Re: [Resin-interest] JMS Exceptions, 3.1.8
I need to investigate further for patterns. On Tue, Jan 20, 2009 at 11:28 AM, Scott Ferguson wrote: > > On Jan 20, 2009, at 9:11 AM, Tyson Weihs wrote: > >> The expected type is an instance of one of our serializable classes. > > Hmm. Is there any pattern when this happens? > >> Fixing that back in 3.1.8/9 would be greatly appreciated. > > Yep. I'll check. > > -- Scott > >> >> >> On Tue, Jan 20, 2009 at 11:01 AM, Scott Ferguson >> wrote: >>> >>> On Jan 18, 2009, at 1:07 PM, tweihs wrote: >>> We finally upgraded to 3.1.8! ("Hallelujah, holy s**t, where's the tylenol" - Clark W. Griswald) The pain moving from 3.1.3, which we stuck ourselves on not knowing it was a dev release, was intense (changed just about every tag, new super classes for many of our objects, made the jump to configuring all our beans via IoC), but it seems to be paying off (better production performance, easier development using annotations, etc.). We're seeing the following exception in production now, though: >>> >>> I've filed this as a bug, although I think we've already fixed it (it >>> might need to be fixed back in 3.1.x) >>> [21:00:07.701] {resin-21} MessageConsumerImpl[FileQueue[executorQueue]]: message listener '_ejb.defaultexecutor.executor__...@149dd5b' failed for message 'ObjectMessageImpl[ID:4U5naCkuy/cAEe3cHn/wAC]' with exception [21:00:07.701] {resin-21} java.lang.RuntimeException: java.lang.NullPointerException [21:00:07.701] {resin-21} java.lang.RuntimeException: java.lang.NullPointerException [21:00:07.701] {resin-21} at com.caucho.jms.connection.JmsSession.commit(JmsSession.java:1018) [21:00:07.701] {resin-21} at com.caucho.transaction.TransactionImpl.commit(TransactionImpl.java: 658) [21:00:07.701] {resin-21} at com .caucho .transaction .TransactionManagerImpl.commit(TransactionManagerImpl.java:271) [21:00:07.701] {resin-21} atmore We're using file queues for storage. We're also seeing class cast exceptions occasionally when pulling the message objects out of the queue, where Message.getObject() is returning a java.lang.Long (and we've checked code to see that we're not injecting Longs). Any ideas? >>> >>> Do you know what the expected type is? It's conceivable that a Byte >>> is getting converted to a Long. >>> >>> -- Scott >>> -- View this message in context: http://www.nabble.com/JMS-Exceptions%2C-3.1.8-tp21532983p21532983.html Sent from the Resin mailing list archive at Nabble.com. ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest >>> >>> >>> >>> ___ >>> resin-interest mailing list >>> resin-interest@caucho.com >>> http://maillist.caucho.com/mailman/listinfo/resin-interest >>> >> >> >> ___ >> resin-interest mailing list >> resin-interest@caucho.com >> http://maillist.caucho.com/mailman/listinfo/resin-interest > > > > ___ > resin-interest mailing list > resin-interest@caucho.com > http://maillist.caucho.com/mailman/listinfo/resin-interest > ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest
Re: [Resin-interest] JMS Exceptions, 3.1.8
On Jan 20, 2009, at 9:11 AM, Tyson Weihs wrote: > The expected type is an instance of one of our serializable classes. Hmm. Is there any pattern when this happens? > Fixing that back in 3.1.8/9 would be greatly appreciated. Yep. I'll check. -- Scott > > > On Tue, Jan 20, 2009 at 11:01 AM, Scott Ferguson > wrote: >> >> On Jan 18, 2009, at 1:07 PM, tweihs wrote: >> >>> >>> We finally upgraded to 3.1.8! ("Hallelujah, holy s**t, where's the >>> tylenol" - >>> Clark W. Griswald) >>> >>> The pain moving from 3.1.3, which we stuck ourselves on not knowing >>> it was a >>> dev release, was intense (changed