Re: [Condor-users] no scheduling with Condor-G 6.7.19 and Globus 4.0.2

[Erik Paulson <epaulson@xxxxxxxxxxx>]

On Mon, Jun 12, 2006 at 12:57:02PM +0200, Jean-Fran?ois SMIGIELSKI wrote:
> I can give some precisions:
> w30176-jaune gathers the indexes of ~rouge and ~noir. Job submitted directly to ~rouge and ~noir executed with no problem.
> Job submitted drectly to ~jaune is held (for unknown reason)
> Job sent with hope that the target will be choosen for me) are "rejected by your job's requirements"
> 
> Globus nodes don't appear in "condor_status".
> 
> --- my submission description file ---------------------------------
> executable = /bin/hostname
> transfer_executable = false
> output = test.$(cluster).$(process).out
> log = test.$(cluster).$(process).log
> error = test.$(cluster).$(process).err
> universe = grid
> 
> grid_resource = gt4 https://w30176-noir:8443 Fork
> queue 1
> grid_resource = gt4 https://w30176-rouge:8443 Fork
> queue 1
> grid_resource = gt4 https://w30176-jaune:8443 Fork
> queue 1
> 
> #requirements = (TARGET.resource_name =!= UNDEFINED) && (TARGET.JobUniverse == 9)
> requirements = (TARGET.JobUniverse == 9)


This is wrong. You're telling the matchmaker "Run me on a resource that
has a JobUniverse attribute set to 9". Your ad file sample below
does not have a JobUniverse attribute, so it's not matching with anything.

Think about putting something like 'Requirements = MyGridSite==TRUE' in your
job submit file, and then in your ad file below put
MyGridSite = TRUE

(note the difference in equal signs - ie the submit file uses == and the
ad file uses = )

The 'my globus nodes don't show up in condor_status' is worrysome. Do you
mean the ads you're trying to use condor_advertise on don't show up in
condor_status? If so, please check your COLLECTOR_LOG on your central
manager.

-Erik