Re: [HTCondor-users] Counting job num of holds for a given reason

[Todd Tannenbaum <tannenba@xxxxxxxxxxx>]

On 1/19/2018 2:55 PM, Parag A Mhashilkar wrote:
> Shouldnât the periodic_release reference RequestMemory job classad attr instead of request_memory?
>
> periodic_release = (HoldReasonCode =?= 34) && (request_memory < (2 * initial_memory))

Yep.  My bad, nice catch!

regards,
Todd





> _____________________________________________
> Parag Mhashilkar
>   
> Fermi National Accelerator Laboratory
> WWW: www.fnal.gov
> Phone: 1 (630) 840-6530
> Fax: 1 (630) 840-3109
> _____________________________________________
>
> On 1/19/18, 2:43 PM, "HTCondor-users on behalf of Parag A Mhashilkar" <htcondor-users-bounces@xxxxxxxxxxx on behalf of parag@xxxxxxxx> wrote:
>
>      [This sender failed our fraud detection checks and may not be who they appear to be. Learn about spoofing at http://aka.ms/LearnAboutSpoofing]
>      
>      Thanks Todd. I wasnât thinking about using MemoryUsage but instead trying to work with NumSystemHolds and NumJobStarts. I think your approach will work and let me do the periodic release exactly once.
>      
>      _____________________________________________
>      Parag Mhashilkar
>      
>      Fermi National Accelerator Laboratory
>      WWW: www.fnal.gov
>      Phone: 1 (630) 840-6530
>      Fax: 1 (630) 840-3109
>      _____________________________________________
>      
>      On 1/19/18, 1:00 PM, "HTCondor-users on behalf of Todd Tannenbaum" <htcondor-users-bounces@xxxxxxxxxxx on behalf of tannenba@xxxxxxxxxxx> wrote:
>      
>      
>          > Hi,
>          >
>          > I want to set a periodic release for a job in case the job is held with
>          > code 34 only so I can request more memory next time when it is released.
>          > However, I want the periodic_release to be triggered only once for these
>          > set of jobs. I got most of the logic figured out, but am missing on how
>          > to put the counter in place for number of times the job was held with
>          > holdcode 34. Is there a macro or standard job classad attribute I am
>          > missing?
>          >
>      
>          No easy way to have a counter like you ask above, but I think you could still achieve the policy you want (only increase the memory once).  How about a submit file like the following?  I didn't test it, but I think it will demonstrate the gist of what you want.  The trick it to record the initial memory request in the ad, and then compare the current memory request to the initial request to decide when to stop releasing the job instead of relying on a (non-existent) counter.
>      
>            executable = foo.exe
>            +initial_memory = 256
>            request_memory = ifthenelse(isUndefined(MemoryUsage),initial_memory,2*MemoryUsage)
>            periodic_release = (HoldReasonCode =?= 34) && (request_memory < (2 * initial_memory))
>      
>          regards,
>          Todd
>          _______________________________________________
>          HTCondor-users mailing list
>          To unsubscribe, send a message to htcondor-users-request@xxxxxxxxxxx with a
>          subject: Unsubscribe
>          You can also unsubscribe by visiting
>          https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users
>      
>          The archives can be found at:
>          https://lists.cs.wisc.edu/archive/htcondor-users/
>      
>      
>      
>      _______________________________________________
>      HTCondor-users mailing list
>      To unsubscribe, send a message to htcondor-users-request@xxxxxxxxxxx with a
>      subject: Unsubscribe
>      You can also unsubscribe by visiting
>      https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users
>      
>      The archives can be found at:
>      https://lists.cs.wisc.edu/archive/htcondor-users/
>      
>
>
> _______________________________________________
> HTCondor-users mailing list
> To unsubscribe, send a message to htcondor-users-request@xxxxxxxxxxx with a
> subject: Unsubscribe
> You can also unsubscribe by visiting
> https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users
>
> The archives can be found at:
> https://lists.cs.wisc.edu/archive/htcondor-users/


--
Todd Tannenbaum <tannenba@xxxxxxxxxxx> University of Wisconsin-Madison
Center for High Throughput Computing   Department of Computer Sciences
HTCondor Technical Lead                1210 W. Dayton St. Rm #4257
Phone: (608) 263-7132                  Madison, WI 53706-1685