On Sep 24, 2008, at 9:27 PM, xuejun gu wrote:
> I tried lamboot, but failed. The attached(bootdebug file) is the
> output information obtained with -d, which I can not interperate.
>
> I tried telnet on the remote host and get 'a Connetction refused'
> error. Does it mean no connection problem on this cluster?
>
> Can anybody give me some tips or help me to figure out what's wrong
> with this cluster? Thanks a lot
As the error message suggests, this is almost always caused by a
firewall running on one of the machines. LAM/MPI uses random port
numbers and requires that all nodes be able to contact all other nodes
in the cluster using both UDP and TCP.
Hope this helps,
Brian
--
Brian Barrett
LAM/MPI Developer
Make today a LAM/MPI day!
|
