Menu

#42 Hundreds of interserver processes running

closed
None
5
2005-03-26
2004-03-23
Jacob Eckel
No

After using Track+ (V2.1) on Firebird (V1.0.2) for a week
it opens hundreds of interserver.exe processes (all
visible in Windows Task Manager). The processes
consume up to 300M virtual memory each and
effectively disable the server, so I need to restart
Tomcat on weekly basis.
Any ideas?

Discussion

  • Nobody/Anonymous

    Logged In: NO

    I am waiting for this problem. :( T+ at the moment don't use
    (and don't have) Connections Pool. It is the reason why
    amount of opened connections so big.

     
  • Alexei Khatskevich

    Logged In: YES
    user_id=940452

    Sorry, forgot to login :). The comment before was from me.

     
  • Joerg Friedrich

    Joerg Friedrich - 2004-04-07

    Logged In: YES
    user_id=364010

    Try to change this line in Torque.properties to a lower value
    like 5 or so:

    torque.dsfactory.track.pool.defaultMaxConnections=80

    Let us know if that helps.

     
  • Joerg Friedrich

    Joerg Friedrich - 2004-09-09

    Logged In: YES
    user_id=364010

    One more comment: I just tested the following setting for the
    new release for Track+ in Torque.properties:

    torque.dsfactory.track.factory=
    org.apache.torque.dsfactory.SharedPoolDataSourceFactory
    torque.dsfactory.track.pool.maxActive=30
    torque.dsfactory.track.pool.testOnBorrow=true
    torque.dsfactory.track.pool.validationQuery=SELECT PKEY
    FROM TSTATE

    which gets rid of the old Torque pool. You have to remove all
    other entries in Torque.properties containing keys with
    torque.dsfactory.track.pool. Please let me know if this gets
    rid of your problem.

     
  • Joerg Friedrich

    Joerg Friedrich - 2004-09-09
    • assigned_to: nobody --> friedj
     
  • Joerg Friedrich

    Joerg Friedrich - 2005-03-26
    • status: open --> closed
     
  • Joerg Friedrich

    Joerg Friedrich - 2005-03-26

    Logged In: YES
    user_id=364010

    Since I have not heard anything on this anymore, I suppose
    the problem is solved. Please reopen if that should not be
    the case.

     

Log in to post a comment.