none
How to purger a 4 Gb HPCCluster database RRS feed

  • Question

  • Hi,

    We have an HPC2008 16 nodes with an SQLExpress 2005 database. The HLCCLuster.mdf is full (4Gb) and we cant submit a new job because this is a limit of SQL Express.

    There is an BIG 2GB LogEntries table 

    We set a 1 day log retention ==> KO

    I try to shrink the database ==> KO, there is really 4 gb data.

    Do you know how can I purge any log ... to reduce the size of the database ?

     

    Tuesday, September 21, 2010 6:16 PM

All replies

  • Hi,

    One other thing you can do is to increase the frequency of old jobs cleanup.

    You can change default values of the following cluster parameters (using 'cluscfg setparams'):

    TtlCompletedJobs                 : 1 (default 5)
    JobCleanUpHour                   : 2

    For above setting, jobs which are in final state and are 1 day old will be deleted from database at 2 AM every day. Job scheduler service (HpcScheduler) is also checking for old jobs to remove at startup, so you can provoke this process by restarting it.

    Regards,
    Łukasz

    Friday, September 24, 2010 8:05 PM
  • Hi,

     

    We have also an HPC2008 with 65 nodes and the HPCCluster.mdf is full (4Go). New jobs are rejected. Any idea to purge quickly and easily the database?

     

    Many Thanks for an answer...

    Friday, October 15, 2010 9:53 AM