[condor-users] globus -> condor matching problem

[Mark Calleja <mcal00@xxxxxxxxxxxxx>]

Hi All,

I'm trying to submit a condor job using Condor-G to a globus gatekeeper
which is part of a condor pool, using the following submit file:

executable = hello.sh
globusscheduler = silica.esc.cam.ac.uk/jobmanager-condor-INTEL-LINUX
universe = globus
output = test.out
log = test.log
queue

The Condor-G -> globus part seems to work, in that a job does indeed get
put into the queue at the remote pool. However, there it sits in an idle
state, and a call to condor_q -ana gives the following reason:

WARNING: Be advised:
     No resources matched request's constraints
     Check the Requirements expression below:

Requirements = (OpSys == "LINUX" && Arch == "INTEL") && (Disk >=
DiskUsage) && ((Memory * 1024) >= ImageSize) && (FileSystemDomain ==
"silica.esc.cam.ac.uk")

Indeed, a quick look at the NegotiatorLog reveals:

10/29 12:33:28   Negotiating with mcal00@xxxxxxxxxxxxxxxxxxxx at <131
.111.41.187:2073>
10/29 12:33:28     Request 00239.00000:
10/29 12:33:28       Rejected 239.0 mcal00@xxxxxxxxxxxxxxxxxxxx <131.
111.41.187:2073>: no match found

Now this is untrue, as I can submit the same job directly from
silica.esc.cam.ac.uk (using a vanilla universe) and it runs on one of
three free matching platforms. 

The question is, am I doing something obviously wrong with my submit
file?

The pool policy has RANK set to 0 for all jobs, with no preferred users
or jobs.

Thanks for any help,

Mark

ps: Some further info: I can submit a similar Condor-G job to a
different jobmanager,  jobmanager-condor-SGI-IRIX, which runs
successfully on the gatekeeper itself (it's part of the condor pool).
The job's a simple "Hello world" sh script, so there are no architecture
conflicts.


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>