JMS Timer Message ClassCastException

Question

          Hi,
          we're using Weblogic's JMS extension of setting a send delay, e.g.:
          ((WLMessageProducer) publisher).setTimeToDeliver(new Long(notification.getDelayMinutes()
          * 60 * 1000).longValue());
          This works quite well. But it seems that the server is unable to delete messages
          once they're triggered.
          In the logs we periodically see something like:
          ####<10.12.2003 15.01 Uhr IRT> <Error> <Kernel> <DP03W01> <PostServer>
          <ExecuteThread: '3' for queue: 'JMS.TimerClientPool'> <<WLS Kernel>> <>
          <BEA-000802> <ExecuteRequest failed
          java.lang.ClassCastException.
          java.lang.ClassCastException
          at weblogic.jms.backend.BETopic._addMessageToConsumers(BETopic.java:409)
          at weblogic.jms.backend.BETopic.access$000(BETopic.java:48)
          at weblogic.jms.backend.BETopic$1.expireTimeout(BETopic.java:343)
          at weblogic.jms.backend.BETimerNode.execute(BETimerNode.java:132)
          at weblogic.kernel.ExecuteThread.execute(ExecuteThread.java:197)
          at weblogic.kernel.ExecuteThread.run(ExecuteThread.java:170)
          >
          So far so good. As long as the server is running the triggered messages won't
          get executed again. But once we restart the server. These messages come back to
          life.
          Does anyone have an idea what could cause such a behaviour?
          Thanks,
          Fabian
          

Answer 1

Hi Fabian,
          That would be a bug, which I suspect it is already a known one.
          Search the customer support web-site, and/or file
          a case with customer support. Meanwhile, since you don't
          want to timer messages to survive a server crash, you may
          want to make sure that your application publishes them
          with a non-persistent QOS to begin with - that may work
          around the problem.
          Tom
          Fabian Crabus wrote:
          > Hi,
          >
          > we're using Weblogic's JMS extension of setting a send delay, e.g.:
          > ((WLMessageProducer) publisher).setTimeToDeliver(new Long(notification.getDelayMinutes()
          > * 60 * 1000).longValue());
          >
          > This works quite well. But it seems that the server is unable to delete messages
          > once they're triggered.
          > In the logs we periodically see something like:
          > ####<10.12.2003 15.01 Uhr IRT> <Error> <Kernel> <DP03W01> <PostServer>
          > <ExecuteThread: '3' for queue: 'JMS.TimerClientPool'> <<WLS Kernel>> <>
          > <BEA-000802> <ExecuteRequest failed
          > java.lang.ClassCastException.
          > java.lang.ClassCastException
          > at weblogic.jms.backend.BETopic._addMessageToConsumers(BETopic.java:409)
          > at weblogic.jms.backend.BETopic.access$000(BETopic.java:48)
          > at weblogic.jms.backend.BETopic$1.expireTimeout(BETopic.java:343)
          > at weblogic.jms.backend.BETimerNode.execute(BETimerNode.java:132)
          > at weblogic.kernel.ExecuteThread.execute(ExecuteThread.java:197)
          > at weblogic.kernel.ExecuteThread.run(ExecuteThread.java:170)
          >
          >
          > So far so good. As long as the server is running the triggered messages won't
          > get executed again. But once we restart the server. These messages come back to
          > life.
          >
          > Does anyone have an idea what could cause such a behaviour?
          >
          > Thanks,
          >
          > Fabian
          >
          >