[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Flocking Grids Together...

Hi Adam,

What?  You don't ask me first?!?  :-)

On Wed, 2004-09-22 at 09:33, asb7@xxxxxxxxxx wrote:
> Hello all!
> Glad to be part of your list.  Here at Lehigh, we have 2 condor grids we would
> like to flock together, but were a little lost as to the actual way to go about
> doing this.  heres what weve got going on...
> public condor consisting of IRIX and Linux boxes appox 65 nodes in all, which is
> 21 Linux nodes and 44 IRIX nodes.
> semipublic condor grid consisting of 20 linux Nodes.
> now i stress that the 2nd is 'semi' public, because while you can ssh into these
> machines, our university uses AFS to authenticate and transfer userfiles, and
> these machines DO NOT have afs running on them.
> so ideally we would like these semipublic nodes to be able to be seen and used
> int he public condor grid, and the public grid usable from the semipublic grid.
> (if you follow)
> Were a little lost on how to actually get this done, so that globally all users
> will always have the ability to submit jobs to all these machines, not a flock
> on a job by job basis.
> is this possible? any help in getting started would be greatly appreciated.

I think what you want to do is "straightforward" flocking.  (I don't
think afs will be an issue)

Look for FLOCK_FROM and FLOCK_TO in the condor_config files of the two
condor_master machines...

(Or you could just have the "semipublic" nodes be nodes in the larger
condor pool...)

I can give you a hand if you want.  Contact me off-list of you want to
get together.


Jeff Linderoth				     O: 610-758-4879
Asst. Professor				     
Industrial and Systems Engineering	     jtl3@xxxxxxxxxx
Lehigh University			www.lehigh.edu/~jtl3