I did some searching on Google and am coming up blank on this particular problem. Before I dig into the JobTracker startup code and Jetty, does anyone know why a jobtracker would start up fine, run jobs successfully, show no problems int he logs, but I would get a 404 error from Jetty (RequestURI=/jobtracker.jsp) when accessing the default jobtracker UI at 50030 either remotely or locally via lynx? It worked fine prior to the restart.
My current thought is that there is some configuration environment variable that isn't set when I'm manually restarting just that instance vs. starting up the whole cluster with cluster scripts.
(I'm running this on an HBase cluster w/ 0.90.0 running Cloudera's hadoop dist 0.20.2+737)