just about every tag, new super >>> classes >>> for many of our objects, made the jump to configuring all our beans >>> via >>> IoC), but it seems to be paying off (better production performance, >>> easier >>> development using annotations, etc.). >>> >>> We're seeing the following exception in production now, though: >> >> I've filed this as a bug, although I think we've already fixed it (it >> might need to be fixed back in 3.1.x) >> >>> >>> >>> [21:00:07.701] {resin-21} >>> MessageConsumerImpl[FileQueue[executorQueue]]: >>> message listener '_ejb.defaultexecutor.executor__...@149dd5b' failed >>> for >>> message 'ObjectMessageImpl[ID:4U5naCkuy/cAEe3cHn/wAC]' with >>> exception >>> [21:00:07.701] {resin-21} java.lang.RuntimeException: >>> java.lang.NullPointerException >>> [21:00:07.701] {resin-21} java.lang.RuntimeException: >>> java.lang.NullPointerException >>> [21:00:07.701] {resin-21} at >>> com.caucho.jms.connection.JmsSession.commit(JmsSession.java:1018) >>> [21:00:07.701] {resin-21} at >>> com.caucho.transaction.TransactionImpl.commit(TransactionImpl.java: >>> 658) >>> [21:00:07.701] {resin-21} at >>> com >>> .caucho >>> .transaction >>> .TransactionManagerImpl.commit(TransactionManagerImpl.java:271) >>> [21:00:07.701] {resin-21} atmore >>> >>> We're using file queues for storage. We're also seeing class cast >>> exceptions occasionally when pulling the message objects out of the >>> queue, >>> where Message.getObject() is returning a java.lang.Long (and we've >>> checked >>> code to see that we're not injecting Longs). Any ideas? >> >> Do you know what the expected type is? It's conceivable that a Byte >> is getting converted to a Long. >> >> -- Scott >> >>> >>> >>> >>> -- >>> View this message in context: >>> http://www.nabble.com/JMS-Exceptions%2C-3.1.8-tp21532983p21532983.html >>> Sent from the Resin mailing list archive at Nabble.com. >>> >>> >>> >>> ___ >>> resin-interest mailing list >>> resin-interest@caucho.com >>> http://maillist.caucho.com/mailman/listinfo/resin-interest >> >> >> >> ___ >> resin-interest mailing list >> resin-interest@caucho.com >> http://maillist.caucho.com/mailman/listinfo/resin-interest >> > > > ___ > resin-interest mailing list > resin-interest@caucho.com > http://maillist.caucho.com/mailman/listinfo/resin-interest ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest
Re: [Resin-interest] JMS Exceptions, 3.1.8
The expected type is an instance of one of our serializable classes. Fixing that back in 3.1.8/9 would be greatly appreciated. On Tue, Jan 20, 2009 at 11:01 AM, Scott Ferguson wrote: > > On Jan 18, 2009, at 1:07 PM, tweihs wrote: > >> >> We finally upgraded to 3.1.8! ("Hallelujah, holy s**t, where's the >> tylenol" - >> Clark W. Griswald) >> >> The pain moving from 3.1.3, which we stuck ourselves on not knowing >> it was a >> dev release, was intense (changed just about every tag, new super >> classes >> for many of our objects, made the jump to configuring all our beans >> via >> IoC), but it seems to be paying off (better production performance, >> easier >> development using annotations, etc.). >> >> We're seeing the following exception in production now, though: > > I've filed this as a bug, although I think we've already fixed it (it > might need to be fixed back in 3.1.x) > >> >> >> [21:00:07.701] {resin-21} >> MessageConsumerImpl[FileQueue[executorQueue]]: >> message listener '_ejb.defaultexecutor.executor__...