none
DPM2012 not backing up Exchange 2010 RRS feed

  • Question

  • Hi all,

    We have a recently upgraded 2012 DPM server backing up our network. Our (single) Exchange server is 2010 (Version 14.2 build 247.5). This problem has occurred since our upgrade. That is to say, exchange was backing up fine after the upgrade. It has only just gone awry.

    What is occurring is, well, nothing. DPM jobs to backup the exchange databases simply do not run as scheduled; either short term to disk or long term to tape. There are no errors; neither on the exchange server nor the DPM server. As a result the job is always a green tick. It is like it never kicks off.

    The some of the protection group settings are as follows

    • Run Eseutil: Both on database and log files.
    • All databases selected for full backup.
    • Retention Range 10 days, Synchronisation frequency 12 hours, running 08:00pm every day except saturday. (almost all out short term backups kick off at 8:00pm)

    The system protection job, schedualed to run at the same time, runs without any problems and the Database backups will run flawlessly if triggered manually. I have tried removing the exchange server from DPM, then re-adding and rebuilding the protection groups.

    I am truly stuck on this one at the moment. I am having trouble even finding a starting point in diagnosing the problem.

    Thanks
    Jack

    Thursday, May 17, 2012 11:58 PM

Answers

  • I have worked it out.

    We did an inplace upgrade from DPM2010 to DPM2012.  In DPM2010 i used a domain service account to run SQL. Some time after the upgrade I remove the old database instance and set the new instance to use that service account (It was done using the SQL Configuration Manager, before you ask). It was this point that the system fell over and any and every schedualed task stopped running.

    Making the SQL service account administrator on the server fixed this issue. Its not a neat solution, i dont like throwing admin permissions around. But as this point I don't care. If something went down without complete backups, I probably would have been flailed and fired.

    I still find it hugely concerning that the system would continue to run without any useful warnings. The only indication that something was amiss was the lack of recovery points for each item and that weekly tape backups stopped running (these I tend to check more thoroughly).

    Thanks
    Jack

    • Marked as answer by CCQ IT Monday, May 21, 2012 10:42 PM
    Monday, May 21, 2012 10:42 PM

All replies

  • Ok, turns out I have lied. Twice.

    • Its not just exchange 2010 failing to backup. A whole bunch of other backup jobs are simply not running.
    • The DPM Alerts event log is awash with errors: 'DPM is out of disk space for the replica. (ID: 58)'

    In DPM 2010 the slightest issue got you an exclaimation mark or a cross. In DPM 2012, it just appears to be a whitewash of green ticks regardless of the actual backup state. For example, i have split a sql database (previously not backing up) into its own protection group. In the DPM 2012 it is hunky-dory. In the event logs I have a whole host of "The replica of XYZ on ABC is inconsistent with the protected data source. All protection activities for data source will fail until the replica is synchronized with consistency check. (ID: 3106) The data source has been marked as Inconsistent during the migration. (ID: 31223)" Clearly this is a major violation of the philosophy behind a green tick.

    Anyway, to the question at hand, I have manually increased the sizes of the replica volumes for my non-running protection groups. If they automatically take backups this evening I will close the thread.

    Auto-grow was enabled; I dont know why it would not have expanded the volumes at some point in the last 3 weeks.

    Thanks
    Jack


    • Edited by CCQ IT Friday, May 18, 2012 5:53 AM
    Friday, May 18, 2012 5:53 AM
  • I have worked it out.

    We did an inplace upgrade from DPM2010 to DPM2012.  In DPM2010 i used a domain service account to run SQL. Some time after the upgrade I remove the old database instance and set the new instance to use that service account (It was done using the SQL Configuration Manager, before you ask). It was this point that the system fell over and any and every schedualed task stopped running.

    Making the SQL service account administrator on the server fixed this issue. Its not a neat solution, i dont like throwing admin permissions around. But as this point I don't care. If something went down without complete backups, I probably would have been flailed and fired.

    I still find it hugely concerning that the system would continue to run without any useful warnings. The only indication that something was amiss was the lack of recovery points for each item and that weekly tape backups stopped running (these I tend to check more thoroughly).

    Thanks
    Jack

    • Marked as answer by CCQ IT Monday, May 21, 2012 10:42 PM
    Monday, May 21, 2012 10:42 PM