I'm seeing this problem as well even running just 4 processes on a
single node (though not as frequently as with higher process counts).
The trick is to force Open MPI to use only tcp,self and nothing else.
Did you try adding this (-mca btl tcp,self) to the runtime parameter
On Apr 18, 2008, at 12:56 PM, Adrian Knoth wrote:
> On Fri, Apr 18, 2008 at 08:04:17AM -0400, Tim Prins wrote:
>> Hi Adrian,
>> After this change, I am getting a lot of errors of the form:
>> mca_btl_tcp_frag_recv: readv failed: Connection reset by
>> peer (104)
>> See for instance: http://www.open-mpi.org/mtt/index.php?do_redir=615
> That's weird. I've tried hello_c.c on about ten machines with
> network configurations, none of them showed any problems at all.
> Do you have a very special setup? And if need be, would it be possible
> to debug on your machine?
>> From all MTT sites, this error only occurs on Odin and Sif. What's so
> special with these clusters?
>> I have found this especially easy to reproduce if I run 16
>> processes all
>> with just the tcp and self btls on the same machine, running the
>> 'hello_c' program in the examples directory.
> Unfortunately, I can't reproduce it that way. If this is related to
> change, then it would mean that mca_btl_tcp_proc_accept() returns
> either after the large loop or in mca_btl_tcp_endpoint_accept().
> Do you have the cycles to add some BTL_VERBOSE-lines to see where
> go wrong? Or even to step through with the debugger?
> If you want me to do it, I would provide you with my ssh key?
> mail: adi_at_[hidden] http://adi.thur.de PGP/GPG: key via keyserver
> Das Sterben wird nur halb so schlimm, rauchst du KIM.
> devel mailing list