[Moabusers] scheduling question

Christopher Samuel samuel at unimelb.edu.au
Sun Mar 18 18:57:39 MDT 2012


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 17/03/12 06:56, Denver Smith wrote:

> One problem that we have here is that it's hard to run jobs that
> are asking for multiple nodes because there are lots of smaller
> single node jobs which the scheduler gives the available nodes to
> as they become available, even if they are queued at a later time.
> What is the best way to deal with this?

We prioritise on number of CPUs and XFACTOR, thus..

# Ignore how long a job has been queued for.
QUEUETIMEWEIGHT 0

# Allow users to reduce job priority.
USERPRIOWEIGHT  1

# Use XFACTOR for prioritisation, and include bypass
SERVICEWEIGHT 10
XFACTORWEIGHT 100
BYPASSWEIGHT 1

# Weight on number of CPUs requested to give wider jobs
# higher priority.
RESWEIGHT 1
PROCWEIGHT 10


- -- 
    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 v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk9mhIMACgkQO2KABBYQAh9+5wCfcAo0AEpiNUUiCjGRANvyO7Gi
L1UAn1EK01S3OiaDvkdOZIEYbXnv+Ibl
=PoVD
-----END PGP SIGNATURE-----


More information about the moabusers mailing list