[HTCondor-users] Request was NOT accepted for claim

[Boris Sadkhin <bsadkhin.anl@xxxxxxxxx>]

Hi,

 

So I have multiple condor workers set up with 4 slots per host. On two of the slots on the hosts, I ran into

 

28327:SchedLog:10/25/18 15:45:13 (pid:64) Request was NOT accepted for claim slot1@hostname

28327-SchedLog:10/25/18 15:45:13 (pid:64) Match record (slot1@hostname) deleted

28327:SchedLog:10/25/18 15:45:14 (pid:64) Request was NOT accepted for claim slot3@hostname

28327-SchedLog:10/25/18 15:45:14 (pid:64) Match record (slot3@hostname) deleted

 

Because of this, jobs were not being matched to the slots. The job would attempt to match, and then fail for some reason, and then restart again, in an endless loop. Because of this, good slots on other hosts were not being considered, and the queue very slowly processed the jobs. Once I narrowed down two condor workers that had some slots that had match failures, I restarted the two condor workers and then jobs began to be matched to workers again. I wasnât able to find much about this in the mailing list search, but I was wondering if anyone had heard of this issue before.

 

Thanks