From: Greg Ercolano <erco@(email surpressed)>
Subject: IRUSH slow to remove cpus from jobs (eg. with "Job Edit")
   Date: Sat, 10 Sep 2005 16:00:17 -0700

Looks like 102.42 and 102.42a has a slowness problem where,
when the user tries to remove many cpus (eg. using IRUSH "Job Edit"),
it takes a long time for irush to complete the operation.

Turns out the problem is with the 'rush -rc .<tid> ..' command,
where the daemon is trying to do hostname lookups on the .<tid>
arguments, which take a few seconds each to fail.

This will be fixed in 102.42b.
I'll have a hot fix for 102.42 and 102.42a rush daemons shortly.

Watch this thread for details.