[Labs-l] Cronjob

[Merlijn van Deen]

Hello Stefan,

On 1 October 2015 at 20:31, Stefan Kühn <kuehn-s at gmx.net> wrote:

> in the last months cronjob work very well. But since maybe one months it
> don't work.
>
> tools.checkpersondata at tools-bastion-01:~$ crontab -l
> 5,15,25,35,45,55 * * * *  /usr/bin/jsub -N cron-tools.checkpersondata-1
> -once -quiet ./run_change.sh
>

>From the error log:

$ tail cron-tools.checkpersondata-1.err
[Thu Oct  1 17:25:04 2015] there is a job named
'cron-tools.checkpersondata-1' already active

i.e. cron is working as expected, but for some reason the job is running
(while it probably should have finished already?)

To get the status of jobs, you can use qstat:

$ qstat
job-ID  prior   name       user         state submit/start at     queue
                     slots ja-task-ID
-----------------------------------------------------------------------------------------------------------------
 501757 0.44782 lighttpd-c tools.checkp r     08/31/2015 20:58:58
webgrid-lighttpd at tools-webgrid     1
1809911 0.52856 cron-tools tools.checkp r     08/14/2015 23:45:07
task at tools-exec-1209.eqiad.wmf     1

which shows the job has been running since 14 august, which is consistent
with your observation it didn't work for at least a month. This is probably
related to technical issues on that day (either the network or the NFS
issues).

The easiest way to solve this is to kill the job, which can be done with

$ qdel 1809911
tools.checkpersondata has registered the job 1809911 for deletion

I just did this, so things should work as expected again!

Best regards,
Merlijn
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.wikimedia.org/pipermail/labs-l/attachments/20151001/c77ca5bc/attachment.html>