Right now the entire infrastructure is set up to depend entirely on which configuration that was chosen for the runtime. The reasons are many. I originally tried having a runtime which was just the server home, and then let the "server" object keep track of which configuration was in use, but then when I wanted to use these runtimes for webtools-type projects, the "runtime" didnt have access to the "server", and I couldn't get a reference to the jars in a specific configuration's folder / deploy directory. So it became clear the runtime needed to hold all the data, including which configuration was being run at that moment. I realize you've been overriding it with command-line arguments, but it's simply not set up right now to do that. This is the reason for all of your listed problems. It's the reason the jndi port keeps getting reset (although that does sound fishy), it's the reason the publish directory is static, and it's probably the reason the xpaths are not showing you the values you expect. In all cases, its drawing from the runtime configuration you've chosen, which seems to be default rather than the ssrv10 you've put in via command line args. In all honesty, I just didn't expect people to be using the infrastructure in this way, and so I didn't plan for it. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3982502#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...