Current status of using grid-control for gridding

[Frank Roscher]

Dear list members,

I'll keep it short and simple: Is there currently ANY known way of running CMSSW jobs on the grid?
This means a setup that allows to create a valid proxy, has a python able to run GC, has working lcg-cp or similar to transfer the sandbox, and has working glite-wms-job-status and friends to monitor the jobs.
I've spent the whole day so far trying different variations and combinations of CMSSW, SLC, grid-envs, GC versions etc, and I don't think I've yet managed to transfer a sandbox - though some of the other stuff worked in various combinations.

Thanks,
Frank

[dha...@cern.ch]

Hey Frank,

can you describe some of the errors you're getting? Do you work at NAF or EKP?

[Frank Roscher]

Hi Dominik,

the errors have been manyfold, I can't really summarize them. I've been working at EKP, but Thomas M says he has a working setup on the NAF, so I'll try and replicate that.

Thank you,
Frank

[Frank Roscher]

Hi again,

okay, one part of my imitation effort is using the stable-fixes branch of GC.
My config files produces this error on that branch (though not on trunk):

GCError: Invalid backend selected! (grid)
GCError: Error while creating instance of type SEStorageManager (grid_control.storage.SEStorageManager)
ValueError: need more than 1 value to unpack

Would be great if anybody could give me a hint what grid-control is trying to tell me here.

Thanks,
Frank

> --
> You received this message because you are subscribed to the Google Groups "grid-control" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to grid-control...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

[Frank Roscher]

Okay, Fred determined the cause of the error. I used a cms:// URL for my output directory. According to the wiki that should work in stable fixes, but apparently it caused this confusing error message instead.

Cheers,
Frank