Hello,
Please don't worry about this for now, the problem stems from
iptable rules. But, I still think putting usb0 into the reject
list should disable the ip-address associated with it.
Regards,
>>>>> "Avinash" == Avinash Malik <ama...@scss.tcd.ie> writes:
Avinash> Hello,
Avinash> I have two mahcines each having 3 live interfaces:
Avinash> lo, eth0 (interanet) and usb0 (internet). eth0 cannot
Avinash> access usb0 on the other machine (and vice-veras). Now,
Avinash> when I try to run the MPI program with these two hosts I
Avinash> cannot get any output, even --mca btl_base_verbose 30 does
Avinash> not give any output. If I set hostfile to have only
Avinash> localhost, then everything runs fine.
Avinash> I tried out the same code and hostfile with two
Avinash> other machines with two interfaces: lo and eth1, which can
Avinash> access each other. The program runs fine on these machines.
Avinash> Next, I setup btl_tcp_if_exclude to lo,usb0 (on the
Avinash> first arch) and also the ip-address/mask, but this does not
Avinash> work either. When I run the program on one machine and do
Avinash> "ps aux | grep mpi" on the other I can see --hnp-uri being
Avinash> set to the usb0's ip-address, which it should not, because
Avinash> I have set usb0 to be exluded in the btl_tcp_if_exclude
Avinash> list. So, what exactly am I doing wrong here?
Avinash> I read the otimization FAQ and saw how openmpi
Avinash> builds the bipartite graphs for connection. But, as I said
Avinash> before, eth0 cannot access usb0's ip and vice-versa, how
Avinash> can I get rid of the usb0 ip-address showing up in
Avinash> --hnp-uri, because this is the only difference between the
Avinash> working and the non-working archs.
Avinash> Regards, -- Avinash Malik
Avinash> _______________________________________________ users
Avinash> mailing list us...@open-mpi.org
Avinash> http://www.open-mpi.org/mailman/listinfo.cgi/users
--
Avinash Malik
