In this case the problem was with the HAPartition channel (cluster-service.xml). Earlier it was the Tomcat clustering channel (tc5-cluster.sar). The state transfer failed because node B didn't get a response to the state transfer request from A within 30 secs. The bits of logs from A don't cover the period when B requested state, so it's hard to say why. Suggest you: 1) Go through the process outlined at http://wiki.jboss.org/wiki/Wiki.jsp?page=TestingJBoss with the configs you use for both cluster-service.xml and tc5-cluster.sar. Most likely there is some connectivity issue. 2) Get the server clocks in sync. Won't fix the problem, but trying to interpret cluster logs with out of sync timestamps is way too much of a pain. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4025546#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...