Mailing List Archives

Public Access

 		[Computer Systems Lab]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Understanding job scheduling
Just to be clear: this does _not_ imply that each time a job finishes on a machine there is a scheduling latency of ~20 seconds before another one can start there. In a situation where you have many similar jobs, the schedd will reuse machines that it has claimed. When one job finishes, another job that matches the machine (and which is from the same user) will be scheduled to run there immediately, without having to go back to the negotiator. 

--Dan

Dan Bradley wrote:
The condor negotiator is responsible for matching new machines to job submitters. Negotiation currently happens in cycles, with a minimum spacing of 20 seconds between cycles. From your description, it sounds like 2 jobs were in the queue when the first negotiation cycle happened, and then when the next negotiation cycle happened, all the rest of the jobs were in the queue. 

--Dan

Ian Stokes-Rees wrote:
Hi, I'm trying to understand how Condor job scheduling works. I have a condor pool with 18 "Unclaimed" slots. I submit 30 "sleep 90" jobs. 2 of them start quite promptly, but the next 16 take 20 seconds before they are matched, then all start at once: 

-- Submitter: abitibi.sbgrid.org : <10.0.10.39:52783> : abitibi.sbgrid.org
ID OWNER SUBMITTED RUN_TIME ST PRI SIZE CMD 93.0 ijstokes 3/21 12:42 0+00:00:56 R 0 0.0 sleep 90 94.0 ijstokes 3/21 12:42 0+00:00:56 R 0 0.0 sleep 90 95.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 96.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 97.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 98.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 99.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 100.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 101.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 102.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 103.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 104.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 105.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 106.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 107.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 108.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 109.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 110.0 ijstokes 3/21 12:42 0+00:00:35 R 0 0.0 sleep 90 

Why does this happen?

Cheers,

Ian

--
Ian Stokes-Rees                            W: http://sbgrid.org
ijstokes@xxxxxxxxxxxxxxxxxxx               T: +1 617 418-4168
SBGrid, Harvard Medical School             F: +1 617 432-5600



------------------------------------------------------------------------

_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users
The archives can be found at: https://lists.cs.wisc.edu/archive/condor-users/ 



_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users
The archives can be found at: https://lists.cs.wisc.edu/archive/condor-users/