Dear Experts,
Could someone recommend some "quick and dirty" HOWTO on condor based cluster?
A did a bunch of cluster setups, I always used PBS (or torque lately). But my professor wants his to be Condor. I started reading Condor documentation, and it is vast. And quick route, like: architecture chart, and parameters I need to configure on master/submit, and compute nodes - somehow manages to escape me. I'm at a real loss now. All works well on a single machine (for quite some time already), but I am clueless about how to add compute nodes. And my web search abilities seem to have left me just on this one: all I'm able to find (together with a heap of how to submit jobs to existing thing) are a couple of "./install" based setups - whereas mine is rpm based - from their yum repository (for which they say: DO NOT USE condor_configure script, but go ahead and edit config files ... - as opposed to "./install")
Any pointers anybody?
Thanks a lot in advance!
Valeri
++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++
On Thu, May 14, 2015 at 06:16:41PM -0500, Valeri Galtsev wrote:
Could someone recommend some "quick and dirty" HOWTO on condor based cluster?
I've run a condor cluster in the past, but I inherited the configuration and only understood probably 10% of it. It is an incredibly confusing and complicated system.
Considering that Condor is the G part of Red Hat Enterprise MRG, you'd think that they'd build a nice interface for configuring Condor, but I've never seen anything.