Re: Debconf10: Batch Queuing Systems BoF -- revisited
Hi,
On Thu, May 10, 2012 at 09:49:11AM +0200, Michael Hanke wrote:
> Unfortunately, I can't attend Debconf this year, but maybe there are
> enough people interested in the topic to have a follow-up BoF to revisit
> the issue.
I won't be there, either.
Note that the Grenoble workshop draft agenda has an item about batch
queueing systems. I might see if I can make it there, if it makes
sense.
> What is the status of similar software in Debian (GridEngine, Slurm,
> ...)?
Mark Hymers recently said that he will not have time to maintain
Gridengine anymore. AIUI, Gridengine has been abondoned as a FLOSS
project by Oracle after the Sun acquisition, but there is a community
fork, http://gridscheduler.sourceforge.net
I will see whether I get/have time to keep maintaining
gridengine/gridscheduler.
I don't know about Slurm, AIUI it is rather mature both upstream and in
Debian, but I never used it (the Leibniz Rechenzentrum is migrating from
Gridengine to Slurm at the moment, though, so I might get to use it soon
:) )
Finally, we now have Torque in Debian,
http://packages.qa.debian.org/t/torque.html
Torque is a PBS/OpenPBS fork, quite mature upstream (and used by lots of
research groups I know), but I am not certain on the Debian status.
So having Condor certainly makes sense, but I think we have several
mature (from a upstream POV) batch queueing systems in Debian now, but
might lack maintenance personell.
Michael
Reply to: