Newsgroups: comp.parallel.mpi
From: Mathias Waack <mathias@mufasa.informatik.uni-mannheim.de>
Subject: Problem starting programm on workstation clusters
Organization: Dept. of Computer Science, University of Mannheim, Germany
Date: Wed, 03 Sep 1997 11:38:58 +0200
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <340D3032.7CD6@mufasa.informatik.uni-mannheim.de>

Hallo,

we're running MPICH on a cluster of SUN-stations. Our login-shell 
is a tcsh and it reads a lot of config-files on startup. Thats why 
it takes some time to start a process on another host with rsh. 
The result is, that mpi is running into a timeout during startup:

mathias@shenzi (~/source/java_mpi) mpirun -pg -p4pg prg.file
/opt/java/JDK/bin/../bin/sparc/green_threads/java Test
p0_29038:  p4_error: Timeout in making connection to remote process on
schalu: 0
bm_list_29055:  p4_error: interrupt SIGINT: 2
P4 procgroup file is prg.file.

A workaround is to move the .cshrc to another location, but it 
isn't a good solution. Is it possible, to solve this problem 
in a better way, for instance to start the tcsh on the remote 
hosts with option -f or to start another shell?

Thanx in advance for your answers,

	Mathias

-- 
Mathias Waack		|     mathias@mufasa.informatik.uni-mannheim.de
Tel.:  +49 621 292 1620  Fax.:  +49 621 292 5597

