Re: [HTCondor-users] Core usage of cluster

[Todd Tannenbaum <tannenba@xxxxxxxxxxx>]

On 8/24/2018 6:15 AM, Sean Crosby wrote:
> Hi all,
>
> I have come from a Torque/Maui background, where it is easy to show how 
> many cores in the cluster are currently being used
>
>     [root@agtorque ~]# showq | grep "Active Jobs" -A1
>       Â108 Active Jobs  Â423 of 452 Processors Active (93.58%)
>                  21 of Â21 Nodes Active   (100.00%)
>
>
> Is there an easy way to get this info in Condor?
>
> I can get slots easily, but that doesn't show how many cores are being used

I'd suggest first passing the "-claimed" flag to condor_status, which 
constrains the output only to slots that are claimed (i.e. slots in use).

Next, you could use "-af" (for 'autoformat') flag to tell condor_status 
to only display the Cpus attribute from each claimed slot (see 'man 
condor_status') like so:

   condor_status -claimed -af cpus

This will output cpu cores in use on each slot, one line per slot.  So 
next you probably want to sum up all those numbers.

My preferred tool for this (and more!) is the awesome 'st' tool ... see
  https://github.com/nferraz/st

But if you don't have st installed, you could use awk (at least awk is 
always installed...) like so:

  condor_status -claimed -af cpus | awk '{n+=$1} END {print n}'

Hope the above helps.

p.s. you may find this talk on monitoring your pool of interest...

http://research.cs.wisc.edu/htcondor/HTCondorWeek2017/presentations/ThuTannenbaum_Monitoring.pdf

regards,
Todd