SQL Server Performance

AS corruption during re-processing

Discussion in 'Analysis Services/Data Warehousing' started by thomas, Jan 15, 2004.

  1. thomas New Member

    A couple of times I have recently had problems with corruptions developing on my AS server during nightly cube re-processing (full processing, not incremental).

    I get errors such as:-

    There was an error during loading dimension Supporter.PostalArea in database MAP_Mart. Dimension file(s) may be corrupted or missing.

    Analysis Services crashes and reports the database as unusable. I then have to restore from the archive of the day before, then re-process.

    Has anyone else seen this? Any ideas as to causes? Hardware looks fine. I've searched the MSKB but it doesn't seem to have anything relevant.

    I realise this isn't strictly a "performance" issue, but I thought this category suited the best.

    TIA!

    Tom Pullen
    DBA, Oxfam GB
  2. satya Moderator

    There is little help available on MS site about AS and its hard to pinpoint root cause.

    I assume AS is patched upto the SP levels including OS.
    Any other information/error in SQL error log and event viewer log?
    Which provider you're using?

    For information review thishttp://msdn.microsoft.com/library/default.asp?url=/library/en-us/trblsql/agtroubleshoot_7dd3.asp link and see whether this helps.

    Satya SKJ
    Moderator
    http://www.SQL-Server-Performance.Com/forum
    This posting is provided “AS IS” with no rights for the sake of knowledge sharing.
  3. thomas New Member

    Thanks Satya, that link doesn't seem to work, though?

    Analysis Services is SP3a and server OS is Windows 2000 SP3 (shortly to be SP4).

    There's not much more in the error log - just related errors such as

    "Due to errors during loading, the database MAP_Mart will not be available for querying. It is recommended that you delete and restore this database from the last known good archive."

    and

    "There was a fatal error during transaction commit in the database MAP_Mart. Server will be restarted to preserve the data consistency."

    What do you mean by provider? Source database is SQL Server.

    Tom Pullen
    DBA, Oxfam GB
  4. satya Moderator

    For provider details I mean to ask about how you're connecting to a Data Source in the AS.
    The later errors are connected to the inital error while processing and while you're getting did you see any other behaviour in the server, which I believe this error is not consistent.

    For information review this linkhttp://www.microsoft.com/technet/tr...rodtechnol/sql/maintain/optimize/ANSvcsPG.asp and though this relates to the optimisation, I'm sure it may lead to some solution.

    HTH

    Satya SKJ
    Moderator
    http://www.SQL-Server-Performance.Com/forum
    This posting is provided “AS IS” with no rights for the sake of knowledge sharing.
  5. stevefromOZ New Member

    It may sound obvious, but have you checked your Event Log for messages? ie run out of disk on the drive holding the temp/building files that AS uses when building cubes. Are you using a SAN? Have you had any problkems where the SAN drives drop off?

    Lastly, have you migrated your repository to MSSQL? Always a good idea, MSaccess is not the best multi-user DB environment...

    HTH,

    Steve.
  6. thomas New Member

    Steve, thanks. Yep the error log looks fine apart from errors directly associated with this problem. We are not using a SAN. We have migrated the repository to SQL Server too. This error is intermittent: for the last 3 nights, cube re-processing has been fine. I think I'll log a support call next time it occurs, to see if they think I'm missing anything else.

    Tom Pullen
    DBA, Oxfam GB
  7. satya Moderator

    Any chance of reproducing the process on other box with similar setup?

    Satya SKJ
    Moderator
    http://www.SQL-Server-Performance.Com/forum
    This posting is provided “AS IS” with no rights for the sake of knowledge sharing.
  8. thomas New Member

    I could set it up on another server but I won't for the time being - this strange problem has mysteriously stopped happening. Not ideal when things get better for no known reason, but it's working, my users aren't bleating, so I'm leaving well alone. Thanks for your help and suggestions.

    Tom Pullen
    DBA, Oxfam GB

Share This Page