From tskirvin at ks.uiuc.edu  Thu Jul  6 10:33:10 2006
From: tskirvin at ks.uiuc.edu (Tim Skirvin)
Date: Thu, 6 Jul 2006 10:33:10 -0500
Subject: [cluster-l] Testing
Message-ID: <20060706153310.GA8535@ks.uiuc.edu>

        Do we have a working cluster-l mailing list?

                            - Tim Skirvin (tskirvin at ks.uiuc.edu)
-- 
   Theoretical and Computational    http://www.ks.uiuc.edu/~tskirvin/
Biophysics, Beckman Institute, UIUC    Senior Systems Administrator
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 185 bytes
Desc: not available
Url : http://www.ks.uiuc.edu/pipermail/cluster-l/attachments/20060706/67ed1bea/attachment.bin 

From tskirvin at killfile.org  Thu Jul  6 13:18:33 2006
From: tskirvin at killfile.org (Tim Skirvin)
Date: Thu, 06 Jul 2006 13:18:33 -0500
Subject: [cluster-l] Testing
Message-ID: <tskirvin.20060706181835$58c8@cairo.ks.uiuc.edu>

        Does this go back to the mailing list the way it should?

                            - Tim Skirvin (tskirvin at ks.uiuc.edu)
-- 
   Theoretical and Computational    http://www.ks.uiuc.edu/~tskirvin/
Biophysics, Beckman Institute, UIUC    Senior Systems Administrator

From beacham at uiuc.edu  Thu Jul  6 13:27:25 2006
From: beacham at uiuc.edu (Jacob Beacham)
Date: Thu, 6 Jul 2006 18:27:25 +0000 (UTC)
Subject: [cluster-l] another test
Message-ID: <slrneaqlgd.e0b.beacham@tijuana.ks.uiuc.edu>

Sorry, we're still trying to work out the mailing list <-> newsgroup
connection.

Jacob

From beacham at ks.uiuc.edu  Thu Jul  6 13:33:02 2006
From: beacham at ks.uiuc.edu (Jacob Beacham)
Date: Thu, 6 Jul 2006 13:33:02 -0500
Subject: [cluster-l] And now vice versa
Message-ID: <20060706183302.GA20813@ks.uiuc.edu>

Will this mailing be routed to usenet?

Jacob

From erikam at uiuc.edu  Thu Jul  6 13:39:05 2006
From: erikam at uiuc.edu (Erika Noerenberg Montgomery)
Date: Thu, 6 Jul 2006 13:39:05 -0500
Subject: [cluster-l] Testing
In-Reply-To: <tskirvin.20060706181835$58c8@cairo.ks.uiuc.edu>
Message-ID: <A1871D6C61F72E43A523CACE6BB46693FB3485@DSBEXCLUSTER.ad.uiuc.edu>

> -----Original Message-----
> From: cluster-l-bounces at ks.uiuc.edu
[mailto:cluster-l-bounces at ks.uiuc.edu]
> On Behalf Of Tim Skirvin
> Sent: Thursday, July 06, 2006 1:19 PM
> To: cluster-l at ks.uiuc.edu
> Subject: [cluster-l] Testing
> 
>         Does this go back to the mailing list the way it should?
> 

Yep.  This is a great idea - thanks for setting it up!

<snipped sig>


--
Erika Montgomery
Network Administrator
Department of Electrical and Computer Engineering
217.244.8032 + erikam at uiuc.edu 



From erikam at uiuc.edu  Thu Jul  6 13:39:05 2006
From: erikam at uiuc.edu (Erika Noerenberg Montgomery)
Date: Thu, 6 Jul 2006 13:39:05 -0500
Subject: [cluster-l] Testing
In-Reply-To: <tskirvin.20060706181835$58c8@cairo.ks.uiuc.edu>
Message-ID: <mailman.1.1152211147.25022.cluster-l@mailhost.ks.uiuc.edu>

> -----Original Message-----
> From: cluster-l-bounces at ks.uiuc.edu
[mailto:cluster-l-bounces at ks.uiuc.edu]
> On Behalf Of Tim Skirvin
> Sent: Thursday, July 06, 2006 1:19 PM
> To: cluster-l at ks.uiuc.edu
> Subject: [cluster-l] Testing
> 
>         Does this go back to the mailing list the way it should?
> 

Yep.  This is a great idea - thanks for setting it up!

