No, in different places, not just gravitation. the first stack trace shows contention during a state transfer for the new buddy. There is stuff being written to cache1 while cache1 is trying to get state to send to cache2. And there is contention here. Perhaps increasing your lock acquisition timeout for getting state for the buddy would help - according to the stack trace, it is set to 400millis!! For the 2nd stack trace, the problem has to do with transactions and 2-phase commits which can happen when there are a lot of frequent writes to the same region. Set your configuration to use synchronous commit and rollbacks and you won't see this. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4152671#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...