This may be a side effect of the current problems where Remoting is looped to establish the connection but it means that we need to cache the users username and password from the first attempt so we don't keep prompting them - the problem then is we do not know if the subsequent call is because their values were invalid or just because of the retries.