Guillermo Marco Puche <[email protected]> writes:

> Hello,
>
> After clearing all.q@compute-0-1 E status.
> I'm getting this error when trying to use 16 slots with MPI:
>
> error: executing task of job 127 failed: execution daemon on host
> "compute-0-1.local" didn't accept task

Assuming it's open-mpi, was that built with gridengine support?
What does the execd log show (preferably at "info" log level)?

> El 13/11/2012 10:01, Guillermo Marco Puche escribió:
>> Hello,
>>
>> Since today and I really don't know why when i request 16 slots in
>> my cluster i'm getting job queued in qw status forever.
>> I've 2 compute nodes each one has 8 cpus (8 slots).
>>
>> When I request 8 sloths with '-pe mpi 8' or '-pe orte 8', it works
>> perfect and allocates task in compute-0-0.

-- 
Community Grid Engine:  http://arc.liv.ac.uk/SGE/

_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users

Reply via email to