Jacek, You should be able to temporarily address these issues by enabling "lockParentForChildInsertRemove", This will force a write a lock on the parent node for insert/update. As brian suggests Scenario 3 can be addressed by using REPEATABLE_READ, although that will increase the possibility of lock contention. The new MVCC locking design that Brian refers to will move to FQN based locking, instead of node locking. -Jason View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4082735#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...