Update submodule pointers - do not commit

[rhc54]

Just testing the configuration changes in PMIx/PRRTE

Signed-off-by: Ralph Castain rhc@pmix.org

[ompiteam-bot]

Can one of the admins verify this patch?

[jsquyres]

ok to test

[rhc54]

@janjust The Mellonox CI is failing on test of an MCA param --mca mca_base_env_list 'XXX_A=1;XXX_B=2;XXX_C;XXX_D;XXX_E'. I suspect this was relying on PMIx to pick it up, which it no longer does as that MCA param has nothing to do with PMIx.

Is this a behavior you wish to preserve? You might need to pick it up in the schizo/ompi component. Note that we do pick up the same behavior from the cmd line in PRRTE - just not as an MCA param.

[rhc54]

@jsquyres @gpaulsen Can you please bring this up on today's telecon? You folks need to decide what to do about Mellanox CI and the path forward so you can commit PRRTE/PMIx updates.

[gpaulsen]

@rhc54 Yes, we're discussing and will discuss at today's Open MPI call.

[gpaulsen]

Our consensus was that we'll implement something in schizo/ompi component for v5.
Shorter term, mellanox will remove the test(s) so that we can update our submodule pointers until we have the new schizo/ompi code, then they will re-enable the mca_base_env_list tests to test long term.

[rhc54]

Okay, thx! I'll close this as it appears everything else is fine.

[artemry-nv]

Disabled the problem test https://github.com/mellanox-hpc/jenkins_scripts/blob/1d66874bec127c22a2248b8d0da4cb5c7bdf5158/jenkins/ompi/ompi_test.sh#L230