[Bioclusters] Distributed Load Management on the desktop

[Botka, Christopher]

Ahh good point. We found that the same number (roughly) were logging in, and the increase in load from more jobs submitted made interactive jobs even worse. 

-----Original Message-----
From: bioclusters-bounces+christopher.botka=joslin.harvard.edu at bioinformatics.org <bioclusters-bounces+christopher.botka=joslin.harvard.edu at bioinformatics.org>
To: Clustering,  compute farming & distributed computing in life science informatics <bioclusters at bioinformatics.org>
Sent: Mon Apr 10 18:09:38 2006
Subject: Re: [Bioclusters] Distributed Load Management on the desktop

Botka, Christopher wrote:
> You can get binaries for Solaris SPARC, Solaris x64/x86, RHE, Windows
> and HPUX at:
> 
> http://www.sun.com/software/gridware/ (Registration required)
> 
> The win executables are not the same rev of N1 v6 as the CVS tree, but
> work fine (in our hands).
> 
> There's pretty simple to follow installation instructions available at:
> 
> http://gridengine.sunsource.net/documentation.html
> 
> The biggest challenge is setting up windows domain users to submit jobs,
> particularly if your cluster does not authenticate against your domain.
> 
> But as mentioned below, facilitating jobs submission does nothing to
> minimize the load on you master node.

Thanks,  And yes, the load on the master/login node is reduced when the
scheduler runs on another node than the users log in to.  I'm trying to keep
users from logging on to the head node to use it as a day-to-day shell server.

adf

-- 
Andrew Fant    | And when the night is cloudy    | This space to let
Molecular Geek | There is still a light          |----------------------
fant at pobox.com | That shines on me               | Disclaimer:  I don't
Boston, MA     | Shine until tomorrow, Let it be | even speak for myself

_______________________________________________
Bioclusters maillist  -  Bioclusters at bioinformatics.org
https://bioinformatics.org/mailman/listinfo/bioclusters