[torqueusers] process using more CPUs than requested

Doug Johnson djohnson at osc.edu
Thu Mar 3 06:45:57 MST 2011


At Thu, 3 Mar 2011 00:33:46 +0000,
Christopher Samuel wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 19/02/11 11:05, Gus Correa wrote:
> 
> > I thought the Torque pam module (pam_pbssimpleauth.so)
> > would prevent users to ssh to the nodes directly.
> 
> That's usually done to allow a user who has a job on the
> node to ssh in (either to check on it, or for broken MPI
> launchers which don't support the Torque TM interface -
> yes Intel MPI, I'm looking at you..).
> 

A slightly off-topic clarification.  Intel MPI uses pmi for process
startup and is supported by OSC mpiexec.  Just use '-comm pmi' on the
command line, or set the env variable MPIEXEC_COMM to pmi.

Doug

> What I would really like to see is a re-architecture of
> the cpuset support which would move to this model:
> 
> /dev/cpuset/torque/$user/$jobID
> 
> Then (SSH security model permitting) it would be nice
> if pam_pbssimpleauth could move such logins into the
> /dev/cpuset/torque/$user cpuset so at least they would
> be contained in the superset of all that users jobs.
> 
> Hmm, actually pbs_mom could even spot user tasks not
> in a cpuset and move them into them.. ;-)
> 
> cheers!
> Chris
> - -- 
>     Christopher Samuel - Senior Systems Administrator
>  VLSCI - Victorian Life Sciences Computation Initiative
>  Email: samuel at unimelb.edu.au Phone: +61 (0)3 903 55545
>          http://www.vlsci.unimelb.edu.au/
> 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
> 
> iEYEARECAAYFAk1u4eoACgkQO2KABBYQAh/fvwCfXJaLdjUqI/EFdh6mUUDejBZH
> EJwAnj5ldlOxngI4ciIA0trZCnpM8KqY
> =Wvr4
> -----END PGP SIGNATURE-----
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers


More information about the torqueusers mailing list