Menu

#11 Extra active-replicas sometimes start?

Bugs
closed-fixed
Agent (10)
4
2001-06-15
2001-03-05
No

Once in a while (or maybe after some amount of time),
the client will start a duplicate set of (at least)
active replica engines. Not sure why. Below is an
excerpt from a client that should have had only two
ARs:
---
{ReplicaActivationDemon

AR THREADS: {SRgroup:
[{ plasmid.replica.ActiveReplica@23ba56 pstate=
{ActiveReplicaPersistent: path C:\tmp\KCRW
pollTime 60000
plasmid.synchronicity.ReconObjSet@5b034538
vHistMapCOM.odi.util.OSHashMap@238d70 },
{ plasmid.replica.ActiveReplica@23ba3d pstate=
{ActiveReplicaPersistent: path C:\tmp\TestLR_1
pollTime 60000
plasmid.synchronicity.ReconObjSet@86ac988e
vHistMapCOM.odi.util.OSHashMap@2392ba }, {
plasmid.replica.ActiveReplica@23ba56 pstate=
{ActiveReplicaPersistent: path C:\tmp\KCRW
pollTime 60000
plasmid.synchronicity.ReconObjSet@5b034538
vHistMapCOM.odi.util.OSHashMap@238d70 }, {
plasmid.replica.ActiveReplica@23ba3d pstate=
{ActiveReplicaPersistent: path C:\tmp\TestLR_1
pollTime 60000
plasmid.synchronicity.ReconObjSet@86ac988e
vHistMapCOM.odi.util.OSHashMap@2392ba }]
}
---

Discussion

  • Joseph Hopfield

    Joseph Hopfield - 2001-04-04

    Logged In: YES
    user_id=4009

    This problem seems transient, and seems to happen when a
    new LR or new engine is discovered. Somehow a complete new
    set of threads is added (at least the old ones stay as
    counted threads).

    A restart of the agent always seems to fix this.

     
  • Joseph Hopfield

    Joseph Hopfield - 2001-04-04

    Logged In: YES
    user_id=4009

    The extra threads are the old dead versions of the engine
    and AR threads. Not really causing a problem (though should
    be zapped to allow gc!).

     
  • Joseph Hopfield

    Joseph Hopfield - 2001-04-04
    • priority: 5 --> 4
     
  • Joseph Hopfield

    Joseph Hopfield - 2001-06-15
    • assigned_to: nobody --> joehopfield
    • status: open --> closed-fixed
     
  • Joseph Hopfield

    Joseph Hopfield - 2001-06-15

    Logged In: YES
    user_id=4009

    Was failing to clear the SessionRunnableGroup's vector
    on shutdown of srgroup (which is done to restart a client
    after a config change).
    Joe

     

Log in to post a comment.

MongoDB Logo MongoDB