[Beowulf] Beowulf cluster mpi problem
chandu_dreams2001 at yahoo.co.in
Fri Jan 6 07:04:54 EST 2006
I have built a Linux cluster with a master and a
slave. The communication medium is ssh. ssh works
without any passwords. this has been acheived using
the public keys. I have installed mpich 1.2.7 p1. I am
using Fedora core 2.
when i run a mpi program on the master..
If i run mpirun nwith 1 processor everything works
[root at master basic]# mpirun -np 1 cpi
Process 0 of 1 on master.mydomain.com
pi is approximately 3.1415926544231341, Error is
wall clock time = 0.000634
But if i run it with 2 procesors it gives me this
[root at master basic]# mpirun -np 2 cpi
rm_3771: p4_error: rm_start: net_conn_to_listener
p0_4293: p4_error: Child process exited while making
connection to remote process on client1.mydomain.com:
p0_4293: (10.253768) net_send: could not write to
fd=4, errno = 32
[root at master basic]#
I do not know where the problem is. I have updated the
machines.LINUX file. it has the client's host name
"client1.mydomain.com". SSH connection exists between
master and client.
Yahoo! DSL Something to write home about.
Just $16.99/mo. or less.
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
More information about the Beowulf