Do you have any more logs about the problem? It seems like somewhere down the interceptor chain you hit a problem (possibly a lock acquisition timeout) and this causes your tx to rollback. Now unfortunately WL's TM throws it's own exception when I call setRollbackOnly() so I don't have a chance to log the real exception. (JBCACHE-1130) Once the above JIRA is fixed, I gsuest getting a snapshot of the 1.4.x branch from CVS and trying with that - it will log more info on the underlying exception. Also, how do you have your data organised? Remember that all locking in JBoss Cache happens on the node level. Having a lot of threads reading/writing to the same node can cause concurrency problems and hence timeouts, tx rollbacks, etc. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4062991#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...