Atomikos Forum

Hi

I'm having an issue using Atomikos 3.9.3 with Spring 4.0.6 and JMS. The application starts a JMS connection to listen for messages. On receiving a message, it will process it and then post it on another JMS destination.

When the process is started, and only if it has processed at least one message, it will throw a warning on closing the ConnectionPool for the publishing connection, like this:

WARN [shutdown] [ConnectionPool] atomikos connection pool 'jms-Publisher': connection is still in use on pool destroy: atomikos pooled connection for resource jms-TradeAllocationsEventP - please check your shutdown sequence to avoid heuristic termination of ongoing transactions!

It will be forcibly closed, but the transaction log is then left in a bad state. The next time the application is started, we get these messages continuously every ten seconds:

[2014-08-06 16:09:30.150]  WARN [Atomikos:7] [XAResourceTransaction] XAResourceTransaction app-localhost0002000005app-localhost17: no XAResource to commit - the required resource is probably not yet intialized?
[2014-08-06 16:09:30.150]  WARN [Atomikos:7] [CommitMessage] Unexpected error in commit
com.atomikos.icatch.HeurHazardException: Heuristic Exception
    at com.atomikos.datasource.xa.XAResourceTransaction.commit(XAResourceTransaction.java:707)
    at com.atomikos.icatch.imp.CommitMessage.send(CommitMessage.java:72)
    at com.atomikos.icatch.imp.PropagationMessage.submit(PropagationMessage.java:83)
    at com.atomikos.icatch.imp.Propagator$PropagatorThread.run(Propagator.java:79)
    at com.atomikos.icatch.imp.Propagator.submitPropagationMessage(Propagator.java:58)
    at com.atomikos.icatch.imp.HeurHazardStateHandler.onTimeout(HeurHazardStateHandler.java:131)
    at com.atomikos.icatch.imp.CoordinatorImp.alarm(CoordinatorImp.java:933)
    at com.atomikos.timing.PooledAlarmTimer.notifyListeners(PooledAlarmTimer.java:112)
    at com.atomikos.timing.PooledAlarmTimer.run(PooledAlarmTimer.java:99)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
    at java.lang.Thread.run(Thread.java:722)

The only way to avoid these messages is to delete the transaction log file before restarting the app.

I have verified that the UserTransactionManager is stopped prior to trying to close the connection pools. The UserTransactionServiceImp is also closed prior to closing the connection pools, and I have tried both, shutdownWait and shutdownForce destroy methods to no avail.

After much hair pulling, we have run out of ideas. Is there a known bug we should be aware of? can anyone offer pointers? Please let me know if you require more specific details

Thanks in advance

I'm facing the same problem. I already checked that init-method is set on resources. Any new findings on this?

This is likely a configuration issue; I recommend getting a support subscription so we can put some expert time into it...

Thanks