Hey Ronald, Thanks for your reply. You're right that the super process class loader isn't propagated to the sub process and I wasn't expecting that (also verified by looking at the source) since that would seriously mess up the versioning semantics. However, I posted precisely because I did include the the class the (sub) process class loader was complaining it couldn't find in the (sub) process archive which meant it should have been able to find it. But... It isn't a bug. I reproduced the problem when testing from within my IDE, created a self-contained archive that I could use when reporting the issue on BugZilla but, surprisingly, running the archive via the command line made the problem disappear. A little housecleaning of referenced jars on our integration server and the problem disappeared there as well. I really don't see how this could fix the problem and that worries me a little, but that's an investigation for another day. Thanks again, /Robin View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4121664#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...