[Pkg-openmpi-maintainers] Bug#494046: running mpi in sbuild (petsc testing)

[Graham Inggs]

In the build logs of a recent rebuild of PETSc [1] for the OpenMPI
1.10 transition, I see the following warning (and the build was
successful):

--------------------------------------------------------------------------
A deprecated MCA variable value was specified in the environment or
on the command line.  Deprecated MCA variables should be avoided;
they may disappear in future releases.

  Deprecated variable: orte_rsh_agent
  New variable:        plm_rsh_agent
--------------------------------------------------------------------------

Changing the line below in debian/rules

export OMPI_MCA_orte_rsh_agent=/bin/false

to

export OMPI_MCA_plm_rsh_agent=/bin/false

silences the warning.  It seems upstream can't make up their minds.


[1] https://buildd.debian.org/status/fetch.php?pkg=petsc&arch=amd64&ver=3.6.2.dfsg1-3%2Bb3&stamp=1455012964


On 9 October 2015 at 12:10, Johannes Ring <johannr at simula.no> wrote:
> Yes, I also get the deprecation warning and OMPI_MCA_plm_orte_agent does not
> work for me either. However, OMPI_MCA_orte_rsh_agent works like a charm.
>
> Johannes
>
> On Fri, Oct 9, 2015 at 11:55 AM Drew Parsons <dparsons at debian.org> wrote:
>>
>> On Fri, 2015-10-09 at 07:01 +0000, Johannes Ring wrote:
>> > I guess this is the same problem I have when building DOLFIN in
>> > pbuilder, which I solve by adding "export
>> > OMPI_MCA_plm_rsh_agent=/bin/false" in debian/rules.
>>
>>
>> Thanks Johannes, that more or less gets it running.  It gives a warning
>>
>> --------------------------------------------------------------------------
>> A deprecated MCA parameter value was specified in the environment or
>> on the command line.  Deprecated MCA parameters should be avoided;
>> they may disappear in future releases.
>>
>>   Deprecated parameter: plm_rsh_agent
>> --------------------------------------------------------------------------
>>
>> but otherwise the test runs fine.
>>
>> There's a suggestion at
>>  http://lists.alioth.debian.org/pipermail/debian-science-commits/2014-J
>> anuary/024112.html
>> that we should use OMPI_MCA_plm_orte_agent instead.  But when I use
>> that, I get the same test error that I got without the variable.
>>
>> Drew
>>
>> >
>> > On Fri, Oct 9, 2015 at 5:24 AM Drew Parsons <dparsons at debian.org>
>> > wrote:
>> >
>> > > p.s. an example of the test failure is:
>> > >
>> > > [schumann:13424] [[INVALID],INVALID] ORTE_ERROR_LOG: Not found in
>> > > file ess_hnp_module.c at line 170
>> > > -----------------------------------------------------------------
>> > > ---------
>> > > It looks like orte_init failed for some reason; your parallel
>> > > process is
>> > > likely to abort.  There are many reasons that a parallel process
>> > > can
>> > > fail during orte_init; some of which are due to configuration or
>> > > environment problems.  This failure appears to be an internal
>> > > failure;
>> > > here's some additional information (which may only be relevant to
>> > > an
>> > > Open MPI developer):
>> > >
>> > >   orte_plm_base_select failed
>> > >   --> Returned value Not found (-13) instead of ORTE_SUCCESS
>>
>