-
Notifications
You must be signed in to change notification settings - Fork 864
WeeklyTelcon_20170926
Geoffrey Paulsen edited this page Jan 9, 2018
·
1 revision
- Dialup Info: (Do not post to public mailing list or public wiki)
- Jeff Squyres (Cisco)
- Geoff Paulsen (IBM)
- David Bernholdt (ORNL)
- Edgar Gabriel
- Geoffroy Vallee (ORNL)
- Howard
- Joshua Hursey
- Todd Kordenbrock
- Joshua Ladd
- Geoffroy Vallee (ORNL)
- Artem (Mellanox)
- Joshua Ladd (Mellanox)
- Thomas Naughton
Review All Open Blockers
Review v2.0.x Milestones v2.0.4
- Going to switch v2.0.x to only Critical fixes only!
- Only Critical fix we know of now is MAdvise fix.
- Ask people to move to v2.1.x or v3.0.0
- If nothing else critical Howard and Jeff will make an RC soon.
- targeting Oct 21st for release.
- Should be pretty easy.
Review v2.x Milestones v2.1.2
- v2.1.3 (unscheduled, but probably jan 19, 2018)
- PR4172 - a mix between feature / bugfix.
Review v3.0.x Milestones v3.0
- v3.0.1 - Opened the branch for bugfixes Sep 18th.
- Looking at Oct 17th
- ortedvm is broken on v3.0.0
- Discussed and pushing to v3.1 due to high number of orted changes.
Review v3.1.x Milestones v3.1](https://github.com/open-mpi/ompi/milestone/27)
- Plan to branch from Master moved to Tuesday Oct 3rd.
- Plan to create first RC Tuesday Oct 3rd after branching.
- gives us 4 weeks to stabilize and release before supercomputing.
- PMIx 2.1 should get in in time for v3.1
- One new feature is cross version compatibility.
- PMIx version 2.x will support one step back, PMIx v1.x Not sure if it support v1.0 and v1.1 and v1.2
- Discuss next week exactly what this supports.
- useful for slurm build with older PMIx.
Review Master Master Pull Requests
- proc_hostname code not coded correctly for 3 years. git bisect from PMIx from 2 weeks ago
- Giles posted a fix
- Someone from PMIX should look at, was this a latent bug that brought forward, or were we just getting lucky?
- Does this affect other branches? Other branches also didn't initialize proc_hostname to NULL.
- Segfaults in Finalize (teardown in proc, tries to free a bogus value
- Related issue: Also discuss preventative programming with calloc vs malloc for proc_hostname type future issues.
- discussion on devel mailing list. Need to understand.
- Howard having issues with reaching out and getting ID from MTT. Josh isn't sure.
- Josh had a breakthrough on Python Client.
- Python 3 users will need this PR to try: https://github.com/open-mpi/mtt/pull/561
- Please try out Python Client if you can.
- This is the future, just a matter of time before everyone should switch.
- Other than Cisco's new proc_hostname issue, looking pretty good.
- Artem is seeing an Out of Resource error (filesystem) on AWS.
- Boris will try to reproduce this, but if can't reproduce, it would be nice if AWS could check what is leftover there.
- Error appears at fallocate(), using it to reserve space for dstore. Get contents of directory.
- Probably not file descriptor.
Review Master MTT testing
- C++ removal - https://github.com/open-mpi/ompi/pull/1389
- If we Pull, it would cause a major version bump, but don't want this to drive a major version bump.
-
Need to see if Attributes are MT - IBM will see if we have any tests to audit.
- Asked, need to get answer back from them.
- Jan / Feb
- Possible locations: San Jose, Portland, Albuquerque, Dallas
- Mellanox, Sandia, Intel
- LANL, Houston, IBM, Fujitsu
- Amazon,
- Cisco, ORNL, UTK, NVIDIA