Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

Subject: [OMPI users] connection problem in btl_tcp_endpoint.c
From: J.S. van Bethlehem (j.s.van.bethlehem_at_[hidden])
Date: 2009-08-31 09:49:31


Dear users,

I'm not sure whether this is the right place to go to with my problem,
but maybe someone can give me some leads. I'm trying to run 'Gadget2'
using OMPI 1.3.3. The installation seems fine; I can run simple programs
on as many machines/nodes I want using a machinefile. I can also run
Gadget2 on the CPUs of a single machine, but when I try to start the
program distributed over multiple machines I get:

[virgo02][[20218,1],0][btl_tcp_endpoint.c:638:mca_btl_tcp_endpoint_complete_connect]
connect() to 172.16.189.1 failed: Network is unreachable (101)
[virgo02][[20218,1],1][btl_tcp_endpoint.c:638:mca_btl_tcp_endpoint_complete_connect]
connect() to 172.16.189.1 failed: Network is unreachable (101)

here virgo02 is the name of the first machine in the machinefile:
$ head -n 1 machinefile
virgo02 slots=2

Since I don't have this problem when running other simple programs, my
initial guess is that it has something to do with Gadget2, however this
particular message is created by OMPI. So has anyone ever seen this
before or does anyone has more concrete ideas about what could be going on?

Greetings,
Jakob van Bethlehem