We have an odd problem where a cron job
sometimes doesnt run.
Its on a daily schedule, and it creates a log file
so we know exactly when it does or doesnt run.
Usually it runs fine.
Just occasionally, maybe a couple times per month,
it doesnt run, and I dont know why.

I have an idea (forgive me if this sounds stupid)
that maybe sometimes the system is so busy running
other high-priority tasks that it misses the
scheduled start-time for our job.
Is that possible, and if so, how would one test
for that?

[ Post a follow-up to this message ]

Chris Coyle <chrisco@bud.sprint.ca> wrote:
> other high-priority tasks that it misses the
> scheduled start-time for our job.

Cron job doesn't (usually) start on-time. The fact that you
schedule it for 2.00 am means only that it will start _around_ 2.00 am.
Most of the time the difference is minimal and the fact that is 2.00 or
2.00 3" isn't noticed. So, no. It's not that your system is too busy.

I'd start checking other possibilities. Like the clock being reset, the
machine hitting a snag or something...

Davide


--
| A year spent in artificial intelligence is enough to make one believe
| in God.
|
|

[ Post a follow-up to this message ]

Chris Coyle wrote:
>
> We have an odd problem where a cron job
> sometimes doesnt run.
> Its on a daily schedule, and it creates a log file
> so we know exactly when it does or doesnt run.
> Usually it runs fine.
> Just occasionally, maybe a couple times per month,
> it doesnt run, and I dont know why.

Is there any chance it's only once per year when daylight
savings time happens?

> I have an idea (forgive me if this sounds stupid)
> that maybe sometimes the system is so busy running
> other high-priority tasks that it misses the
> scheduled start-time for our job.
> Is that possible, and if so, how would one test
> for that?

Also check that cron is still running.  Every so often
it crashes and needs to be restarted.  Less likely each
year but worth checking.

Make sure the permissions are right, the filesystem
is not full, the script doesn't exit with some error
before it creates the log.

Look in the cron logs.  You don't mention your version
so I don't know where they are.  Maybe /var/cron.  You
may need to create the direct and/or add a switch to
the daemon to get it to log what it runs.

[ Post a follow-up to this message ]