0 Replies Latest reply on Jul 28, 2016 4:33 AM by jon286

    ESM alert table partitions purging too soon

    jon286

      I'm having an issue with the alert table partitions being purged when I think they shouldn't be. This has happened once before around the time another fault was being resolved, but not to such a level and it seemed to sort itself out.

       

      Our event counts have increased so I'm not expecting to retain that many days in normalized events, but less than a week isn't particularly useful. As I type, this morning it has purged the last two inactive partitions, leaving that empty, and has just started to fill the 5th active partition. Previously, it would always tick over with 4 inactive partitions sitting in the storage volume.

       

      Has anyone experienced this and/or has any knowledge how to stop this happening?

       

      We're using 9.5.2 and I have a 217GB share configured for inactive partitions.

      Database>Data Retention is set to keep all data allowed by the system, and the NitroTID output shows the following;

       

      Partition Type          - Time based partition
        Total Partitions        - 5
        Total Active            - 5
        Total Inactive          - 0
        Partitioning Field      - LastTime
        Partitioning Time Unit  - 1 day(s)
        Min Records / Partition - 25,000,000
        Max Records / Partition - 25,000,000
        Allowed Attached        - 115,000,000 record(s) OR 5 partition(s)
        Max Before Deletion     - 265,000,000 record(s) OR 11 partition(s)
        Max Emtpy Gap           - 30 partition unit(s)
      

       

      Partitions on average cover ~30 hours weekdays, 40 or so weekends.