Re: MPI problem?

[Jean-Noël Grad]

Hi,

The OpenMPI FAQ has more details on that warning 
(https://www.open-mpi.org/faq/?category=openfabrics#ofa-fork). I looked 
at the cluster specs (https://www.hpc.uni-freiburg.de/nemo), but they 
don't say much about the software. I also looked at the OpenMPI issue 
tracker 
(https://github.com/open-mpi/ompi/issues?utf8=%E2%9C%93&q=is%3Aissue) 
but couldn't find your error message.

Do you know if the OpenMPI version changed on the cluster on November 
29th, 2019? If so, you could look at the release notes 
(https://github.com/open-mpi/ompi/blob/master/NEWS) to see what changed 
and investigate the PRs of the corresponding milestone if you need even 
more details (https://github.com/open-mpi/ompi/milestones?state=closed).

Best,
JN

On 2/25/20 2:40 PM, Andreas Härtel wrote:
> Dear ESPResSo mailing list,
>
> Maybe anyone has an idea to help us:
> We are running ESPResSo on a cluster (BW cluster nemo).
> Starting with November 29th, 2019 we have seen messages from the cluster 
> management system saying: "Job does not exist on node". This does 
> happend only for a few percent of the jobs!
>
> Further, we see the message when starting ESPResSo:
> --------------------------------------------------------------------------
> An MPI process has executed an operation involving a call to the
> "fork()" system call to create a child process.  Open MPI is currently
> operating in a condition that could result in memory corruption or
> other system errors; your MPI job may hang, crash, or produce silent
> data corruption.  The use of fork() (or system() or other calls that
> create child processes) is strongly discouraged.
>
> The process that invoked fork was:
>
>    Local host:          n3765 (PID 172209)
>    MPI_COMM_WORLD rank: 0
>
> If you are *absolutely sure* that your application will successfully
> and correctly survive a call to fork(), you may disable this warning
> by setting the mpi_warn_on_fork MCA parameter to 0.
> --------------------------------------------------------------------------
>
> We do not know about the existence of a connection between both.
> Can anyone give us a hint?
>
> Thank you and best regards,
> Andreas