ALINKA Linux Clustering Letter, October 3rd. 2001 (fwd)

Eugene Leitl Eugene.Leitl at lrz.uni-muenchen.de
Thu Oct 4 08:59:17 EDT 2001



-- Eugen* Leitl <a href="http://www.lrz.de/~ui22204/">leitl</a>
______________________________________________________________
ICBMTO: N48 04'14.8'' E11 36'41.2'' http://www.lrz.de/~ui22204
57F9CFD3: ED90 0433 EB74 E4A9 537F CFF5 86E7 629B 57F9 CFD3

---------- Forwarded message ----------
Date: Thu, 04 Oct 2001 14:44:23 +0200
From: Antoine Brenner <abrenner at alinka.com>
To: clustering at alinka.com
Subject: ALINKA Linux Clustering Letter, October 3rd. 2001
Resent-Date: Thu,  4 Oct 2001 14:44:25 +0200 (CEST)
Resent-From: clustering at alinka.com


The ALINKA Linux Clustering Letter,

Wednesday, October the 3rd. 2001

Dear readers, I am happy to send you this week's edition of
clustering at alinka.com

clustering at alinka.com is a free weekly e-mail newsletter on linux
clustering. It provides a summary of the weekly activity in
mailing-lists relative to linux clustering (such as beowulf,
linux virtual server or linux-ha) and general clustering news.
For more information about ALINKA, see: http://www.alinka.com


News from the High Performance world, by Dr Laurent Gatineau
(lgatineau at alinka.com)
======================================================================
Software for Beowulf cluster
========
SCE V1.2 (Scalable Computing Environment) [1]
Scalable    Cluster Environment  (SCE) is    a   set of  interoperable
opensource tools that  enable users to build  and  use Beowulf cluster
effectively to solve their problems.
[1] http://prg.cpe.ku.ac.th/research/sce/

Tips and tricks from the Beowulf mailing list
========
* Bruno Richard [m1] posted a link (the  good link is there [1]) about
  Linpack   scalability  on  a   mainstream  cluster.   Definition  of
  mainstream cluster (vs other cluster) is given there [m2].
[1] http://www.hpl.hp.com/techreports/2001/HPL-2001-206.html
[m1] http://www.beowulf.org/pipermail/beowulf/2001-September/001255.html
[m2] http://www.beowulf.org/pipermail/beowulf/2001-September/001343.html

* David Vos  [m1] is  looking for  information  about Gaussian   98 on
  RedHat  7.x.   Gary Stiehr  [m2]  reported that  it run  fine with a
  RedHat 7.1.
[m1] http://www.beowulf.org/pipermail/beowulf/2001-September/001358.html
[m2] http://www.beowulf.org/pipermail/beowulf/2001-October/001378.html

* Sebastien  Cabaniols [m1] is  looking for information about swapping
  over the network (for diskless  nodes).  Mark Hahn [m2] is wondering
  if its a good idea because doing network needs memory, he wrote that
  NBD [1] should work.   Donald  Becker [m3]  wrote that  Linux kernel
  2.2.x could run without swap.
[1] http://atrey.karlin.mff.cuni.cz/~pavel/nbd/nbd.html
[m1] http://www.beowulf.org/pipermail/beowulf/2001-October/001366.html
[m2] http://www.beowulf.org/pipermail/beowulf/2001-October/001371.html
[m3] http://www.beowulf.org/pipermail/beowulf/2001-October/001373.html

* Brian  Korsedal    [m1] wants  to   use free   CPU  of  this  office
  station. Some Beowulfers    [m2, m3, m4]   pointed him  to  SGE [1],
  Condor [2].
  Eray Ozkural   [m5, m6] is  looking  for information  on  how to run
  parallalel batch jobs  on cluster, preventing  users to log on nodes
  and run their own   application.   Gary Stiehr [m7]  answered   that
  OpenPBS  answer to  its first    question,  but he will need    some
  middleware for the second point.
[1] http://www.sun.com/software/gridware/
[2] http://www.cs.wisc.edu/condor
[m1] http://www.beowulf.org/pipermail/beowulf/2001-September/001328.html
[m2] http://www.beowulf.org/pipermail/beowulf/2001-September/001339.html
[m3] http://www.beowulf.org/pipermail/beowulf/2001-October/001395.html
[m4] http://www.beowulf.org/pipermail/beowulf/2001-September/001340.html
[m5] http://www.beowulf.org/pipermail/beowulf/2001-October/001380.html
[m6] http://www.beowulf.org/pipermail/beowulf/2001-October/001383.html
[m7] http://www.beowulf.org/pipermail/beowulf/2001-October/001387.html

* Nick  Gregory  [m1]   wants   the    best  CPU for   1U node  (AMD,
  P4...). Actually  there  is no  way to  have  P4 in   1U [m2].   Rob
  Myers [m3] explained that compilers  and cooling are  more important
  than CPU.
[m1] http://www.beowulf.org/pipermail/beowulf/2001-September/001291.html
[m2] http://www.beowulf.org/pipermail/beowulf/2001-September/001331.html
[m3] http://www.beowulf.org/pipermail/beowulf/2001-September/001341.html

* Joel Jaeggli  [m1] and Donald Becker  [m2] gave adivecs  about Fiber
  Gigabit Ethernet cards.
[m1] http://www.beowulf.org/pipermail/beowulf/2001-September/001321.html
[m2] http://www.beowulf.org/pipermail/beowulf/2001-September/001346.html


