8 Replies Latest reply on Aug 15, 2007 10:12 AM by kukeltje

    multiple jobexecutor threads?

    ax666

      hi, I'm having trouble to have the JobExecutor run multiple threads. there seems to be an issue with hibernate sync. is there anything I can do about it (configuration?) ?
      the flow I'm testing it with has sub processes (which are launched after a fork node) which I'd like to run parallel. everything's fine with just 1 JobExecutor thread, but then jobs get executed one by one. it is a tomcat application, the JobExecutor is started using JMX (jconsole), classloader for the JobExecutor is set to the webapp's classloader.
      any quick hints?

      thanks, alex

        • 1. Re: multiple jobexecutor threads?
          ax666

          all nodes in process and subprocesses are async.

          14:49:16,412 JbpmJobExector:192.168.41.136:5 ERROR AbstractFlushingEventListener:108 - Could not synchronize database state with session
          org.hibernate.StaleObjectStateException: Row was updated or deleted by another transaction (or unsaved-value mapping was incorrect): [org.jbpm.job.ExecuteNodeJob#12]
           at org.hibernate.persister.entity.AbstractEntityPersister.check(AbstractEntityPersister.java:1765)
           at org.hibernate.persister.entity.AbstractEntityPersister.update(AbstractEntityPersister.java:2407)
           at org.hibernate.persister.entity.AbstractEntityPersister.updateOrInsert(AbstractEntityPersister.java:2307)
           at org.hibernate.persister.entity.AbstractEntityPersister.update(AbstractEntityPersister.java:2607)
           at org.hibernate.action.EntityUpdateAction.execute(EntityUpdateAction.java:92)
           at org.hibernate.engine.ActionQueue.execute(ActionQueue.java:248)
           at org.hibernate.engine.ActionQueue.executeActions(ActionQueue.java:232)
           at org.hibernate.engine.ActionQueue.executeActions(ActionQueue.java:140)
           at org.hibernate.event.def.AbstractFlushingEventListener.performExecutions(AbstractFlushingEventListener.java:298)
           at org.hibernate.event.def.DefaultFlushEventListener.onFlush(DefaultFlushEventListener.java:27)
           at org.hibernate.impl.SessionImpl.flush(SessionImpl.java:1000)
           at org.hibernate.impl.SessionImpl.managedFlush(SessionImpl.java:338)
           at org.hibernate.transaction.JDBCTransaction.commit(JDBCTransaction.java:106)
           at org.jbpm.persistence.db.DbPersistenceService.commit(DbPersistenceService.java:253)
           at org.jbpm.persistence.db.DbPersistenceService.close(DbPersistenceService.java:211)
           at org.jbpm.svc.Services.close(Services.java:222)
           at org.jbpm.JbpmContext.close(JbpmContext.java:139)
           at org.jbpm.job.executor.JobExecutorThread.acquireJobs(JobExecutorThread.java:141)
           at org.jbpm.job.executor.JobExecutorThread.run(JobExecutorThread.java:56)
          


          • 2. Re: multiple jobexecutor threads?

            Alex,

            If you search for "StaleObjectStateException", you'll see other instances on this forum where it occurs after a fork. I don't know if it has the same cause as yours, because you didn't report any details about what you're doing. Unfortunately, no one bothered to follow up with a resolution or workaround, if one was found.

            Also, see http://jira.jboss.com/jira/browse/JBPM-626. I'm not sure if it's relevant. It's reported in enough detail that you can probably tell. But again, unfortunately, no resolution or workaround was reported, if found.

            I suggest that you prove to yourself that when your async events kick off, you're picking up the different node instances - not accidentally grabbing the same node instance, somehow.

            Beyond that, sorry, I got nuthin'.
            Please report what you find - this will be my problem soon, too!

            -Ed Staub

            • 3. Re: multiple jobexecutor threads?
              ax666

              seems like different threads try to work on the same job. in general concurrent execution is obviously not well tested. I hope our problem is limited to JobSession.getFirstAquireableJob...

              • 4. Re: multiple jobexecutor threads?
                ax666

                ok, after synchronizing JobExecutorThread.acquireJobs() it seems to work

                • 5. Re: multiple jobexecutor threads?

                  Alex,

                  Can you be more specific?
                  I read your note as saying you changed the method signature to

                  protected synchronized Collection acquireJobs()

                  but that shouldn't help, because it will synchronize on the thread object. But this won't prevent other threads from hitting it at the same time. Unless acquireJobs() is being called from a subclass of JobExecutorThread, the only place it is called from is run(), so there's not even a remote possibility of a cross-thread invocation somehow.

                  If I understood you correctly, and this made the problem go away, I suspect it's just hiding for a while...

                  -Ed Staub

                  • 6. Re: multiple jobexecutor threads?
                    ax666

                    http://jira.jboss.com/jira/browse/JBPM-974

                    what I did as workaround is to synchronize on a static variable inside of acquireJobs(), this way locks are applied and concurrency works.
                    don't know if thats a good solution as I'm not much into jbpm internals...

                    • 7. Re: multiple jobexecutor threads?
                      ax666

                      well, next problem, on a join node all incoming branches try to update the parent token, which fails, I think I read about this already somewhere in the forum... altogether it seems that multithreaded async processes have never been tested...

                      • 8. Re: multiple jobexecutor threads?
                        kukeltje

                        for anyone interested:

                        http://jira.jboss.com/jira/browse/JBPM-995

                        closed....