@149dd5b' failed >> for >> message 'ObjectMessageImpl[ID:4U5naCkuy/cAEe3cHn/wAC]' with exception >> [21:00:07.701] {resin-21} java.lang.RuntimeException: >> java.lang.NullPointerException >> [21:00:07.701] {resin-21} java.lang.RuntimeException: >> java.lang.NullPointerException >> [21:00:07.701] {resin-21} at >> com.caucho.jms.connection.JmsSession.commit(JmsSession.java:1018) >> [21:00:07.701] {resin-21} at >> com.caucho.transaction.TransactionImpl.commit(TransactionImpl.java: >> 658) >> [21:00:07.701] {resin-21} at >> com >> .caucho >> .transaction >> .TransactionManagerImpl.commit(TransactionManagerImpl.java:271) >> [21:00:07.701] {resin-21} atmore >> >> We're using file queues for storage. We're also seeing class cast >> exceptions occasionally when pulling the message objects out of the >> queue, >> where Message.getObject() is returning a java.lang.Long (and we've >> checked >> code to see that we're not injecting Longs). Any ideas? > > Do you know what the expected type is? It's conceivable that a Byte > is getting converted to a Long. > > -- Scott > >> >> >> >> -- >> View this message in context: >> http://www.nabble.com/JMS-Exceptions%2C-3.1.8-tp21532983p21532983.html >> Sent from the Resin mailing list archive at Nabble.com. >> >> >> >> ___ >> resin-interest mailing list >> resin-interest@caucho.com >> http://maillist.caucho.com/mailman/listinfo/resin-interest > > > > ___ > resin-interest mailing list > resin-interest@caucho.com > http://maillist.caucho.com/mailman/listinfo/resin-interest > ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest
Re: [Resin-interest] JMS Exceptions, 3.1.8
On Jan 18, 2009, at 1:07 PM, tweihs wrote: > > We finally upgraded to 3.1.8! ("Hallelujah, holy s**t, where's the > tylenol" - > Clark W. Griswald) > > The pain moving from 3.1.3, which we stuck ourselves on not knowing > it was a > dev release, was intense (changed just about every tag, new super > classes > for many of our objects, made the jump to configuring all our beans > via > IoC), but it seems to be paying off (better production performance, > easier > development using annotations, etc.). > > We're seeing the following exception in production now, though: I've filed this as a bug, although I think we've already fixed it (it might need to be fixed back in 3.1.x) > > > [21:00:07.701] {resin-21} > MessageConsumerImpl[FileQueue[executorQueue]]: > message listener '_ejb.defaultexecutor.executor__...@149dd5b' failed > for > message 'ObjectMessageImpl[ID:4U5naCkuy/cAEe3cHn/wAC]' with exception > [21:00:07.701] {resin-21} java.lang.RuntimeException: > java.lang.NullPointerException > [21:00:07.701] {resin-21} java.lang.RuntimeException: > java.lang.NullPointerException > [21:00:07.701] {resin-21} at > com.caucho.jms.connection.JmsSession.commit(JmsSession.java:1018) > [21:00:07.701] {resin-21} at > com.caucho.transaction.TransactionImpl.commit(TransactionImpl.java: > 658) > [21:00:07.701] {resin-21} at > com > .caucho > .transaction > .TransactionManagerImpl.commit(TransactionManagerImpl.java:271) > [21:00:07.701] {resin-21} atmore > > We're using file queues for storage. We're also seeing class cast > exceptions occasionally when pulling the message objects out of the > queue, > where Message.getObject() is returning a java.lang.Long (and we've > checked > code to see that we're not injecting Longs). Any ideas? Do you know what the expected type is? It's conceivable that a Byte is getting converted to a Long. -- Scott > > > > -- > View this message in context: > http://www.nabble.com/JMS-Exceptions%2C-3.1.8-tp21532983p21532983.html > Sent from the Resin mailing list archive at Nabble.com. > > > > ___ > resin-interest mailing list > resin-interest@caucho.com > http://maillist.caucho.com/mailman/listinfo/resin-interest ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest