[torqueusers] Torque job accounting: figuring out run duration

Matthew Britt msbritt at umich.edu
Fri Feb 8 07:48:40 MST 2013


Nick - although this is outside of just using the torque logs, this was the principal reason we started using Gold a while ago - to have a good database of usage.   It makes looking at job usage pretty easy. As an example, looking at my jobs that I've started in the last month:

[root at nyx services (master)]# glsusage -u msbritt -s 2013-01-08 --show Instance,User,Project,Processors,Memory,Duration  | head
Instance    User    Project       Processors Memory Duration 
----------- ------- ------------- ---------- ------ -------- 
9374899     msbritt cacstaff_flux 1000       1024   13       
9366325     msbritt cacstaff_flux 4          1024   40       
9366328     msbritt cacstaff_flux 4          1024   40       
9366391     msbritt cacstaff_flux 1          1024   40       
Moab.404[1] msbritt msbritt_fluxm 6          333    279      
9370097[1]  msbritt msbritt_fluxm 6          333    291      
9396027     msbritt msbritt_flux  1          9      320      
9396031     msbritt msbritt_flux  1          768    240      

We've since updated to use MAM (gold's replacement), but you might find it useful.

Thanks,
   - Matt 

--------------------------------------------
Matthew Britt
CAEN HPC Group - College of Engineering
msbritt at umich.edu


On Feb 7, 2013, at 5:44 PM, Nick Lindberg <nlindberg at mkei.org> wrote:

> Hello,
> 
> I have the following enabled in qmgr:
> 
> record_job_info = True
> record_job_script = True
> 
> Per those parameters, I am familiar with the job information being stored in:
> 
> /var/spool/torque/server_priv/Accounting/ 
> 
> The basic format is:
> 
> 01/17/2013 17:34:10;S;38254.hpc01;user=nlindberg group=nlindberg jobname=Job175 queue=batch ctime=1357945580 qtime=1357945580 etime=1357945580 start=1358465650 owner=nlindberg exec_host={compute node names} Resource_List.neednodes=1 Resource_List.nodect=1 Resource_List.nodes=1 Resource_List.procs=232 Resource_List.walltime=240:00:00
> 
> But I haven't been able to figure out how to figure out how long a job ran for.  Does anybody know what ctime, qtime, and etime and start stand for?  I can wager a guess at etime and qtime, but I still don't see anywhere that shows me job time.
> 
> I was hoping to be able to go back and look at individual jobs and how long they ran for use in accounting (more specifically, being able to subtract core-hours for a job that was either run accidentally or as a test)
> 
> Perhaps I'm looking in the wrong spot, or I don't have something turned on correctly.  I also use Moab, but thought it might be easier to query the Torque accounting.
> 
> Thanks,
> 
> Nick Lindberg
> Sr. Consulting Engineer, HPC
> Milwaukee Institute
> 414.727.6413 (W)
> http://www.mkei.org
> 
> 
> 
> 
> 
> 
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers



More information about the torqueusers mailing list