You should produce a Stack trace of your server when things are slowing down. Add -XX:+PrintClassHistogram to your VM Options. and call kill -3 when the process is getting slow (before it crashed). -XX:+PrintClassHistogram or kill -3 wouldn't hurt your production system. But if you can replicate this on a test environment it's better. This link will provide you some extra information: http://www.jboss.org/community/wiki/StackTrace You are probably running out of memory. I used to work with Solaris years ago, and running out of memory would certainly crash the VM on Solaris. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4232072#... Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...