tags 332285 moreinfo
quit
* Petter Reinholdtsen
> When installing munin-node as part of the debian-edu installation on a
> qemu virtual machine, the munin-node cron job take so long to run (or
> the clock runts too fast) that several cronjobs are started in
> parallel, slowly bringing the machine to a complete halt.
>
> The problem is that the cron job do not check if another cron job is
> running before starting a new task. The first cron job is still
> running when the next starts, slowing down the machine even further
> making sure both are still running when the third cron job starts, and
> so no.
>
> I suggest teaching the cron job to check if the same cron job is still
> running as another process, and either kill the previous one, or not
> start when the script already is running.
The underlying scripts that are run from "munin-cron" are supposed to
check their lock files at
/var/run/munin/munin-{update,graph,html,limits}.lock. This seems to
work here, e.g.:
[EMAIL PROTECTED] :) sudo -u munin /usr/share/munin/munin-graph & sudo -u
munin /usr/share/munin/munin-graph ( ~
[1] 19988
Lock already exists: /var/run/munin/munin-graph.lock. Dying.
[1] + 19988 done sudo -u munin /usr/share/munin/munin-graph
There's also a global timeout of three minutes for the munin-update
job.
Can you try the above command and see if you get the same output? I
must admit I'm not familiar with the virtualisation technology you're
using, but unless something is amiss with the locking implementation it
sounds as if the scripts are getting stuck somewhere before checking
their locks. I think I need you to help with some more debugging in
order to get anywhere with this bug, I'm afraid...
Regards
--
Tore Anderson
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
