failover Clustering (Cluster Shared Volume 'Volume1' ('Cluster Disk 2') is no longer directly accessible from this cluster node)
    Event ID 304, hpeaadsm (An unrecoverable path failure occurred on SCSI address (0.0.0.1))

    Wednesday, May 18, 2011 8:48 AM
  • If the suggested answer does not help in resolving this thread then please re-open it.
    Cheers, Tyler F [MSFT] - This posting is provided "AS IS" with no warranties, and confers no rights.
    Friday, May 20, 2011 6:48 PM
    Moderator
  • We're having the same problem with our hyper-v 3-node cluster on server2008R2 SP1 connected to a HP P4300 ("Lefthand") SAN. Snapshots are taken with the HP P4000 hardware snapshot provider (most current version to date). Unfortenately it's not possible to assign a server host type like "windows server 2008" in the storage configuration of the HP P4300.
    Thursday, August 18, 2011 9:11 AM
  • I get the same thing, just 1 or 2 while the snapshot kicks off after that it is OK, using Equallogic.
    Thanks, -Lindy
    Monday, September 5, 2011 11:54 PM
  • Same thing with EMC provider and CX3 SAN. The 5121 is closely followed by a 5122 to show access is back in 'normal' mode, but it's not pretty looking at the logs on my cluster. Is there a definitive answer from MS regarding hardware providers and 5121 errors? It seems that even in this forum there are divided opinions as to whether it's normal or an issue with the provider? (or possibly with DPM, if you raise a call with your SAN vendor....)
    • Edited by farscapes Friday, September 16, 2011 7:24 AM
    Friday, September 16, 2011 6:47 AM
  • I have the same problem mit Fujitsu Eternus DX60 VSS Hardware Provider.
    Tuesday, November 8, 2011 11:42 AM
  • We use an EqualLogic as well.  Once the backup kicks off it produces the error then once the backup is done it's back to normal.  I have not seen any of the VM's on the CSV fail though.  It would be nice to not see any errors in the log although it seems like these are more just informational messages.
    Cheers, Chris Holley, MCITP: Virtualization
    Wednesday, November 23, 2011 4:15 PM