Newsgroups: comp.parallel.mpi
From: Lothar Thum <thum@mathpool.uni-augsburg.de>
Subject: LAM: hboot-Problem
Organization: Leibniz-Rechenzentrum, Muenchen (Germany)
Date: Tue, 03 Sep 1996 13:52:46 +0200
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <322C1C0E.41C6@mathpool.uni-augsburg.de>

Hello,

because of our system (OSF DCE) I cannot use 'lamboot' to start the LAM
deamons (because there is no 'rsh' possible) and cannot use other
useful tools either.
I have to start the deamons manually with 'hboot'.
That's no big problem, but sometimes when I create a 'virtual machine'
with more than two nodes it runs into a segmentation fault. E.g.

I've got three machines in a LAN:
aixcip02
aixcip03
aixcip10

Then I try to start the LAM deamon on machine aixcip02:

% hboot -vc conf.lam -I"-n0 -o0 aixcip03 1 aixcip10 2"

it prints out:

% Segmentation fault(coredump)

(a core file is created)


Can anyone explain or help ?

Thanks in advance,
Lothar.

