Red Hat Bugzilla – Bug 474446
init script restart doesn't always restart condor
Last modified: 2009-02-04 11:06:22 EST
Description of problem:
[root@north-11 condor]# /etc/init.d/condor restart
Stopping Condor daemons: [ OK ]
Starting Condor daemons:
[root@north-11 condor]# ps -ef | grep condor
root 27760 27153 0 15:10 pts/2 00:00:00 grep condor
Version-Release number of selected component (if applicable):
Steps to Reproduce:
This is likely the result of switching to -QUIT the condor_master. When killproc is given a signal it never sleep while stopping the daemon. The lack of sleep means that on a restart (which is stop + start) the condor_master still exists when start is executed. The result is start bails out and the condor_master eventually exists.
It's probably sufficient to add a sleep 5 (which is the killproc delay) to stop() in the init script. Though that's kinda sad.
Actually, when you pass a signal to killproc() it no longer rm's the pidfile for you, so stop() must do that itself.
This is address in dffb2e11, will be part of 7.2.0-0.8
I'll buy this, based on code inspection and the fact that I haven't run into this recently. No state "SEEMS_TO_BE_VERIFIED", so moving to VERIFIED.
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.