[gridengine users] Sudden DRMAA error
reuti at Staff.Uni-Marburg.DE
Tue Mar 1 15:55:36 UTC 2011
Am 01.03.2011 um 16:50 schrieb Prentice Bisbal:
> Reuti wrote:
>> Am 28.02.2011 um 23:29 schrieb Prentice Bisbal:
>>> Dave Love wrote:
>>>> Reuti <reuti at staff.uni-marburg.de> writes:
>>>>> This doesn't look like a DRMAA error per se - the job requests something which can't be satisfied.
>>>> I.e. DRMAA has a default implicit `-w e' if I recall correctly.
>>> I did some googling before posting here for help, and on eof my hits
>>> mentioend the same thing and recommend adding '-w n' to my job
>>> submission script, to override the default validation behavior of the
>>> DRMAA. Didn't help.
>>> Of course I added that to my submission script, which starts the first
>>> Mathematica job. I really need to find a way to add it to the
>>> Mathematica script so I can debug why the DRMAA calls are failing. If I
>>> could do that, '-w p' or '-w v' should identify the problem right away.
>>> Is there anyway to force '-w p' or '-w v' for all DRMAA jobs?
>> You could attach it to your ~/.sge_request file or the global one. When it's only happening for parallel jobs: the PE is attached to a queue? No RQS in the way?
> Reuti, I try adding '-w p' and '-w v' to my global sge_request (not at
> the same time, mind you). And it wasn't to helpful. It just give output
> like this for the initial job:
> $ qsub factor.sh
> verification: found suitable queue(s)
When you override the ~/.sge_request on the commandline, the job should enter the system.
> But it doesn't launch the job, so it never gets to the point where the
> DRMAA is invoked. I need to do this only for the DRMAA invocations, and
> I think Lao Tsao pointed out where to do this in a separate e-mail.
> users mailing list
> users at gridengine.org
More information about the users