Re: large number of short jobs?

Andreas Schäfer <[email protected]>
Newsgroups gmane.linux.cluster.openmosix.general
Message-ID <20060615111321.GA20002@wintermute>
On 19:03 Wed 14 Jun     , Yanming Di wrote:
> Now I have a cluster of 20 computers. I organize my jobs using a bash script. If I simply submit all the jobs from one computer, that computer become very heavily loaded. Is there a easy solution for situations like this.

Well, if you get those 10000 jobs to run as one sequential process
(rather than creating 10000 small and short-lived processes) then oM
should distribute them among your 20 nodes pretty evenly. 

If not I'd suggest you to use a DRM system (Distributed Resource
Management, e.g. Sun Grid Engine). Such a system can start the
short-lived jobs directly on the target machine. 

That ain't as beautiful as the transparent oM approach, but oM has a
problem with short-lived processes: a process can only be migrated
after it's creation. If it doesn't run at all it's migration will be
more expensive than the local execution.

Regards
-Andreas

_______________________________________________
openMosix-general mailing list
openMosix-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
https://lists.sourceforge.net/lists/listinfo/openmosix-general
signature.asc (application/pgp-signature, 189 B)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)

iD8DBQFEkUDRxQzSAzXqsyYRAlNqAJ92Icw3gCvvWBXztB+sFrEX4CB0GQCdFRKO
mu6TT7DL4/M7oJ1jwXuq8tM=
=92B3
-----END PGP SIGNATURE-----
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.