Re: [Condor-users] condor_rm failing for one user because of credentialproblem

["Matt Hope" <matthew.hope@xxxxxxxxx>]

On 10/30/06, Buhl, Marshall <Marshall_Buhl@xxxxxxxx> wrote:
> I can't answer your question, but it raises another issue we have.  We
> had a student visit us a couple of years ago and one of his jobs is
> still in the queue.  We cannot figure out how to get rid of it.  How can
> it be done on a Windows pool?

Find the offending schedd where the job is located.
and ask it:

condor_config_val -n <scheddhostname> HOSTALLOW_OWNER
condor_config_val -n <scheddhostname> QUEUE_SUPER_USERS

This should tell you which machine/user combinations are allowed owner access.

If you can achieve both then your in luck and can just condor_rm from
the machine/user combination.

If you can't (or the config_val doesn't work) then get on the box
where the schedd is.
Stop the condor service.
Locate the job_queue.log on that machine and delete it - this wipes
*all* jobs from that schedd.
If you can do that then you can likely get on the box as system which
should have privilege to remove the job anyway (though it might
require the -forcex option on condor_rm)

Matt