<snipped sig>


--
Erika Montgomery
Network Administrator
Department of Electrical and Computer Engineering
217.244.8032 + erikam at uiuc.edu 



From gnelson at uiuc.edu  Fri Jul  7 10:44:03 2006
From: gnelson at uiuc.edu (Gerald Nelson)
Date: Fri, 7 Jul 2006 10:44:03 -0500
Subject: [cluster-l] cluster setup wiki
Message-ID: <bbdeb457.d42a7e1f.81af500@expms1.cites.uiuc.edu>

I took the cluster workshop last fall and now have 6 pcs in a
cluster that has been running for about 2 weeks. We have setup
a wiki to document the process for us. The purpose of this
post is to see what kind of interest there is in a. seeing
this documentation, b. setting up a more general wiki so
others could also document their successes (and failures!). At
the moment the wiki is available only to UIUC urls. Its
address is agecon192.agecon.uiuc.edu/wiki. 

Comments welcome. Feel free to edit the wiki too!

Regards, Jerry Nelson
Gerald Nelson
Professor, Dept. of Agricultural and Consumer Economics
University of Illinois, Urbana-Champaign
office: 217-333-6465 
cell: 217-390-7888
315 Mumford Hall
1301 W. Gregory
Urbana, IL 61801

From gnelson at uiuc.edu  Fri Jul  7 11:56:30 2006
From: gnelson at uiuc.edu (Gerald Nelson)
Date: Fri, 7 Jul 2006 11:56:30 -0500
Subject: [cluster-l] 'improved' wiki url
Message-ID: <befb08dc.d431203b.841f400@expms1.cites.uiuc.edu>

I gave the wrong url for the wiki I posted earlier. The
correct one is 
http://agec192.agecon.uiuc.edu/wiki/

Sorry about that. 

Jerry
Gerald Nelson
Professor, Dept. of Agricultural and Consumer Economics
University of Illinois, Urbana-Champaign
office: 217-333-6465 
cell: 217-390-7888
315 Mumford Hall
1301 W. Gregory
Urbana, IL 61801

From tskirvin at ks.uiuc.edu  Mon Jul 10 09:19:13 2006
From: tskirvin at ks.uiuc.edu (Tim Skirvin)
Date: Mon, 10 Jul 2006 09:19:13 -0500
Subject: [cluster-l] Welcome to cluster-l!
Message-ID: <20060710141913.GA2669@ks.uiuc.edu>

        Welcome to cluster-l!  Or whatever it is that we want to call this
group; I know that that name isn't very elegant.  

        Some background: I'm the head sysadmin for the Theoretical and
Computational Biophysics Group at the Beckman Institute.  I'm interested
in Linux clusters primarily because a good portion of my job is to run the
things; we've currently got six 48-processor clusters running in our group, 
and are currently putting together a budget to buy several more.  

        Over the last year or so, my co-workers and I have put together
a series of cluster-building workshops.  The point of these workshops
was to share some of the hard-earned knowledge that we had gained from
putting together four generations of Linux clusters, rather than letting
others make the same mistakes that we had once made.  The focus has been
on small- to medium-sized clusters, or anything that a research group
could reasonably fund by writing a grant.  These workshops were quite
successful, or at least that's what our feedback said...

        ...but in the grand scheme of things, this was only a start.
Our advice was helpful towards putting together your first cluster, or
learning more effectively what decisions had gone into the cluster that
you're running on; but there wasn't much on maintenance, or planned
growth, or applications, or any of the myriad of other cluster-related
problems that we face every day, and which are a whole lot more difficult
to run a workshop on.  

        That's where this list comes in.

        What I hope to put together with this mailing list:


        o A community of users that can share their experiences (and
horror stories) regarding their cluster experiences from all angles,
including purchase and design, operations, applications, decommisioning,
etc.  

        o A resource for anybody that's looking to put together a new
cluster to ask questions, including ones that we couldn't include in the 
workshop (like "what specific hardware should I buy?", which varies by
time and application dramatically).

        o Semi-regular (monthly?  Quarterly?  Semesterly?) meetings to
discuss this stuff in person, possibly over... beverages.


