ALINKA Linux Clustering Letter, September 26th. 2001 (fwd)
Eugene.Leitl at lrz.uni-muenchen.de
Thu Sep 27 08:32:11 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, 27 Sep 2001 14:25:04 +0200
From: Antoine Brenner <abrenner at alinka.com>
To: clustering at alinka.com
Subject: ALINKA Linux Clustering Letter, September 26th. 2001
Resent-Date: Thu, 27 Sep 2001 14:25:08 +0200 (CEST)
Resent-From: clustering at alinka.com
The ALINKA Linux Clustering Letter,
Wednesday, September the 26th. 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)
Tips and tricks from the Beowulf mailing list
* Delcides Flávio de Sousa Jr. [m1] is wondering which kind of
configuration he needs to start its beowulf cluster. As Greg Lindahl
[m2] try to answer to its questions, it depends on applications.
* Shane Canon [m1] has problem to configure linpack, Greg Lindahl [m2]
and Patrick Geoffray [m3] explained how to set the problem size.
* Jalal Haddad [m1] is looking for cheap dual P4 motherboard, Joel
Jaeggli [m2] answered that actually the only available chipset is
the intel 860 which is expensive. Ron Chen [m3] gave two
links [1,2] where Dual P4 and Dual Athlon are compared, showing that
performances depends on the compiler and the code [m4]. About PCI
bus speed, Greg Lindahl [m5] posted a link  where some comparison
have been done.
News from MOSIX mailing list by Benoit des Ligneris
<bligneri at physique.usherb.ca>
* Jame Troup ask for support [m1] in order to use Mosix in the Linux
Terminal Server Project . An old debate about development of mosix
for the ``last'' linux kernel arise.
Only last versions of Mosix are developped and patches are not ported
to old version. As stated by Thomas Webb [m2] it's not good for a
production environment !
* Discussion lauched by John Strange about GFS + Mosix [m1]. One of the problem
was mexecd that did not work on recent Mosix kernels. Paul Mundt told [m2]
that he has almost all rewritten for recent mosix. Then Jacob Gorn Hansen
[m3] says that CODA/intermezzo will surely have better performance.
* Tim Chipman asked [m1] for a simple queue management for mosix cluster.
John Strange [m2] suggested Gnu Queue  or Sun Grid  and lots of other
posters gave advices for a "home made" solution but nothing concrete and
News from the High Availability world
DRBD devel by Guillaume GIMENEZ (ggimenez at alinka.com)
* Martin Bene posted a patch [m1] to solve some troubles. it applies
against drbd 0.61-pre2 and is useful for redhat users.
Linux-HA dev by Rached Ben Mustapha (rached at alinka.com)
* Alan Robertson proposed [m1] that dependencies on resources be
implemented in future versions of linux-ha software.
Linux-HA by Rached Ben Mustapha (rached at alinka.com)
* Jean-Yves Bouet asked [m1] what was the role of job of each thread
in heartbeat. Alan Robertson replied [m2] this:
master control process: read requests to send messages from a FIFO,
and sends them to the master status proces
Master status process, tracks status of nodes, initiates takeovers,
handles heartbeat protocol, etc.
Each medium declared in the ha.cf file has a read and a write process.
LVS by Rached Ben Mustapha (rached at alinka.com)
* malalon at poczta.onet.pl asked [m1] what was the problem with testing
a NAT-based service, with rr scheduling. Joseph Mack replied [m2]
that the recommended method to test lvs was with telnet.
* Dmitry Dan Brovkovich asked [m3] why trying to telnet to the VIP
resulted in an inactive connection, and the client not receiving
anything. Joseph Mack replied [m4] that the realserver probably
didn't have its root set properly to the client.
News on the Filesystems front
Coda by Ludovic Ishiomin (lishiomin at alinka.com)
* Jan Harkes posted a patch specific to Coda kernel code on Linux that makes
Coda client be able to use other filesystem than ext2 [1m].
* Phil Nelson gave information about Coda on Solaris 8 [2m].
* Jan Harkes posted comments about Coda and Intermezzo [3m].
Intermezzo by Ludovic Ishiomin (lishiomin at alinka.com)
* Shirish Phatak posted status about Reiserfs support with Intermezzo [1m].
JFS by Ludovic Ishiomin (lishiomin at alinka.com)
* Steve Best announced JFS 1.0.5 [1m].
* Dave Kleikamp posted a patch to the latest release of JFS [2m].
* Anthony Liu confirmed that this drop won't compile with 2.4.10 kernel because
of major VM changes [3m].
XFS by Ludovic Ishiomin (lishiomin at alinka.com)
* There was a long thread about mainstream kernel inclusion of XFS [1m].
and the followings
News on other cluster related topics
Linux Cluster by Bruno Muller (bmuller at alinka.com)
* Lars Marowsky-Bree announced the Linux Kongress 2001 Clustering
LVM by Bruno Muller (bmuller at alinka.com)
* AJ Lewis posted a patch to configure LVM 1.0.1-rc2 [m1].
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
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