<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Here is a paper from TG07, that compares GRAM2 with GRAM4. The
conclusion of the paper are (copied and pasted from the paper at
<a class="moz-txt-link-freetext" href="http://www.globus.org/alliance/publications/papers/TG07-GRAM-comparison.pdf">http://www.globus.org/alliance/publications/papers/TG07-GRAM-comparison.pdf</a>):<br>
<ul>
<li>GRAM4 provides vastly better functionality than GRAM2, in
numerous respects.</li>
<li>GRAM4 provides better scalability than GRAM2, in terms of the
number of concurrent jobs that can be sup-port. It also greatly reduces
load on service nodes, and permits management of that load.</li>
<li>GRAM4 performance is roughly comparable to that of GRAM2. (We
still need to improve sequential submission and file staging
performance, and we have plans for doing that, and also for other
performance optimizations.)</li>
</ul>
You can draw your own conclusions once you read the paper. I also bet
Stu has more numbers than were reported in this paper. From what I
heard, GRAM2 will be optional in GT4.2, and will be phased out
completely in GT4.4, so the upgrade to GRAM4 is inevitable.<br>
<br>
Ioan<br>
<br>
<br>
Mihael Hategan wrote:
<blockquote cite="mid:1201664295.32688.40.camel@blabla.mcs.anl.gov"
type="cite">
<pre wrap="">I'm becoming confused now. Last time I spoke to Yong about WS-GRAM, it
was less scalable and slower (although that varied) than gt2 gram.
So unless I see some numbers, I personally won't believe either of the
statements.
On Tue, 2008-01-29 at 21:25 -0600, Ioan Raicu wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Yong and I ran most of our tests (from Swift) using WS-GRAM (aka GRAM4)
on UC/ANL TG, and I use Falkon on the same cluster using only WS-GRAM.
If I am not mistaken, all TG sites support WS-GRAM.
Ioan
Michael Wilde wrote:
</pre>
<blockquote type="cite">
<pre wrap="">MikeK, this may be obvious but just in case:
On 1/29/08 8:47 PM, Mihael Hategan wrote:
</pre>
<blockquote type="cite">
<pre wrap="">That and/or try using ws-gram:
<jobmanager universe="vanilla" url="tg-grid1.uc.teragrid.org" major="4"
minor="0" patch="0"/>
</pre>
</blockquote>
<pre wrap="">(this goes in the sites.xml file)
Q for the group: is ws-gram supported on uc.teragrid?
</pre>
<blockquote type="cite">
<pre wrap="">
On Tue, 2008-01-29 at 20:42 -0600, Mihael Hategan wrote:
</pre>
<blockquote type="cite">
<pre wrap="">You may want to try to lower throttle.score.job.factor from 4 to 1.
That
will cap the number of jobs at ~100 instead of ~400.
Mihael
</pre>
</blockquote>
</blockquote>
<pre wrap="">for info on setting Swift properties, see "Swift Engine Configuration"
in the users guide at:
<a class="moz-txt-link-freetext" href="http://www.ci.uchicago.edu/swift/guides/userguide.php#properties">http://www.ci.uchicago.edu/swift/guides/userguide.php#properties</a>
- MikeW
</pre>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">On Tue, 2008-01-29 at 18:31 -0800, Mike Kubal wrote:
</pre>
<blockquote type="cite">
<pre wrap="">sorry, long day : )
--- Mihael Hategan <a class="moz-txt-link-rfc2396E" href="mailto:hategan@mcs.anl.gov"><hategan@mcs.anl.gov></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">On Tue, 2008-01-29 at 20:02 -0600, Michael Wilde
wrote:
</pre>
<blockquote type="cite">
<pre wrap="">MikeK, no attachment.
Ive narrowed the cc list, and need to read back
</pre>
</blockquote>
<pre wrap="">through the email thread
</pre>
<blockquote type="cite">
<pre wrap=""> on this to see what Mihael observed.
</pre>
</blockquote>
<pre wrap="">Let me summarize: too many gt2 gram jobs running
concurrently = too many
job manager processes = high load on gram node. Not
a new issue.
