LWG Minutes 2017-09-07

Attendance

Cray: Ben Evans, Cory Spitz
ORNL: James Simmons, Dustin Leverman
SuperMicro: Abe Asraoui
Sandia: Ruth Klundt
HPE: Olaf Weber, Christopher Voltz, Robert Clark
Intel: Joe Gmitter, Peter Jones, Andreas Dilger

Actions

New Actions Captured:

• None

Existing Open Actions:

• None

Actions Recently Closed:

• None

Minutes

2.10.1 Release Update
Peter

• The 2.10.1 release remains in a holding pattern due to waiting on the CentOS 7.4 being release.

2.11.0 Release Update
Peter

• There has been quite a bit of activity on lockahead and it appears to be trending well. Oleg has it in master-next for further testing.
• The next item from Intel that will be going in is Data on MDT, which is currently under testing and reviews.
• James: I see that tracepoint is also in master-next, so this should be appearing in master soon as well.
• James: Is there any more LNet Multi-Rail work going in?
  • Olaf: We are working on a bug found in Oleg's testing, but nothing other than bug fixes to the already landed code is planned.

Upstream Lustre Client
James

• The next window is opening up soon.
• Have been spending about a week or so fixing and repairing changelog and sysfs bugs.
• Andreas: Is the upstream Lustre client in a useable state?
  • James: It does work, but it can't really be tuned due to the sysfs breakage. LU-8837 is tracking the work to start to make master buildable against the upstream client.

lustre.org
Ken

• We are currently in the process of transitioning the remaining architecture pages on the old wiki (approximately 70) and their attachments to static versions as discussed previously. Once this is finished the wiki migration effort will be complete and we'll begin the process of formally shutting down the old wiki.
• From an internals standpoint, let's plan on discussing LNET and LDLM on the following call.

Other Business

• Cory: What is happening with openSFS test cluster? Does IU still host it?
  • Peter: Ken brought it up and is asking if it is still being used by anyone since evidence shows that is hasn't been used for a long period of time. Therefore, they are discussing taking it out of service soon.
  • Cory: Can we find a use for it if IU is willing to continue to host it?
    • Andreas: It has a very old generation of flash devices. When we were testing DNE2 for performance it was not useful. In the Intel side, functional testing is able to be done for internally, so we have not been using it at all for that purpose either.

Next meeting will be on 2017-09-21 at 11:00am Pacific