cron goes mad?

[Dave Alden]

On Thu, Jan 15, 2004 at 07:19:10AM -0800, David Jackson wrote:
> > Hi,
> >   I've got the exact same problem.  :-(  The load on this workstation is
> > up to 478.  :-)  I received no mail fron cron and there is nothing out
> > of the ordinary in any of the logfiles in /var/log.  I did a:
> >
> > find / -xdev -name core | xargs ls -ld
> >
> > and there are no core files from within the last week.
> 
> I previous poster said that core dumps are turned off by default? Did
> change your  settings?

Oops, no.  I've now modified my /etc/init.d/crond to add:

ulimit -S -c unlimited

after source /etc/init.d/functions -- we'll see what that does...


> Have you tried shutting down one cron job and a time, and run?

No, mainly because I can't recreate the problem.  I have 2 machines that
are exactly the same, the other one has never had this problem.  Both have
been running FC1 for ~1 month.  This is the first time this occurred.

In a seperate message you asked what I meant by "additional cron jobs".
All I meant was that in my /etc/cron.d directory I have a file called
cfengine that runs cfengine once every 5 minutes.  This script has been
running on 40 other linux boxes (a mixture of Redhat 9 and FC1) without
any problems -- and I've been running the same script for several years,
so I doubt it's the script.

Looking into it a little more, it looks like it occurred yesterday around
11:09am -- I went into the /proc and looked at several of the CROND processes
and they all seem to have gone in to disk wait around 11:09am.


> If a dump was created where would it be /tmp?

I don't know -- that's why I did the find.  :-)


> Has anyone checked the dev-list or bugzilla?

I checked bugzilla and didn't find anything.  I'm not ready to create a
new bugzilla report myself (I would like a little more info), maybe
someone else will (or I will if I can get it to do it again :-).

...dave