Check if groups is really running - could be java/NMC thing. Check for savegrp process. If it is there it is possible that it can't kill session from client for some reason. Maybe your job stopped, but what you wished to say is that you still have request for tape (which for example normal when it comes to initiated Oracle backups) - I thought of mentioning this as lately you had Oracle related thread too.

Try restarting your gstd / NMC daemons. Most likely to be a java thing. You could also check if save, or the appropriate backup command in case of any modules, is running on the client.

When you start to see those kind of things its normally the gstd daemon playing up, just stop and start it using:

/etc/init.d/gstd stop
/etc/init.d/gstd start

Should normally fix this. It may be that the jobsdb is struggling to keep up - go into Configuration, right click on your server and check the values for jobsdb retention and jobsdb maximum size; retention should be 3 days and max size a lot higher than EMC's documentation states, we have ours set to 200,000KB...