        Initially, all of this is directed at the UIUC community; but
there's no reason that it would have to stay so local.  Contributors from
the community are welcome, and ideally, we could spread out to be useful
to the world at large...  Academic clusters are not a UIUC-specific
problem, after all.

        As with many things, this list is going to be what its members
make of it.  If anything useful comes of this list at all, I'll be happy.
So... discuss!  Document!  Suggest interesting problems that you haven't
solved, or interesting problems that you *have* solved!  And, all in all,
welcome aboard...

                            - Tim Skirvin (tskirvin at ks.uiuc.edu)
-- 
   Theoretical and Computational    http://www.ks.uiuc.edu/~tskirvin/
Biophysics, Beckman Institute, UIUC    Senior Systems Administrator
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 185 bytes
Desc: not available
Url : http://www.ks.uiuc.edu/pipermail/cluster-l/attachments/20060710/9ac45aec/attachment.bin 

From tskirvin at ks.uiuc.edu  Fri Jul 21 12:36:18 2006
From: tskirvin at ks.uiuc.edu (Tim Skirvin)
Date: Fri, 21 Jul 2006 12:36:18 -0500
Subject: [cluster-l] Types of clusters
Message-ID: <20060721173618.GA13497@ks.uiuc.edu>

        I ask this partially as a conversation starter, and mostly because
I really am curious:
        
        How many people on this list are running some kind of compute
cluster?  For those running one or more, what are you running - size,
processor type, purpose, cooling method?  

        Our setup is here:

        http://www.ks.uiuc.edu/Development/Computers/Cluster/

                            - Tim Skirvin (tskirvin at ks.uiuc.edu)
-- 
   Theoretical and Computational    http://www.ks.uiuc.edu/~tskirvin/
Biophysics, Beckman Institute, UIUC    Senior Systems Administrator
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 185 bytes
Desc: not available
Url : http://www.ks.uiuc.edu/pipermail/cluster-l/attachments/20060721/24c00c7b/attachment.bin 

From alazarev at itg.uiuc.edu  Fri Jul 21 13:04:18 2006
From: alazarev at itg.uiuc.edu (Alexander Lazarevich)
Date: Fri, 21 Jul 2006 13:04:18 -0500 (CDT)
Subject: [cluster-l] Types of clusters
In-Reply-To: <20060721173618.GA13497@ks.uiuc.edu>
References: <20060721173618.GA13497@ks.uiuc.edu>
Message-ID: <Pine.LNX.4.63.0607211239500.5056@zeus.itg.uiuc.edu>

We are in the process of assembling the following 8 node cluster:

8 x supermicro A+ Server AS1020A-T chassis
16 x AMD Opteron 250 processors
32 x 1GB dimms

We buy hardware from parts, in this case the mobo comes mounted in the 
chassis so we only have to assemble the procs and memory. We will be 
running microCT reconstructions and maya rendering on this cluster. Rack 
is a basic 42U from RackSolutions.com, cooling is 2.5 ton ceiling mounted 
leibert, room is 4271A Beckman which was recently renovated for ITG when 
we were kicked out of B650R. No cool website of our machine room, yet...

Alex
-----                                                            -----
   Alex Lazarevich | Systems Administrator | Imaging Technology Group
     Beckman Institute | University of Illinois | www.itg.uiuc.edu
                405 North Mathews Avenue Urbana IL 61801
            Ph: (217) 244-1565 | Email: alazarev at itg.uiuc.edu
-----                                                            -----

On Fri, 21 Jul 2006, Tim Skirvin wrote:

>        I ask this partially as a conversation starter, and mostly because
> I really am curious:
>
>        How many people on this list are running some kind of compute
> cluster?  For those running one or more, what are you running - size,
> processor type, purpose, cooling method?
>
>        Our setup is here:
>
>        http://www.ks.uiuc.edu/Development/Computers/Cluster/
>
>                            - Tim Skirvin (tskirvin at ks.uiuc.edu)
> --
>   Theoretical and Computational    http://www.ks.uiuc.edu/~tskirvin/
> Biophysics, Beckman Institute, UIUC    Senior Systems Administrator
>