I maintain the osc/sm component but did not write the pscw
synchronization. I agree that a counter is not sufficient. I have a fix
in mind and will probably create a PR for it later this week. The fix
will need to be applied to 1.10, 2.x, and master.
-Nathan
On Fri, Sep 18, 2015 at 10:33:18AM +0
Patrick,
thanks for the report.
can you confirm what happened was
- you defined
OMPI_MCA_plm_rsh_agent=oarshmost
- oarshmost was not in the $PATH
- mpirun silently ignored the remote nodes
if that is correct, then i think mpirun should have reported an error
(oarshmost not found, or cannot remot
Hi Gilles,
I've done a big mistake! Compiling the patched version of openMPI and creating a
new module, I've forgotten to add the path to oarshmost command while
OMPI_MCA_plm_rsh_agent=oarshmost was set
OpenMPI was silently ignoring oarshmost command as it was unable to find it and
so only
- Mensaje original -
> De: "Ralph Castain"
> Para: "Open MPI Users"
> Enviado: Lunes, 21 de Septiembre 2015 1:42:08
> Asunto: Re: [OMPI users] send() to socket 9 failed with the 1.10.0 version
> but not with 1.8.7 one.
>
> Okay, let’s try doing this:
>
> mpirun -mca oob_tcp_if_include
The project I’m working on is a larger toolchain (written in Python) to run
regression tests. The part that does the data comparison is fairly small. Speed
is not crucial but doing the data comparison in python was incredibly slow. So
we went with a C++ extension. For everything else python work
Okay, let’s try doing this:
mpirun -mca oob_tcp_if_include br0 …
This will restrict us to the br0 interface that is common to the two nodes. I
note that your “node1” has two interfaces on the same subnet (192.168.1), which
is usually a “no-no” that can cause trouble. Let’s see if removing that
