Re: [HTCondor-users] All jobs go to a single node (with enough slots)

[Alexander Prokhorov <prokher@xxxxxxxxx>]

Thank you very much!

All the best,

Alexander A. Prokhorov

prokher@xxxxxxxxx

On 2 Dec 2019, at 18:37, Carsten Aulbert <carsten.aulbert@xxxxxxxxxx> wrote:

Hi

On 12/2/19 4:26 PM, Alexander Prokhorov wrote:

In my experiments I see that all jobs go to a single node (which has
enough slots configured), but I would like to distribute these jobs
across the cluster more or less uniformly even if there are enough slots
on a single machine. Can you please give me a hint how to achieve this?

The condor wiki[1] is a treasure trove and I think your question is
answered there already:

https://htcondor-wiki.cs.wisc.edu/index.cgi/wiki?p=HowToFillPoolBreadthFirst

HTH

Carsten

[1] https://htcondor-wiki.cs.wisc.edu/index.cgi/wiki?p=HowToAdminRecipes
--
Dr. Carsten Aulbert, Max Planck Institute for Gravitational Physics,
CallinstraÃe 38, 30167 Hannover, Germany
Phone: +49 511 762 17185