[Beowulf] Re: WW Fedora Student questions help

Robert G. Brown rgb at phy.duke.edu
Wed Feb 23 12:15:13 PST 2005


On Wed, 23 Feb 2005 streich at uwm.edu wrote:

> *SNIP*
> > but part of this class is also setting up monitoring and benchmarking tools
> > and tests to get overall proformance and what not of my cluster...
> 
> Check out MRTG and Ganglia on the net for monitoring.  I know that MRTG is
> avalible as a RedHat package that can be installed from distro's install CDs,
> but it is probably best to get it in source code and configure it to meet your
> needs.  Ganglia is powerful, it is used by a lot of clusters running ROCKs.

Or you can try wulfware (xmlsysd and e.g. wulfstat or wulflogger).
Depends on what you want to do with the data.  wulflogger makes it very
easy to log a variety of cluster load metrics into a file at a selected
interval, in case you want to actually run programs to analyze it or
plot it with standalone tools.

It should be prebuilt for FC2, RH9 and Centos 3.3 here:

  http://www.phy.duke.edu/~rgb/Beowulf/wulfware.php

(or available in source rpm or tarball if you prefer).  I've tried to
make this a yummified repository, so you can if you wish autoupdate from
it via yum, or of course you can just grab source or binary rpms and put
them into a local repository.

    rgb

> 
> > And setting up and running some parrallel Applications ... Im very interested
> in
> > getting more into clusters so was wondering if anyone has any tools or scripts
> > or anything i can setup and test on my fedora warewulf setup just to get
> > experience Also i really need some computational or Sim type app's or any type
> > of cluster parrallel application i can play with to get the hang of them so i
> > can move on to makeing my own applications this would be a HUGE help
> 
> Hot topics that are using clusters include astronomical models, biology (human
> genome stuff), and all sorts of things.  Our cluster is dedicated to
> atmospheric research (in particular studying fluffy white clouds most
> atmospheric researchers aren't as intrested in as storms and such) running
> COAMPs and wrf (MPI based programs).
> 
> I suppose the real question is: what type of problem are you intrested in?
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
> 

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu





More information about the Beowulf mailing list