News from MOSIX mailing list by Benoit des Ligneris
<bligneri at physique.usherb.ca>
=====================================================================
* Gregory R. Wames announced [m1] the release of a new ClusterNFS version [1].
  It's a project very useful for managing diskless clusters (MOSIX ones in
  particular).
[m1] http://www.mosix.cs.huji.ac.il/month-arch/2001/Sep/0132.html
[1]  http://clusternfs.sourceforge.net/

* Tim Chipman [m1] share his statistics on the scaling of MOSIX clusters and
  find a "Saturation Effect" nodes/jobs. Andreas Korn says that the some
  syscalls get very slow when executed on the home node. Eric Whiting
  suggested some explanations and tests to proove it.
[m1] http://www.mosix.cs.huji.ac.il/month-arch/2001/Oct/0007.html
[m2] http://www.mosix.cs.huji.ac.il/month-arch/2001/Oct/0008.html
[m3] http://www.mosix.cs.huji.ac.il/month-arch/2001/Oct/0011.html



News from the High Availability world
======================================================================
Failsafe by Guillaume GIMENEZ (ggimenez at alinka.com)
========
* Joachim Gleissner announces [m1] new release of failsafe for SuSE
is available [1]

[1] ftp://ftp.suse.com/pub/projects/failsafe
[m1] http://lists.community.tummy.com/pipermail/linuxfailsafe/2001-September/001237.html


Linux-HA dev by Rached Ben Mustapha (rached at alinka.com)
========
- Alan Robertson posted [m1] a little summary of linux-ha software
  progress.

- Alan Robertson announced [m2] that version 0.4.9.0a was up for testing
  on the web site.

[m1] http://marc.theaimsgroup.com/?l=linux-ha-dev&m=100208889131477&w=2
[m2] http://marc.theaimsgroup.com/?l=linux-ha-dev&m=100198954331171&w=2


Linux-HA by Rached Ben Mustapha (rached at alinka.com)
========
- Lorn Kay asked [m1] if two pairs of heartbeat servers could use the
  same physical network without generating lots of "failed authentication".
  Alan Robertson replied [m2] that to keep heartbeat broadcasting he
  should use separate port numbers, or use multicast with different multicast
  groups.

- Lars Marowsky-Bree posted [m3] a call for participation for the Linux
  Kongress 2001 Clustering Workshop, on November 26 - 28 2001, at
  Enschede, NL.
  More info is available on the mail archive.

[m1] http://marc.theaimsgroup.com/?l=linux-ha&m=100179377712154&w=2
[m2] http://marc.theaimsgroup.com/?l=linux-ha&m=100179487914051&w=2
[m3] http://marc.theaimsgroup.com/?l=linux-ha&m=100151424401828&w=2


LVS by Rached Ben Mustapha (rached at alinka.com)
========
- Alexandre Cassen announced [m1] the release 0.0.3 of LVSGSP [l1],
  that supports LVS FWMARK virtual server.

- Serge Sozonoff launched here [m2] a thread discussing about LVS and
  ethernet Bridging, which would be somewhere between VS-NAT and VS-DR
  in performance terms.

[m1] http://marc.theaimsgroup.com/?l=linux-virtual-server&m=100197878908286&w=2
[m2] http://marc.theaimsgroup.com/?l=linux-virtual-server&m=100161489013483&w=2
[l1] http://linuxvirtualserver.org/~acassen/lvsgsp-0.0.3.tar.gz



News on the Filesystems front
======================================================================
reiserfs by Guillaume Gimenez (ggimenez at alinka.com)
========
* Nikita Danilov advise us [m2] that there is a patch [2] to support
  immutable and noatime filesystem attributes.

[m2] http://marc.theaimsgroup.com/?l=reiserfs&m=100218568727101&w=2
[2] ftp://ftp.namesys.com/pub/misc-patches/linux-2.4.5-inode-attrs-1.d.gz


JFS by Ludovic Ishiomin (lishiomin at alinka.com)
========
* Steve Best announced JFS 1.0.6 [1m].

[1m] http://oss.software.ibm.com/pipermail/jfs-discussion/2001-September/000605.html



News on other cluster related topics
======================================================================
linux-ia64 by Guillaume GIMENEZ (ggimenez at alinka.com)
========
linux-ia64
========
* Keith Owens announces [m3] kdb v1.9 for IA64  is available [3]
[3] ftp://oss.sgi.com/projects/kdb/download/ia64/
[m3] https://external-lists.valinux.com/archives//linux-ia64/2001-October/002235.html


LVM by Bruno Muller (bmuller at alinka.com)
=====
* Jim Cromie had a problem with lvcrete on LVM 1.0.1-rc2 and kernel
  2.4.10.[m1] John Marquart answered that he is running into a problem he
  had and there is a patch.[m2]
* AJ Lewis annouced that LVM 1.0.1-rc4 is available.[m3]
* Andre Margis posted a patch for lvm-1.0.1-rc4 to work under
  linux-2.4.10-ac4[m4]
[m1] http://lists.sistina.com/pipermail/linux-lvm/2001-September/008979.html
[m2] http://lists.sistina.com/pipermail/linux-lvm/2001-September/008980.html
[m3] http://lists.sistina.com/pipermail/linux-lvm/2001-October/009071.html
[m4] http://lists.sistina.com/pipermail/linux-lvm/2001-October/009088.html


======================================================================
To subscribe to the list, send e-mail to clustering at alinka.com from the
address you wish to subscribe, with the word "subscribe" in the subject.

To unsubscribe from the list, send e-mail to clustering at alinka.com from
the address you wish to unsubscribe from, with the word "unsubscribe" in
the subject.

Alinka is the editor of the ALINKA ORANGES and ALINKA RAISIN
administration software for Linux clusters.
(Web site: http://www.alinka.com )



_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list