Re: [Condor-users] "Failed to open as standard output" error

[Jaime Frey <jfrey@xxxxxxxxxxx>]

On Feb 14, 2006, at 1:20 PM, Ilya Narsky wrote:

> On Tue, 14 Feb 2006, Jaime Frey wrote:
>
> > On Feb 10, 2006, at 3:56 PM, Ilya Narsky wrote:
> >
> > > We installed condor-6.7.13.x86_rh_9 on a testbed cluster at
> > > Caltech. Now I am trying to submit a globus job:
> > >
> > > [narsky@citgrid3 OSG]$ globus-job-run
> > > citgrid3.cacr.caltech.edu:2119/jobmanager-condor /bin/date
> > >
> > > The job becomes idle and never finishes. StarterLog.vm1 on the  
> > > worker
> > > node shows this error:
> > ...
> > > 2/10 12:21:05 Failed to open
> > > '/home/narsky/.globus/job/citgrid3.cacr.caltech.edu/
> > > 18520.1139599210/stdout'
> > > as standard output: No such file or directory (errno 2)
> > > 2/10 12:21:05 Failed to open
> > > '/home/narsky/.globus/job/citgrid3.cacr.caltech.edu/
> > > 18520.1139599210/stderr'
> > > as standard error: No such file or directory (errno 2)
> >
> > Is /home/narsky/.globus/job on a shared filesystem?
>
> Yes, nfs.
>
> We reinstalled condor (mostly because we wanted to move it to a  
> different
> location), and now there is another error about UidDomain before  
> the old
> 'Failed to open' error in StarterLog.vm1 for the worker node.
> UID_DOMAIN is set to 'local' in both headnode and global condor_config
> files.
...
> 2/14 11:02:21 Submitting machine is "citgrid3.cacr.caltech.edu"
> 2/14 11:02:21 ERROR: the submitting host claims to be in our UidDomain
> (local), yet its hostname (citgrid3.cacr.caltech.edu) does not match

The UidDomain must match the end of the machine's hostname. See the  
Condor Manual for details: http://www.cs.wisc.edu/condor/manual/ 
v6.7/3_3Configuration.html#10366

> 2/14 11:02:21 Failed to open
> '/home/narsky/.globus/job/citgrid3.cacr.caltech.edu/5231.1139943730/ 
> stdout'
> as standard output: Permission denied (errno 13)
> 2/14 11:02:21 Failed to open
> '/home/narsky/.globus/job/citgrid3.cacr.caltech.edu/5231.1139943730/ 
> stderr'
> as standard error: Permission denied (errno 13)

I see the error has changed. Now it's permission denied. I'll bet the  
starter is accessing the files as user nobody, because of the  
UidDomain error.

+--------------------------------+-----------------------------------+
|           Jaime Frey           | I used to be a heavy gambler.     |
|       jfrey@xxxxxxxxxxx        | But now I just make mental bets.  |
| http://www.cs.wisc.edu/~jfrey/ | That's how I lost my mind.        |
+--------------------------------+-----------------------------------+