</pre>
<blockquote type="cite">
<pre wrap="">- MikeW
On 1/29/08 8:00 PM, Mike Kubal wrote:
</pre>
<blockquote type="cite">
<pre wrap="">The attachment contains the swift script, tc
</pre>
</blockquote>
</blockquote>
<pre wrap="">file,
</pre>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">sites file and swift.properties file.
I didn't provide any additional command line
arguments.
MikeK
--- Michael Wilde <a class="moz-txt-link-rfc2396E" href="mailto:wilde@mcs.anl.gov"><wilde@mcs.anl.gov></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">[ was Re: Swift jobs on UC/ANL TG ]
Hi. Im at OHare and will be flying soon.
Ben or Mihael, if you are online, can you
investigate?
Yes, there are significant throttles turned on
</pre>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">by
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">default, and the system opens those very gradually.
MikeK, can you post to the swift-devel list
</pre>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">your
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">swift.properties file, command line options, and your swift source
</pre>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">code?
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">Thanks,
MikeW
On 1/29/08 8:11 AM, Ti Leggett wrote:
</pre>
<blockquote type="cite">
<pre wrap="">The default walltime is 15 minutes. Are you
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">doing
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">fork jobs or pbs jobs?
</pre>
<blockquote type="cite">
<pre wrap="">You shouldn't be doing fork jobs at all. Mike
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">W, I
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">thought there were
</pre>
<blockquote type="cite">
<pre wrap="">throttles in place in Swift to prevent this
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">type
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">of overrun? Mike K,
</pre>
<blockquote type="cite">
<pre wrap="">I'll need you to either stop these types of
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">jobs
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">until Mike W can verify
</pre>
<blockquote type="cite">
<pre wrap="">throttling or only submit a few 10s of jobs at
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">a
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">time.
</pre>
<blockquote type="cite">
<pre wrap="">On Jan 28, 2008, at 01/28/08 07:13 PM, Mike
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">Kubal
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">wrote:
</pre>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">Yes, I'm submitting molecular dynamics
</pre>
</blockquote>
</blockquote>
<pre wrap="">simulations
</pre>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">using Swift.
Is there a default wall-time limit for jobs
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">on
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">tg-uc?
</pre>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">--- joseph insley <a class="moz-txt-link-rfc2396E" href="mailto:insley@mcs.anl.gov"><insley@mcs.anl.gov></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Actually, these numbers are now
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">escalating...
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">top - 17:18:54 up 2:29, 1 user, load
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">average:
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">149.02, 123.63, 91.94
Tasks: 469 total, 4 running, 465 sleeping,
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap=""> 0
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">stopped, 0 zombie
insley@tg-grid1:~> ps -ef | grep kubal | wc
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">-l
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap=""> 479
insley@tg-viz-login1:~> time globusrun -a -r
tg-grid.uc.teragrid.org
GRAM Authentication test successful
real 0m26.134s
user 0m0.090s
sys 0m0.010s
On Jan 28, 2008, at 5:15 PM, joseph insley
</pre>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">wrote:
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">Earlier today tg-grid.uc.teragrid.org (the
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">UC/ANL
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">TG GRAM host)
</pre>
<blockquote type="cite">
<pre wrap="">became unresponsive and had to be rebooted.
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap=""> I
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">am
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">now seeing slow
</pre>
<blockquote type="cite">
<pre wrap="">response times from the Gatekeeper there
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">again.
