none
DPM 2012: "Backup using child partition" is not getting the latest version of files RRS feed

  • Question

  • I am running DPM 2012 build 4.0.1920.0 on Windows Server 2008. I'm backing up three Hyper-V guests on one server with "Backup using child partition." The Hyper-V server is also WS 2008 (not R2). I discovered a while ago that, although my jobs were running without errors and reporting success, nothing was actually being backed up anymore.

    I've completely uninstalled and reinstalled DPM and am still having this same problem. I can create the replicas just fine. But then when the regular recovery point creation job runs, it says it's synchronizing, but it doesn't seem to be. Nothing that has changed since the replica was created is found in the recovery point. If I schedule a daily consistency check, I get a warning saying that the replica is already consistent. But if I run a consistency check manually, it finds inconsistencies and then I do get the changed files.

    It used to work. As near as I can tell, the failures started at the same time as the July 2012 Patch Tuesday updates. I don't know if that means anything or is just a coincidence though.

    Tuesday, January 22, 2013 7:08 PM

All replies

  • Hi,

    Can you please make a new recovery point with integrity check using the following steps, it still applies to DPM 2012.

    http://technet.microsoft.com/en-us/library/cc627335.aspx

    New Information

    Cmdlet support for integrity check
    DPM 2007 now allows you to trigger an integrity check on a new recovery point through cmdlets through a new parameter WithDataIntegrityCheck. This parameter does not take a value, but if used at the time of creating a new recovery point by using Expressfull backup, it triggers an integrity check on the recovery point.

    Syntax

    New-RecoveryPoint –Datasource name of data source -Disk -BackupType expressfull –WithDataIntegrityCheck

    Example – Integrity Check
    The following sequence of commands allows you to create a new recovery point and perform an integrity check on it, so do that for a guest and see what we get.

    $dpmname = (&HOSTNAME)
    $pg = Get-ProtectionGroup -DPMServerName $dpmname
    $pg
    $ds = Get-Datasource -ProtectionGroup $pg[X]         #(WHERE X IS A ZERO BASED INDEX FROM THE OUTPUT OF $PG) 
    $ds
    $recoverypt = New-RecoveryPoint -Datasource $ds[Y] -Disk -BackupType expressfull –WithDataIntegrityCheck  #(WHERE Y IS A ZERO BASED INDEX FROM THE OUTOUT OF $DS)

    If it finds any mismatch, it fixes the replica for the same and it raises the alert that corruption has been detected.


    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.

    Tuesday, January 22, 2013 11:32 PM
    Moderator
  • I did that, and it worked. The new files were detected and backed up. I didn't see any alert, though.

    Thursday, January 24, 2013 6:25 PM
  • Hi,

    OK - let it run normal synchronizations for a day, then re-run the power shell command and see what happens.  This basically checks if the DPM filter is missing any block level changes, or if replica data is mismatched with source data.


    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.

    Thursday, January 24, 2013 6:30 PM
    Moderator
  • I re-ran the recovery point with integrity check and the same thing happened. Also, this time I found the warning that popped up about the inconsistent replica.
    Monday, January 28, 2013 6:18 PM
  • Hi,

    On the DPM 2012 Server, please run these commands:

    Note if this was an upgrade from DPM 2010, path will be: C:\Program Files\Microsoft DPM\DPM\Temp

    CD "C:\Program Files\Microsoft System Center 2012\DPM\DPM\Temp"
    Find /I “Corruption detected”  *.errlog

    If this returns anything, then we need to investigate some thing on the protected server.


    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.

    Monday, January 28, 2013 7:12 PM
    Moderator
  • The search came back with 0 matches.
    Tuesday, January 29, 2013 5:31 PM
  • Ok,

    So If understand your original problem statement,  Scheduled Synchronizations and recovery point jobs run and complete successfully, but no data seems to be transferred for the successful jobs.  If you run a CC, then you see lots of data transferred ?

    NOTE: CC jobs will include normal agent communications including CRC checking as part of the "data transferred" number, so two back to back CC's will show roughly the same amount of data transferred, even if no real data changed on the guest.  (IE: If you shut down a guest, or put in a saved state, then run two CC jobs back to back, you will see the same amount of data transferred, although, none of that represents any changed data from the saved VM.)


    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.

    Tuesday, January 29, 2013 6:14 PM
    Moderator
  • Recovery point jobs run and complete successfully but do not seem to transfer any data. Scheduled consistency checks claim that the replica is consistent, even though I could run a CC five minutes later and the inconsistency is found and fixed.
    Tuesday, January 29, 2013 8:29 PM
  • Hi,

    <snip>
    even though I could run a CC five minutes later and the inconsistency is found and fixed.
    >snip<

    What are you looking at to determine that the on-demand CC found inconsistencies ?


    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.

    Tuesday, January 29, 2013 9:27 PM
    Moderator
  • Looking at the job in the Monitoring tab.
    Tuesday, January 29, 2013 9:33 PM
  • Hi,

    As noted earlier, CC jobs will include normal agent communications including CRC checking as part of the "data transferred" number, so two back to back CC's will show roughly the same amount of data transferred, even if no real data changed on the guest.  (IE: If you shut down a guest, or put in a saved state, then run two CC jobs back to back, you will see the same amount of data transferred, although, none of that represents any changed data from the saved VM.)

    So that is not a reliable way of knowing if any real changes came over.  Anyhow, I have not heard of the problem that you describe, please open a support case to investigate the behavior.


    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.

    Tuesday, January 29, 2013 9:48 PM
    Moderator
  • I understand, but in my experience, the on-demand CC causes data to be transferred over that had not been prior to the CC. I'm looking at whether specific files are available for recovery, not the amount of data transferred.

    I will open a support case. Thank you.


    • Edited by jroth_uiuc Tuesday, January 29, 2013 10:26 PM
    Tuesday, January 29, 2013 10:21 PM
  • Hi,

    Ok - I think I understand, so you are looking inside the VM after a CC and new recovery point is made and seeing file difference inside the guest that should have been brought over during earlier synchronizations ?

    If that is the condition, it may be a Anti-virus filter driver incompatibility issue.  To verify that, you would need to un-install (not just disable) the antivirus software from the protected server and reboot, then see if synchronizations work.


    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.

    Tuesday, January 29, 2013 10:34 PM
    Moderator
  • Hi,

    Ok - I think I understand, so you are looking inside the VM after a CC and new recovery point is made and seeing file difference inside the guest that should have been brought over during earlier synchronizations ?

    Yes, that's right.

    There's no AV software on the protected server.

    Tuesday, January 29, 2013 10:39 PM
  • Hi,

    Can you open an administrative command prompt and run the following and paste the results here:

    c:\>fltmc instances


    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.

    Tuesday, January 29, 2013 11:22 PM
    Moderator
  • Filter                Volume Name                              Altitude        Instance Name      Frame  VlStatus
    --------------------  -------------------------------------  ------------  ---------------------  -----  --------
    DpmFilter             C:                                      385300       DpmFilter Instance       1
    DpmFilter             \Device\HarddiskVolume11                385300       DpmFilter Instance       1    Detached
    DpmFilter             \Device\HarddiskVolume12                385300       DpmFilter Instance       1    Detached
    DpmFilter             \Device\HarddiskVolume13                385300       DpmFilter Instance       1    Detached
    DpmFilter             \Device\HarddiskVolume14                385300       DpmFilter Instance       1    Detached
    DpmFilter             \Device\HarddiskVolume15                385300       DpmFilter Instance       1    Detached
    luafv                 C:                                      135000       luafv                    0

    • Edited by jroth_uiuc Wednesday, January 30, 2013 4:33 AM
    Wednesday, January 30, 2013 4:31 AM
  • Hi,

    Curious to know where the effected VM lives because the volumes that don't have drive letters are not being tracked by the DPMFilter - note the filter is detached.

    What are those volumes ?  Are they NTFS ?


    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, January 30, 2013 2:46 PM
    Moderator