Re: [Condor-users] The submitting PC suffers power failure and running jobs

[Robert Futrick <rfutrick@xxxxxxxxxxxxxxxxxx>]

Rob,

Yes, the scheduling node must be operational for the jobs to keep 
running, b/c the starter and the shadow communicate while the job runs. 
The submit file attribute, +JobLeaseDuration=<#seconds>, tells Condor 
how long to allow the shadow and the starter to be out of contact before 
killing the job. If you set it to something longer (1200-2400 seconds), 
the occasional reboot of your scheduler shouldn't affect running jobs, 
provided the condor_schedd restarts during the lease duration time you 
specified in the submit file.

This issue is a good reason to have a centralized scheduler that isn't 
your workstation. You could run a scheduler node which is always on, and 
use condor_submit -r | -n | -s  to remotely spool/submit your jobs 
depending on your set-up.

I hope that helps.

Regards,
Rob

Rob Stevenson wrote:
> Dear all,
> First, many thanks for previous advice.
>
> I have an issue where it appears that the submitter PC needs to be on
> permanently during the whole duration of the run.
>
> I'm running in Windows in the vanilla universe with jobs that, if
> suspended or preempted, will restart from the beginning. Therefore I
> completely disallow all preempts/suspends which is no problem.
>
> However, I've noticed that when a submitter PC is powered off or crashes
> and has to be restarted, any jobs that have been submitted from this but
> not yet started (I) will not begin until the submitter is back online.
> Also that sometimes, though not always, jobs that are currently active
> (r) will stop and restart from the beginning.
>
> I can imagine that starting jobs that haven't yet started may require
> the submitter pc to be on, but I'm surprised that already running jobs
> occasionally fail. Is this usual?
>
> Best Regards,
> Rob
>
>
>
> **********************************************************************
> HR Wallingford uses Faxes and Emails for confidential and 
> legally privileged business communications.  They do not of
> themselves create legal commitments.  Disclosure to parties 
> other than addressees requires our specific consent.  We are
> not liable for unauthorised disclosures nor reliance upon them.  
> If you have received this message in error please advise us 
> immediately and destroy all copies of it.
>
> HR Wallingford Limited
> Howbery Park, Wallingford, Oxon, OX10 8BA, UK
> Registered in England  No. 02562099 
>
> **********************************************************************
>
> _______________________________________________
> Condor-users mailing list
> To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
> subject: Unsubscribe
> You can also unsubscribe by visiting
> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
>
> The archives can be found at: 
> https://lists.cs.wisc.edu/archive/condor-users/
>   

> --
>
> ===================================
> Rob Futrick
> main: 888.292.5320
>
> Cycle Computing, LLC
> Leader in Condor Grid Solutions
> Enterprise Condor Support and CycleServer Management Tools
>
> http://www.cyclecomputing.com
> http://www.cyclecloud.com
>
>