Re: two nodes see each other as "down", no job migration

Michal, you are correct.  The current ClusterKnoppix is not matched to
the official openMosix version.  
For your cluster to work, the openMosix version used on each machine
must match. 
I would suggest that you also consider using Quantian which may offer a
better match for your needs. 
The openMosix in Quantian does match that in ClusterKnoppix.  Bruce

>>> Michal Szymanski <[email protected]> 2/19/2006 6:46:02 AM >>>
Hello,

I want to setup a cluster of 5-10 machines, using OM. To make some
tests
and gain experience, I started with a mini-cluster at home, on two
laptops connected via switch acting as a DHCP server, too.

When I boot ClusterKnoppix (ver. 3.6, kernel 2.4.27-om) on both
machines, the clustering works fine, jobs do migrate, openmosixview
shows everything perfectly.

As a next step, I tried to install OM stuff (RPMS from OM home pages,
kernel 2.4.26-openmosix1 - there is no 2.4.27 there, plus OM tools
plus
openmosixview) on one of the laptops which still has Fedora Core 1
hard-installed. Installation worked fine, the services seem to run (in
autodiscovery mode - I did not do any changes in the configuration
that
was autoinstalled with the RPMs) The other laptop has FC3 installed,
so
I stayed with LiveCD ClusterKnoppix on it.

Sadly, in this configuration OM does not work. Both machines see the
other node as "down", see below. Obviously, there is no job migration
etc.

Any ideas? Browsing the mailing list archive I have found hints
suggesting that *all* of the nodes *must* run the very same version of
the kernel/modules - could that be the reason? Would the OM be that
self-incompatible? 

BTW, it is a pity that "mosctl" does not recognize hex format (as
output
by showmap)

regards, Michal.
-- 

  Michal Szymanski (msz at astrouw dot edu dot pl)
  Warsaw University Observatory, Warszawa, POLAND

==========================================
On FC1+OM

>setpe -r
This is OpenMosix node #357
Network protocol: 2 (AF_INET)
OpenMosix range   357-357   begins at guagua
OpenMosix range   360-360   begins at guagua2
Total configured: 2

>showmap
My Node-Id: 0x0165

Base Node-Id Address          Count
------------ ---------------- -----
0x0165       192.168.1.101    1
0x0168       192.168.1.104    1

>mosctl stat 357
up.
>mosctl stat 360
down.

=================================================
on ClusterKnoppix:

$ setpe -r
This is OpenMosix node #360
Network protocol: 2 (AF_INET)
OpenMosix range   360-360   begins at 192.168.1.104
OpenMosix range   357-357   begins at 192.168.1.101
Total configured: 2

$ showmap
My Node-Id: 0x0168

Base Node-Id Address          Count
------------ ---------------- -----
0x0168       192.168.1.104    1
0x0165       192.168.1.101    1

$ mosctl stat 360
up.
$ mosctl stat 357
down.
=================================================



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log
files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD
SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642

_______________________________________________
openMosix-general mailing list
openMosix-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org 
https://lists.sourceforge.net/lists/listinfo/openmosix-general


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642