Messaging DLQ not clustering correctly

[tnine]

Hi All,
I have an old EJB 2 MDB that listens from a queue and processes messages. When a message fails, it goes to the default DLQ. I have 2 nodes in my cluster, prodjboss1, and prodjboss2. We have a process that runs on a timer and wakes up every 30 seconds to check the DLQ for failed messages. Those messages are then read from the queue, and written to a file on disk for manual inspection. We ran into an issue today, where the DLQ on prodjboss1 was empty but the DLQ on prodjboss2 had a message in it. I've configured the DLQ to be clustered, but it doesn't appear that the nodes are clustering. Whenever the client wakes up and inspects the DLQ, it never receives any messages if they're not on prodjboss1. Here is my DLQ config from both nodes as well as my client's JNDI properties. Is something configured incorrectly for clustering of the DLQ? All my other queues that are clustered work as intended, messages are shared across nodes and clients.

Thanks,
Todd

DLQ config

<mbean code="org.jboss.jms.server.destination.QueueService"
 name="jboss.messaging.destination:service=Queue,name=DLQ"
 xmbean-dd="xmdesc/Queue-xmbean.xml">
 <depends optional-attribute-name="ServerPeer">
 jboss.messaging:service=ServerPeer
 </depends>
 <depends>jboss.messaging:service=PostOffice</depends>
 <attribute name="Clustered">true</attribute>
 </mbean>

Client config

java.naming.factory.initial=org.jnp.interfaces.NamingContextFactory
java.naming.factory.url.pkgs= org.jboss.naming:org.jnp.interfaces
#Development host
#java.naming.provider.url=jnp://localhost:1100
#Production host
java.naming.provider.url=jnp://prodjboss1:1100,jnp://prodjboss2:1100

Versions:
JBoss 4.2.2 GA custom config. All configuration with web serivces removed
JBoss Messaging 1.4
Java 1.5
Redhat 5

[kenwdelong]

Did you ever figure this out? We have the same problem.

[tnine]

Unfortunately not. The only way we got around it was to make sure that prodjboss1 starts before prodjboss2. It seems that the first node in the cluster receives messages that fall into the DLQ. While this is a hack, it does seem to work for us. Unfortunately it kills the whole point of messaging, which was to have a true peer messaging service. If you come up with anything better, please post it here. I'd love to know how to fix this.