DPM Primary and Secondary Protection RRS feed

  • Question

  • Hi:

    Just wondering what is a best practice when using DPM for a file server.  Scenario...

    Primary DPM - disk only

    Secondary DPM - Disk and tape libary

    Multi TB, 20 million plus files file server backed up to primary.  I have it protected on the 2ndary DPM server as well.  File server synchronizes hourly to the primary DPM server, recovery point occurs once a day.  I'm attempting hourly synchronizations from the primary DPM server to the 2ndary DPM server for the file servers volume and a recovery point once a day (offset from the daily recovery point on the primary DPM server). If I want to accomplish a backup of the primary DPM to the secondary server and get a weekly backup done to tape....what should my synchronizations be set for on the 2ndary DPM server for the file server node?   I'm wondering if synchronizing in too many places at the same time and doing a tape backup on the 2ndary DPM server is causing the consistency checks I'm getting on both DPM servers  for this file server's volume.

    Problem with this ...if the primary server detects issues, it will start a consistency check.  2ndary detects issues on wants to do a consistency check as well.  Of course..while a consistency check goes on....really there aren't any backups.  Is MS working on a "better" way to allow DPM to keep a firm handle on the protected file systems without spawning a consistency check?

    I've doubled the change journal size on each DPM server as well.

    So questions:

    1. Is there a best practice for protecting file servers with primary and 2ndary DPM protection?

    2. Is there going to be a "fix" for long consistency checks. 





    Friday, October 28, 2011 8:51 PM

All replies

  • Hi,

    I think the correct approach is to figure out why your synchronizations are failing on the Primary DPM server which is leading to the required consistency check.  This seems to be the first domino to fall which causes other problems down the chain.   Our best practice recomendation would be for you to use volume mount points to place subset of data across several smaller volumes so a consistency check on one portion of youir data does not effect other porttions. This also helps speed up the CC should on be necessary.

    In my example below,  Host_Volume (U:) is just a very small ntfs volume that only holds the mount point directories and shares.  

    UserShare# is a folder that is shared for user access and is the empty mountpoint folder.   Mounted_Volume# is the underlying volume that holds the users directories and data.  DPM protects U: and the Mounted Volumes.   That way a chkdsk and / or a DPM CC only needs to run against the subset of users files on one volume and won’t take as long.

    Host_Volume (U:)
       UserShare1 -->Mounted_Volume1
       UserShare2 -->Mounted_Volume2

    You can “grow” the Mounted_volume at the SAN level if you need more space on that disk for users data, then use diskpart.exe command to grow the NTFS file system into the new free space, that way you can start with smaller LUN’s and grow them over time as new users / data gets added. 




    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Saturday, October 29, 2011 2:41 PM
  • OK....but what do you do with a 8 TB/20 million plus fileshare.  Splitting isn't an option.  I realize this is most likely the problem, as any backup software will have problems with this if it loses a grip on it.  Our old software though...if its journal db became invalid, the backup software would comb through the entire system (although it would take days) least you would be sure you'd be getting some data backed up.  DPM.... what can I restore if it takes 4-5 days to do a consistency check?  DPM seems to want a consistency check periodically, although not as often as DPM 2007.  I'm wondering if this will be improved with DPM 2012????

    As for the other question... do hourly synchronizations by the 2ndary dpm server affect the primary server doing synchronizations in any way?






    Monday, October 31, 2011 9:19 PM
  • Hi,

    DPM prioritizes restores over all other jobs, so you can always restore from a recovery point, even if a CC is ongoing.  DPM will only require a CC is something goes wrong with a synchronization job, or the server buckchecks and our DPM filter detects that it lost tracking a file change. 

    Go to the jobs tab and create a custom filter.  Select time from "7 days ago", deselect all jobs except recovery point, Synchronizations, and consistency checks. Under protection, deselect all data sources except that one problematic volume, then say OK.  When the results are displayed, group By computer.  That will give you a chronological view of all sync, RP, and CC jobs that ran in the last 7 days.  You will notice that it's always the first Synchronization job that fails that leads to the inconsistent replica. You need to work on the cause of the failed synchronizations. If CC's are also failing, what is the cause of those failures.  

     By default, auto consistency check will only run once, then you need to kick it off manually.  You can adjust how soon after a CC fails and how many times it runs using these settings, this should help getting back into a consistent state easier.

    Copy the below in notepad, save as autoCC.reg - then right-click and select merge, or open with registry editor to add it to the registry.


    Windows Registry Editor Version 5.00

    [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Data Protection Manager\Configuration]


    Secondary protection should not effect primary protection.

    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Monday, October 31, 2011 10:17 PM
  • I realize that you can do restores while a consistency check is occurring, the issue I have is:

    If the volume initiates a consistency check.  I do not have daily backups occurring ...the consistency check will run between 48->96 hours depending on how busy the server is.  if something happens during the consistency check, and it feels it needs to initiate another one, I might have days/weeks of no backups.  That is a concern, as it does not create recovery points until the conistency check for that volume is completed.  I realize there may be some underlying issues that cause the consistency check in the first place, but when I was at a conference earlier this year...this concern was raised by multiple users all with the same problem...all on file servers.  We had many more of these on DPM 2010 until we replaced the DPM server disk system (went with much faster disk).  So consistency checks are fewer, don't take as long (they were taking upwards of a week in the past), but still I worry about the conistency check and not having recovery points during this time.

    Wednesday, November 2, 2011 4:58 PM
  • Hi,

    Yes, this is under discussions, DPM should be able to determine which files have been updated since last successful synch by using the usn journal, then only perform CC on those specifc files.   DPM currently scans all files in alphabetical order, so you can open process monitor and see what folders are being processed by DPMRA during the CC.   Files already processed are up to date on the replica, so you can aways make a manual recovery point without synchronization and the latest versions of those files will be available for restore oncce the manual RP is created.   

    Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.
    Wednesday, November 2, 2011 6:23 PM