This isn't from HA-JNDI, which doesn't use the UnifiedInvoker. I assume it's from the SLSB invocation. Suggest you post on the Remoting forum at http://www.jboss.com/index.html?module=bb&op=viewforum&f=222 to see if there's some config option you can set in conf/jboss-service.xml, jboss.remoting:service=Connector,transport=socket mbean to get the remoting client to fail over more quickly. Looks like the client is cycling through a pool of connections or something trying each one before finally giving up, throwing the exception and allowing the clustering logic to fail over. The individual attempts fail promptly enough, but the fact that its trying so many certainly adds up. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4152185#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...