tech-cluster archive

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

Re: bproc?



Hello all...

While BProc does look interesting, and I have used it under Linux in the
past, I think you will find that the majority of beowulf-styled clusters
end up controlling computational job submission with a queuing system such
as condor, PBS (either OpenPBS or ScalablePBS and possibly combined with
the Maui scheduler) or Sun Grid Engine.  By using BProc you still run into
problems such as users over-scheduling compute nodes which effects
everyones performance.

C3 isn't a bad set of distributed shell tools but there are many others
out there too.  There are various packages refereed to as 'psh' along with
'dsh', 'pdsh' all leveraging rsh hosts.equiv or ssh rsa keys.

-Mike

--
hanulec%hanulec.com@localhost                           cell: 516.410.4478
https://secure.hanulec.com            EFnet irc && aol im: hanulec

On Tue, 9 Dec 2003, Jan Schaumann wrote:

> Hi all,
>
> Is there anybody here who has considered porting bproc[1] to NetBSD or --
> given that bproc is GPL -- implementing a similar concept from scratch?
> What, if anything, are other people currently using to manage processes
> across cluster installations?
>
> At the moment, I don't have any tools for this on my cluster, so people
> need to actually rsh into each node if they want to see what processes
> are running there, for example.  Obviously, that's tedious.
>
> I've started packaging C3[2], but have not used it yet.  What do you
> guys use?
>
> -Jan
>
> [1] http://bproc.sourceforge.net
> [2] http://www.csm.ornl.gov/torc/C3/
>
> --
> I seem to be having this tremendous difficulty with my lifestyle.
>



Home | Main Index | Thread Index | Old Index