Pete and Mike, There must be at least a third element involved in the deadlock you described. If there is one read/write lock and one of the threads acquire the read lock, while a second one waits on a write lock, the first thread will finish execution and the second will be resumed right away. A deadlock wouldn't occur also if one thread holds the write lock while the second one is waiting to retrieve the read lock. I need to see the piece that is missing. Where are the threads that hold these locks, and why won't they let this lock go. Is there a way I can reproduce this? If you could attach your app to a Jira issue, it would be much appreciated. Cheers View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4222133#... Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...