Re: [Condor-users] How to run a job in all machines of a CondorPoll, but only once?

[Dave STREET <davey_street@xxxxxxxxxxx>]

could you not make up a class add on all the executate machines
 
So make up the class add attrib
 
Patched =
 
create a job that adds this attritubate to the machine it runs on and sets it to
 
patched = "YES"
 
then in the submit file for the job tell only run the job on machines that do not have PATCHED = "yes"
 
 
then que it more times than clients you have
 
in theroy as I understand it, it would then run once on each machine. (as long as you make sure it runs per machine and not per slots)
 
 
I don't know if this would work jsut my thoughts on it :)
 
Aaron
 
 
 
> Date: Tue, 29 Jun 2010 13:14:59 +0100
> From: john.kewley@xxxxxxxxxx
> To: condor-users@xxxxxxxxxxx
> Subject: Re: [Condor-users] How to run a job in all machines of a CondorPoll, but only once?
>
> > I tend to do this with a PERL script. Parse condor_status, find the
> central
> > manager and construct submit files for each machine (except for the
> manager)
>
> BTW you shouldn't have to have code to exclude the manager, either:
>
> A) your manager is also an execute machine in which case you'll want to
> run the test on there too or
> B) it isn't, in which case using the -execute (or similar - a while
> since I used it) flag to condor_status will only match execute nodes and
> therefore omit the manager and any other nodes in pool which aren't
> execute ones.
>
> JK
> --
> Scanned by iCritical.
> _______________________________________________
> 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/

---

Get a new e-mail account with Hotmail - Free. Sign-up now.