I'm giving this one a nudge - to see if anyone has experience similar issues. It appears that I'm more likely to get messages stuck if the receiver client attempts to fail-over and then times out after 60 seconds. It is a though something loses track of the correct node IDs to use when pulling messages across the cluster and it's only by seqeuentially restarting the cluster that they reset correctly. I'd appreciate if someone can say what the client does after timing out waiting for fail-over to start. Thanks. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4211592#... Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...