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">Authenticating to
</pre>
<blockquote type="cite">
<pre wrap="">the gatekeeper should only take a second or
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">two,
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">but it is
</pre>
<blockquote type="cite">
<pre wrap="">periodically taking up to 16 seconds:
insley@tg-viz-login1:~> time globusrun -a
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">-r
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">tg-grid.uc.teragrid.org
</pre>
<blockquote type="cite">
<pre wrap="">GRAM Authentication test successful
real 0m16.096s
user 0m0.060s
sys 0m0.020s
looking at the load on tg-grid, it is
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">rather
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">high:
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">top - 16:55:26 up 2:06, 1 user, load
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">average:
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">89.59, 78.69, 62.92
</pre>
<blockquote type="cite">
<pre wrap="">Tasks: 398 total, 20 running, 378
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">sleeping,
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">0
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">stopped, 0 zombie
</pre>
<blockquote type="cite">
<pre wrap="">And there appear to be a large number of
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">processes
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">owned by kubal:
</pre>
<blockquote type="cite">
<pre wrap="">insley@tg-grid1:~> ps -ef | grep kubal | wc
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">-l
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap=""> 380
I assume that Mike is using swift to do the
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">job
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">submission. Is
</pre>
<blockquote type="cite">
<pre wrap="">there some throttling of the rate at which
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">jobs
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">are submitted to
</pre>
<blockquote type="cite">
<pre wrap="">the gatekeeper that could be done that
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">would
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">lighten this load
</pre>
<blockquote type="cite">
<pre wrap="">some? (Or has that already been done since
</pre>
</blockquote>
<pre wrap="">earlier today?) The
</pre>
<blockquote type="cite">
<pre wrap="">current response times are not
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">unacceptable,
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">but
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">I'm hoping to
</pre>
<blockquote type="cite">
<pre wrap="">avoid having the machine grind to a halt as
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">it
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">did
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">earlier today.
</pre>
<blockquote type="cite">
<pre wrap="">Thanks,
joe.
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">===================================================
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">joseph a.
insley
<a class="moz-txt-link-abbreviated" href="mailto:insley@mcs.anl.gov">insley@mcs.anl.gov</a>
mathematics & computer science division
</pre>
</blockquote>
<pre wrap="">(630) 252-5649
</pre>
<blockquote type="cite">
<pre wrap="">argonne national laboratory
</pre>
</blockquote>
<pre wrap=""> (630)
</pre>
<blockquote type="cite">
<pre wrap="">252-5986 (fax)
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">===================================================
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">joseph a. insley
<a class="moz-txt-link-abbreviated" href="mailto:insley@mcs.anl.gov">insley@mcs.anl.gov</a>
mathematics & computer science division
</pre>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">(630)
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">252-5649
argonne national laboratory
(630)
252-5986 (fax)
</pre>
</blockquote>
<pre wrap="">
</pre>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">=== message truncated ===
____________________________________________________________________________________
Looking for last minute shopping deals? Find them fast with Yahoo!
Search.
<a class="moz-txt-link-freetext" href="http://tools.search.yahoo.com/newsearch/category.php?category=shopping">http://tools.search.yahoo.com/newsearch/category.php?category=shopping</a>
</pre>
</blockquote>
<pre wrap="">_______________________________________________
Swift-devel mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Swift-devel@ci.uchicago.edu">Swift-devel@ci.uchicago.edu</a>
<a class="moz-txt-link-freetext" href="http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel">http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel</a>
</pre>
</blockquote>
<pre wrap="">_______________________________________________
Swift-devel mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Swift-devel@ci.uchicago.edu">Swift-devel@ci.uchicago.edu</a>
<a class="moz-txt-link-freetext" href="http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel">http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel</a>
</pre>
</blockquote>
<pre wrap="">_______________________________________________
Swift-devel mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Swift-devel@ci.uchicago.edu">Swift-devel@ci.uchicago.edu</a>
<a class="moz-txt-link-freetext" href="http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel">http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel</a>
</pre>
</blockquote>
</blockquote>
<pre wrap=""><!---->
</pre>
</blockquote>
<br>
<pre class="moz-signature" cols="72">--
==================================================
Ioan Raicu
Ph.D. Candidate
==================================================
Distributed Systems Laboratory
Computer Science Department
University of Chicago
1100 E. 58th Street, Ryerson Hall
Chicago, IL 60637
==================================================
Email: <a class="moz-txt-link-abbreviated" href="mailto:iraicu@cs.uchicago.edu">iraicu@cs.uchicago.edu</a>
Web: <a class="moz-txt-link-freetext" href="http://www.cs.uchicago.edu/~iraicu">http://www.cs.uchicago.edu/~iraicu</a>
<a class="moz-txt-link-freetext" href="http://dev.globus.org/wiki/Incubator/Falkon">http://dev.globus.org/wiki/Incubator/Falkon</a>
<a class="moz-txt-link-freetext" href="http://www.ci.uchicago.edu/wiki/bin/view/VDS/DslCS">http://www.ci.uchicago.edu/wiki/bin/view/VDS/DslCS</a>
==================================================
==================================================
</pre>
</body>
</html>