Mailing List Archives
Authenticated access
|
|
|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] How to run a job in all machines of a CondorPoll, but only once?
- Date: Tue, 29 Jun 2010 13:00:48 +0100
- From: <john.kewley@xxxxxxxxxx>
- Subject: Re: [Condor-users] How to run a job in all machines of a CondorPoll, but only once?
I had a script which took a param that was the CLASSAD to do a foreach over.
So you could do
$ test_each MACHINE
$ test_each NAME
$ test_each OpSys
$ test_each YOUR_CLASSAD
That kind of thing
JK
> -----Original Message-----
> From: condor-users-bounces@xxxxxxxxxxx [mailto:condor-users-bounces@xxxxxxxxxxx]
> On Behalf Of Steven Platt
> Sent: Tuesday, June 29, 2010 10:18 AM
> To: Condor-Users Mail List
> Subject: Re: [Condor-users] How to run a job in all machines of a CondorPoll,
> but only once?
>
> Angel,
>
> 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)
>
> #!/usr/bin/perl
> #FARM OUT PROCESS ONCE TO EACH MACHINE
> chomp($nodelist = `condor_status -format "%s " name -format "%s\t" opsys`);
> @allnodes = split(/\t/,$nodelist);
> foreach (@allnodes) {
> ($machine,$opsys) = split(/ /,$_);
> if ($machine =~ /@/) { ($x,$nod) = split(/@/, $machine) } else { $nod =
> $machine; }
> $nodes{$nod}=$opsys;
> }
> #GET NAME OF CENTRAL MANAGER
> chomp($condor_host = `condor_config_val CONDOR_HOST`);
> foreach $node (keys %nodes) {
> if ($node ne $condor_host) {
> # LOOPS ONCE PER MACHINE **NOT** ONCE PER NODE/SLOT
> # EDIT THE PRINT STATEMENT TO CREATE THE REQUIRED condor_submit FILE
> open(LOCATESUB,">locate_$node.sub")||die "Can't open locatesub : $! ";
> print LOCATESUB "universe=vanilla
> executable = XXXXXXX
> arguments = XXXXXXX
> notification = NEVER
> output = locateis.\$\$(Name).stdout
> error = locateis.\$\$(Name).stderr
> Log = locateis.log
> Requirements = Machine == \"$node\"
> queue
> ";
> close (LOCATESUB) || die "Can't close locatesub : $! ";
> #SUBMIT JOB TO NODE
> system("condor_submit locate_$node.sub 1>/dev/null");
> }
> }
>
> There's probably a better way but this has worked well our small cluster.
>
> Steve
>
> Dr Steven Platt
> Bioinformatics
> Health Protection Agency
> UK
>
>
> -----Original Message-----
> From: condor-users-bounces@xxxxxxxxxxx [mailto:condor-users-bounces@xxxxxxxxxxx]
> On Behalf Of Ángel de Vicente
> Sent: 29 June 2010 09:43
> To: condor-users@xxxxxxxxxxx
> Subject: [Condor-users] How to run a job in all machines of a Condor Poll,but
> only once?
>
> Hi all,
>
> from time to time I find that I would like to run a job in all machines in
> my Condor pool (mostly to find misbehaving ones), but I haven't found an
> easy way of doing it. In the past I've been using the Hawkeye stuff, but
> this is a bit cumbersome, as it implies modifying the configuration files,
> restarting the Startd daemon, etc.
>
> Any ideas on how to easily run a job (only once) on all (as per the
> current pool state) the machines of a Condor pool?
>
> Thanks,
> Ángel de Vicente
> --
> +---------------------------------------------+
> | |
> | http://www.iac.es/galeria/angelv/ |
> | VoIP -> sip:angelv@xxxxxxxxx |
> | |
> | High Performance Computing Support PostDoc |
> | Instituto de Astrofísica de Canarias |
> | |
> +---------------------------------------------+
> _______________________________________________
> 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/
> -----------------------------------------
> **************************************************************************
> The information contained in the EMail and any attachments is
> confidential and intended solely and for the attention and use of
> the named addressee(s). It may not be disclosed to any other person
> without the express authority of the HPA, or the intended
> recipient, or both. If you are not the intended recipient, you must
> not disclose, copy, distribute or retain this message or any part
> of it. This footnote also confirms that this EMail has been swept
> for computer viruses, but please re-sweep any attachments before
> opening or saving. HTTP://www.HPA.org.uk
> **************************************************************************
> _______________________________________________
> 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/
--
Scanned by iCritical.