In the JMS tests we're now using a small wrapper around the org.jnp.server.Main class which introduces a small delay on stop().

This is because there appears to be a bug in the class which means if you start() it again quickly after stopping it, you can get a

java.rmi.server.ExportException: Port already in use: 1098; nested exception is:
java.net.BindException: Address already in use

exception, as the socket doesn't seem to get closed synchronously with the close() method. I think this has been causing intermittent test suite failures