Red Hat Bugzilla – Bug 130996
crond crashes intermittently - need debug help
Last modified: 2007-11-30 17:10:48 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6)
Description of problem:
Crond seems to crash intermittently. I need some help obtaining debug
information to post. Simply starting it with the service command puts
me back in business.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Start crond
2. Wait for crash ( can't give amount of time
3. Restart it
You should really upgrade to FC2 - FC1 has been "retired" .
You are running vixie-cron-3.0.1-76 ?
vixie-cron-4.1-FC2_3 is the latest recommended version for Fedora.
If you can't upgrade, or you upgrade and the crash still occurs,
please do the following :
$ grep crond /var/log messages | tail -10 > /tmp/cron_crash.log
$ tail -10 /var/log/cron >> /tmp/cron_crash.log
and attach the resulting cron_crash.log file to this bug - that
should give us a way to start determining the problem.
I have already checked messages for goofy errors and am not getting
any. Can I enable a debug level in crond somehow? I can't apply that
latest RPM because FC1 doesnt have SELINUX support.
Is there any reason not to upgrade to FC2 ?
You can see exactly what's happening as follows - as root:
$ service crond stop
$ ulimit -c unlimited
$ cd /var/spool
$ crond -nx 'ext,sch,proc,pars,load,misc,test,bit' 2>&1 | tee
(all on one line).
This will run crond in the foreground, so you won't get another
prompt until it exits.
Then wait for the crash - how long does it take? - and then do:
$ gzip -9 /tmp/cron_dbg.log
and append the resulting /tmp/cron_dbg.log.gz file to this bug .
If there is a 'core.*' file in the /var/spool directory, please
also compress! the latest! core.* file and append the core.gz to
This is a UML box so upgrading is pretty much starting over. It
hasn't crashed yet since I last started it so That probably won't
work. I need to find out how often this is happening.
OK, well if you don't mind a huge /var/log/messages file, do this:
1. Enable core dumps: edit /etc/profile to say:
# No core files by default
# ulimit -S -c 0 > /dev/null 2>&1
ulimit -c unlimited >/dev/null 2>&1
2. Edit /etc/init.d/crond - where it says "daemon crond" in start(),
make it say:
daemon crond -x 'ext,sch,proc,pars,load,misc,test,bit'&
All the debug info will now appear in /var/log/messages,
and if crond cores, the corefile will be in /var/spool.
To find out how often it happens, create a cron job:
'* * * * * /bin/date>>/tmp/cron.alive.txt'
The last date & time string written to /tmp/cron.alive.txt
will be the last time cron ran its jobs.
This really doesn't happen with current version !
I'll work on getting a update to vixie-cron-4.1 out to
Fedora 1 & 2 updates.
Just clearing out old bugs here. This one is definitely no longer a