[Labs-l] Getting SIGTERM on grid engine jobs

Anthony Di Franco di.franco at gmail.com
Tue Feb 17 20:28:12 UTC 2015


Thanks Merlijn, increasing the memory limit seems to help for now.
I tried the -ma flag both with and without an argument and jsub and jstart
reject it, and it doesn't appear in the man pages or command line help for
either one. Am I missing something, or are you sure you had the right thing
in mind?
Anthony

On Mon, Feb 16, 2015 at 11:47 PM, Merlijn van Deen <valhallasw at arctus.nl>
wrote:

> Typically: out of memory. The easiest way to check is using
>
> qacct -j "$name"
>
> and checking for maxvmem.
>
> Alternatively, if you pass -ma ('send me an e-mail on abort') to jsub,
> you'll get an e-mail with (somewhat cryptic) information on what happened.
> That information could help to debug the issue.
>
> Merlijn
>
> On 17 February 2015 at 08:18, Anthony Di Franco <di.franco at gmail.com>
> wrote:
>
>> Hi all,
>>  We're getting SIGTERM sent to some of our cocytus jobs which we are
>> running under jsub as follows:
>> jsub -N "$name" -mem 640m -e $logdir -o $logdir -continuous $command
>> Can anyone say what the possible causes of this might be?
>> Thanks
>> Anthony
>>
>> _______________________________________________
>> Labs-l mailing list
>> Labs-l at lists.wikimedia.org
>> https://lists.wikimedia.org/mailman/listinfo/labs-l
>>
>>
>
> _______________________________________________
> Labs-l mailing list
> Labs-l at lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/labs-l
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.wikimedia.org/pipermail/labs-l/attachments/20150217/193b78fb/attachment.html>


More information about the Labs-l mailing list