Re: [HTCondor-users] ShadowLog not accessible for Shadows, jobs not starting/not ending up in the right slot for their Shadow

[Jaime Frey <jfrey@xxxxxxxxxxx>]

> On Feb 5, 2020, at 3:55 AM, Thomas Hartmann <thomas.hartmann@xxxxxxx> wrote:
> 
> Hi Michael,
> 
> thanks for the hints.
> 
> Unfortunately, the log dir and logs all belong to condor:condor.
> 
> But interestingly, the message still pops up endlessly, when moving the
> ShadowLog to the ram disk [1]?!
> 
> Cheers,
>  Thomas
> 
> 
> [1]
> root@grid-arcce1: [~] tail -F /dev/shm/ShadowLog
> 02/05/20 10:25:10 (20489576.0) (3710922): File transfer completed
> successfully.
> 02/05/20 10:25:10 (20489576.0) (3710922): WriteUserLog checking for
> event log rotation, but no lock
> 02/05/20 10:25:10 (20489576.0) (3710922): WriteUserLog checking for
> event log rotation, but no lock

These messages are harmless and should be ignored. They occur when Condor is configured to have a global event log (set via EVENT_LOG) and to not do file locking when writing new events to that log (set via EVENT_LOG_LOCKING). Not locking has been the default since Condor 8.3.6, when we made appending to the event log an atomic write operation.

We should remove these messages in a future release, since they donât indicate a problem.

 - Jaime