[torqueusers] Green computing - Power on/off
Joshua Bernstein
jbernstein at penguincomputing.com
Fri Aug 6 12:32:26 MDT 2010
Hello Alap,
This really isn't something you want happening for each and every job.
Sometimes for very short running jobs, you don't need a whole bunch of
nodes booting and then shutting down right away. Instead, this is
probably best implemented as a separate daemon, or even shell script
daemon that monitors the TORQUE status and does power management from
say, IPMI. This is how it was first implemented in Scyld ClusterWare two
years ago.
http://www.penguincomputing.com/software/scyld_clusterware
The reworked version is much more tightly integrated with TORQUE, and
rather then using shell scripting, it speaks directly to TORQUE using
the native C tm interface.
You might be able to come up with something that works and runs from the
prologue, but remember the prologue runs only after the job is actually
started. I have also seen it implemented as part of a health check
script. The later is probably your best bet.
Also, the Moab schedular has 'green' support built in. Moab of course is
a commercial product, but among green support, it has a ton more features.
http://www.clusterresources.com/products/mwm/docs/18.0greencomputing.shtml
-Joshua Bernstein
Penguin Computing
alap pandya wrote:
> *For the power saving purpose we want to call one script / program for
> each job submission.
> **This script should run only on head node and perform job to turn on
> extra nodes if free nodes are not sufficient to run a job.*
> *
> *Another script/program will turn-off idle nodes .
> How to do that ...Will appreciate if you share any idea for implementation.
>
>
>
>
> With Regards,
> Alap Pandya
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers
More information about the torqueusers
mailing list