Re: [HTCondor-users] PERIODIC_HOLD is applied extremely infrequently

[Ben Cotton <ben.cotton@xxxxxxxxxxxxxxxxxx>]

Vlad,

On Mon, May 11, 2015 at 12:18 PM, Vladimir Brik
<vladimir.brik@xxxxxxxxxxxxxxxx> wrote:

> For testing, I tried using this:
> SYSTEM_PERIODIC_HOLD = (JobStatus == 2 && JobUniverse == 5 && Owner ==
> "vbrik")
>
> The interesting thing about the expression above is that it puts *some* jobs
> on hold immediately after they start running (as expected), but jobs that
> weren't put on hold immediately after starting are never put on hold.
>
I don't have an answer in mind asking these questions, I'm just trying
to tease out a few details that will hopefully spark a moment of
brilliance in someone's mind.

First, what version of HTCondor are you running?

Is the value of "some" consistent (either in raw terms or as a
percentage) across multiple tests?

I know you said that you checked with `condor_q -const <your
expression>`. Was that with your real expression, your test expression
or both?

Is it possible that some of your test jobs define `periodic_hold` in
the submit file? A job's periodic_hold expression takes precedence
over SYSTEM_PERIODIC_HOLD.


Thanks,
BC

-- 
Ben Cotton
main: 888.292.5320

Cycle Computing
Better Answers. Faster.

http://www.cyclecomputing.com
twitter: @cyclecomputing