Red Hat Bugzilla – Bug 153741
gulm init script should not fail if tag doesn't exist in cluster.conf
Last modified: 2009-04-16 15:56:45 EDT
Description of problem:
Starting lock_gulmd: no <gulm> section detected in /etc/cluster/cluster.conf
A better messages might be:
Starting lock_gulmd: Skipping because no <gulm> section detected in
Version-Release number of selected component (if applicable):
[root@tank-03 ~]# lock_gulmd -V
lock_gulmd 1.0-0.pre26 (built Mar 28 2005 18:04:15)
Copyright (C) 2004 Red Hat, Inc. All rights reserved.
Except for the fact that lock_gulmd didn't actually startup. "OK" is just as
wrong as "FAILED".
Trouble shooting working systems that are reporting failures are easier to deal
with than broken systems reporting success.
Will look into a different error message to appease both camps
gulm init.d script now prints WARNING instead of FAILURE.
Checking in gulm/init.d/lock_gulmd;
/cvs/cluster/cluster/gulm/init.d/lock_gulmd,v <-- lock_gulmd
new revision: 184.108.40.206; previous revision: 220.127.116.11
where did the "no <gulm> section detected in /etc/cluster/cluster.conf" go?
Now it just says:
> where did the "no <gulm> section detected in /etc/cluster/cluster.conf" go?
The warning message is sent to syslog.
Since lock_gulmd does not start, it should not print "OK". It instead fails to
start in a manor that is determined to be acceptable (hence why it's exit status
is 0 and not 1) but in a way that still provides the user with a clue that they
might have a configuration problem.
The bug states that lock_gulmd should not fail, so it doesn't. It just provides
the user with feedback to start looking into what the problem might be. If they
don't like that message, they should disable lock_gulmd with chkconfig.
It is my opinion that this will be much easier to trouble shoot than pretending
that everything is OK.
I see the message in the syslog, fix verified.