The generated condor job description is not placed in $HOME/.globus.
But by modifying the condor perl module you can see what condor job
description is generated.
Add
system("cp $script_filename /tmp/condorSubmitScript.cjd");
after
close(SCRIPT_FILE);
in $GLOBUS_LOCATION/lib/perl/Globus/GRAM/JobManager/condor.pm
and check the generated condor job description in /tmp after
a submission.
(you don't need to restart the container for this)
Martin
> Hi again,
> I'm trying different solutions because "I need to submit a job via GRAM
> now". But I don't know what's going wrong.
> I've read that the script condor.pm generates is placed at $HOME/.globus
> directory. I have some id-job-folder, but they're empty, so I cannot check
> how the script works.
>
> Do you have any idea? I don't know what could be my next step...
>
> Thank you in advance
>
> 2008/2/19, Miguel Julián <[EMAIL PROTECTED]>:
>>
>> Thank you (again) for your time. I've tried to submit a simple job via
>> Condor. And everything seemed OK
>> [EMAIL PROTECTED]:~$ condor_submit /condor_jobs/true.sub
>> Submitting job(s).
>> 1 job(s) submitted to cluster 183846.
>> [EMAIL PROTECTED]:~$ condor_q | grep true
>> 183846.0 username 2/19 09:47 0+00:00:00 I 0 9.8
>> true
>> [EMAIL PROTECTED]:~$
>>
>> I've also tried to submit that work on the following way:
>> [EMAIL PROTECTED]:~$ condor_submit $HOME/condor_jobs/true.sub
>>
>> ERROR: Failed to open command file (No such file or directory)
>>
>> That's the same error I was obtaining when I ask at the email list. Now,
>> I
>> suppose that it happens because computerA-$HOME is different than the
>> computerB-$HOME so condor isn't able to get the description file. But,
>> At
>> which source file does GRAM submits a work via Condor? Maybe I could
>> take a
>> look and find out the problem.
>>
>> Any help is very wellcome :P
>>
>> Thank you for your time!
>>
>> 2008/2/18, [EMAIL PROTECTED] <[EMAIL PROTECTED]>:
>> >
>> > Miguel,
>> >
>> > i see that your RFT configuration seems to be incomplete.
>> > You may want to fix that: see
>> >
>> > http://www.globus.org/toolkit/docs/4.0/admin/docbook/quickstart.html#q-rft-configure
>> > for that.
>> >
>> > (see errors like
>> >
>> > 2008-02-15 00:55:18,956 ERROR service.ReliableFileTransferImpl
>> > [main,<init>:68] Unable to setup database driver with
>> pooling.Aconnection
>> > error has occurred: FATAL: no hay una lÌnea en pg_hba.conf para
>> > "155.210.142.126", usuario "globus", base de datos "rftDatabase", SSL
>> > inactivo
>> >
>> > in the container logfile)
>> >
>> > But this is not the problem with your submission.
>> > Can you try a submission using condor on your GT4 machine?
>> > I mean: if you create a condor job description with /bin/true as
>> > executable and submit it via condor_submit: does that work?
>> > I wonder if you might not have /bin/true as executable on the
>> execution
>> > node.
>> >
>> > Martin
>> >
>> >
>> > > Firstly, thank you for your attention.
>> > > I've attached the output of the container in debug mode and I also
>> > checked
>> > > that condor_submit and condor_rm were pointing to the correct
>> > executables.
>> > >
>> > > Thank you
>> > >
>> > > 2008/2/14, [EMAIL PROTECTED] <[EMAIL PROTECTED]>:
>> > >>
>> > >> I think i never saw this.
>> > >> Can you please start the container in debug mode
>> > >> (set log4j.category.org.globus to DEBUG in
>> > >> $GLOBUS_LOCATION/container-log4j.properties)
>> > >> restart the container and send the container output (zipped)
>> > >>
>> > >> Also: please check that
>> > >> $GLOBUS_LOCATION/lib/perl/Globus/GRAM/JobManager/condor.pm
>> > >> has correct pointers to condor executables.
>> > >> Look for
>> > >>
>> > >> BEGIN
>> > >> {
>> > >> $condor_submit = '/path/to/condor_submit';
>> > >> $condor_rm = '/path/to/condor_rm';
>> > >> }
>> > >>
>> > >>
>> > >> Martin
>> > >>
>> > >>
>> > >>
>> > >> > Hi,
>> > >> > I'm having problems while trying to submit a job to Condor using
>> > the
>> > >> > globusrun-ws command.
>> > >> > My scenario:
>> > >> > -Computer A, with the 4.03 GT installation and Condor Scheduler
>> > >> Adapter
>> > >> > installed.
>> > >> > -Computer B: It has Condor installed
>> > >> > -ComputerA can execute condor_submit,... and those commands (by
>> > ssh)
>> > >> will
>> > >> > be
>> > >> > executed at ComputerB.
>> > >> >
>> > >> > So I've tryed:
>> > >> > ./globusrun-ws -submit -F
>> > >> > http://hostname:port/wsrf/services/ManagedJobFactoryService -Ft
>> > Condor
>> > >> -c
>> > >> > /bin/true
>> > >> > Submitting job...Done.
>> > >> > Job ID: uuid:6f283686-daed-11dc-8685-000476cd404c
>> > >> > Termination time: 02/15/2008 11:10 GMT
>> > >> > Current job state: Failed
>> > >> > Destroying job...Done.
>> > >> > globusrun-ws: Job failed: The executable could not be started.
>> > >> >
>> > >> > ERROR: Failed to open command file (No such file or directory)
>> > >> >
>> > >> > If I try the same instruction without the -Ft option everything
>> > goes
>> > >> fine.
>> > >> >
>> > >> > Any idea?
>> > >> >
>> > >> > Thank you
>> > >> >
>> > >> > ---
>> > >> > Miguel Julián Ramos
>> > >> >
>> > >>
>> > >>
>> > >>
>> > >
>> > >
>> > > --
>> > > Miguel Julián Ramos(Miky)
>> > > http://www.mikysblog.com
>> > > http://www.youtube.com/watch?v=d6gmp8pf8k4
>> > >
>> >
>> >
>> >
>>
>>
>> --
>> Miguel Julián Ramos(Miky)
>> http://www.mikysblog.com
>> http://www.youtube.com/watch?v=d6gmp8pf8k4
>>
>
>
>
> --
> Miguel Julián Ramos(Miky)
> http://www.mikysblog.com
> http://www.youtube.com/watch?v=d6gmp8pf8k4
>
