I'd also like to re-emphasize something Andreas said earlier: SIGTERM
*usually* means that some external entity is killing your
application.  It *could* be coming from within the application itself,
but that's not too common.

You might want to look into that to find out where the SIGTERM is
coming from.  The Microtar maintainers might have some better ideas.

On May 30, 2008, at 9:17 AM, Andreas Schäfer wrote:

> On 12:28 Fri 30 May     , Lee Amy wrote:
>> 2008/5/29 Andreas Schäfer <gentryx@gmx.de>:
>> Thank you very much. If I do a shorter job it seems run well. And
>> the job
>> dosen't repeatedly fail at the same time, but it will fail at this
>> error
>> messages. Anyway, I'm not using a scheduling system. So any
>> suggestions?
>
> At least no easy ones, sorry. ;-) You could ask the Microtar guys if
> they know anything about that problem. And of course you could use a
> debugger to dig into Microtar and find the problem yourself. ^^ Open
> MPI has some doc how to attach gdb to a parallel job: (and how to use
> valgrind etc.)
>
> http://www.open-mpi.org/faq/?category=debugging
>
> Good luck!
> -Andi
>
>
> --
> ============================================
> Andreas Schäfer
> Cluster and Metacomputing Working Group
> Friedrich-Schiller-Universität Jena, Germany
> PGP/GPG key via keyserver
> I'm a bright... http://www.the-brights.net
> ============================================
>
> (\___/)
> (+'.'+)
> (")_(")
> This is Bunny. Copy and paste Bunny into your
> signature to help him gain world domination!

> _______________________________________________
> users mailing list
> users@open-mpi.org

> http://www.open-mpi.org/mailman/listinfo.cgi/users


--
Jeff Squyres
Cisco Systems

_______________________________________________
users mailing list
users@open-mpi.org

http://www.open-mpi.org/mailman/listinfo.cgi/users