Re: [HTCondor-users] Fwd: Re: DAG error: "BAD EVENT: job (...) executing, total end count != 0 (1)"

[John M Knoeller <johnkn@xxxxxxxxxxx>]

I think the EVENT_LOG knob needs to be set in the HTCondor configuration, not the dag configuration. 
This knob causes the shadow to use a different code path that bypasses the bug, but the *shadow*
needs to see the knob for that to work. 

-tj

-----Original Message-----
From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> On Behalf Of Nicolas Arnaud
Sent: Wednesday, February 20, 2019 5:24 AM
To: Mark Coatsworth <coatsworth@xxxxxxxxxxx>; giuseppe.dibiase <giuseppe.dibiase@xxxxxxxxx>
Cc: Franco Carbognani <franco.carbognani@xxxxxxxxx>; HTCondor-Users <htcondor-users@xxxxxxxxxxx>; EGO Helpdesk <service@xxxxxxxxx>
Subject: Re: [HTCondor-users] Fwd: Re: DAG error: "BAD EVENT: job (...) executing, total end count != 0 (1)"


Hi Mark,

> (...)
> In the meantime, there's a workaround you can use that is safer than the 
> one we talked about previously. If you set a value for your EVENT_LOG 
> configuration knob (which is empty by default), that should force the 
> shadow to run the code that refreshes the job ID. The value should be 
> the path to a file where events can get written. More info here:
> 
> http://research.cs.wisc.edu/htcondor/manual/v8.8/ConfigurationMacros.html#x33-1870003.5
> 
> Please give this a try, and we'll keep you posted on getting this 
> problem fixed.
> (...)

So my dag file starts by

> CONFIG /data/procdata/web/dqr/test_20190220_narnaud/dag/dag.config

and the contents of that file are

> DAGMAN_ALLOW_EVENTS = 5
> EVENT_LOG = /data/procdata/web/dqr/test_20190220_narnaud/dag/event.log

Yet, when I run the dag, there is no particular problem but nothing gets 
written in EVENT_LOG.

In the dagman.dag.out file I see

> (...)
> 02/20/19 12:14:42 Using DAGMan config file: /data/procdata/web/dqr/test_20190220_narnaud/dag/dag.config
> (...)
> 02/20/19 12:14:42 allow_events (DAGMAN_ALLOW_EVENTS) setting: 5
> (...)

but searching for "EVENT_LOG" or "event.log" (the name of the EVENT_LOG 
file) in that file doesn't return anything.

=> What am I missing to get the EVENT_LOG file filled?

=> Could there another setting that would mask the setup of "EVENT_LOG"?

Thanks in advance,

Nicolas
_______________________________________________
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/