[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [condor-users] Condor-G: failed to open stdout (but works withglobus-job-run)



Hi,

I'm no expert, but it might still be a filter problem (though you say
you've checked already...):
http://www.nsf-middleware.org/documentation/NMI-R2/0/globustoolkit/admin/guide-faq.html#stdout

Do you have the GLOBUS_TCP_PORT_RANGE set in the server?

You may want to compare the env variables interactively in the server
and when doing:
globus-job-run grid1.lab.ac.uab.edu /usr/bin/env

Sebastian


On Mon, 9 Feb 2004, Zach Garner wrote:

> I have one Globus-enabled resource, running GT3 and GT2 services (from
> Globus 3.0.2) on RedHat 9 with Condor 6.6.0.
>
> When I run a job via globus-job-run, everything works:
>
> [zach@nori condor]$ globus-job-run grid1.lab.ac.uab.edu /bin/date
> Mon Feb  9 11:35:11 CST 2004
>
>
> But when I run it through Condor-G I get:
>
> [zach@nori condor]$ condor_q -l|grep HoldReason
> HoldReason = "Globus error 73: the job manager failed to open stdout"
> HoldReasonCode = 2
> HoldReasonSubCode = 73
>
>
> I've double checked the things that I know to check: globus-hostname,
> DNS, firewall. The file being executed is on a shared filesystem, in a
> common UID space.
>
>
> Does anyone know of any misconfiguration that would prevent Condor-G
> from submitting correctly, but still allow regular Globus commands to
> work?
>
>
> -----
> My submit script for Condor-G is:
>
> | executable = /bin/date
> | Transfer_Executable = false
> | globusscheduler = grid1.lab.ac.uab.edu/jobmanager
> | universe = globus
> | output = test.out
> | log = test.log



Condor Support Information:
http://www.cs.wisc.edu/condor/condor-support/
To Unsubscribe, send mail to majordomo@xxxxxxxxxxx with
unsubscribe condor-users <your_email_address>