[HTCondor-users] PERIODIC_HOLD is applied extremely infrequently

[Vladimir Brik <vladimir.brik@xxxxxxxxxxxxxxxx>]

Hello,

IceCube's OSG submitters have a problem where their SYSTEM_PERIODIC_HOLD 
expression is applied extremely rarely (days+) and only to a subset of 
matching jobs. It may be that matching jobs are actually placed on hold 
only right after condor is restarted (and even then not always).

Running 'condor_q -con "$(condor_config_val system_periodic_hold)"' 
displays the right jobs, so I don't think I have a typo in the 
expression. I played with various PERIODIC_EXPR timing settings but 
couldn't fix the problem.

The only unusual thing about the affected servers I can thing of is that 
they are OSG submitters, so all their jobs flock to other pools, and 
most run on glideins.


Does anybody know what could be going wrong or how to debug this?

Thanks,

Vlad