Hi guys,
I'm currently integrating LAM 7.0.2 for the next release of CLIC
distribution (this a real clustering oriented clustering distribution
http://clic.mandrakesoft.com)
I'm facing the following problem:
My server owns 2 NIC, one for computing, one for the administrative
tasks.
Each NIC is in a separated network one in 172.16.X.X (admin), the other
in 10.0.X.X (computing).
Each node & the server own to IP address, each IP address is matched to
a DNS alias.
The problem is facing is: I'm asking my server to start lam (using
lamboot) to the nodes using theirs computing name (nodex.domcomp.com =
10.0.1.253).
So lamboot is rshing my first node but in answer nodes are trying to
answer to the IP address that match the hostname of my server
(172.16.1.253).
So the nodes are trying to answer using their admin interface & IP
address. So lamboot is rejecting this answer (unexpected connection).
It is possible to ask lamboot to give the nodes the IP address of the
NIC I want and not giving the IP Address that equals my hostname ?
Of course, I've tried by changing the hostname of the nodes and it
works.
One part of the traces...
[..]
n0<22000> ssi:boot:rsh: attempting to execute "/usr/bin/rsh
compute1.domcomp.com -n hboot -t -c lam-conf.lamd -d -s -I "-H
172.16.1.253 -P 44039 -n 0 -o 9""
[..]
n0<21092> ssi:boot:rsh: successfully launched on n0
(compute1.domcomp.com)
n0<21092> ssi:boot:base:server: expecting connection from finite list
n0<21092> ssi:boot:base:server: got connection from 172.16.1.1
n0<21092> ssi:boot:base:server: unexpected connection; dropping
--
Erwan Velu
Linux Cluster Distribution Project Manager
MandrakeSoft
43 rue d'aboukir 75002 Paris
Phone Number : +33 (0) 1 40 41 17 94
Fax Number : +33 (0) 1 40 41 92 00
Web site : http://www.mandrakesoft.com
OpenPGP key : http://www.mandrakesecure.net/cks/
|
