Core Committer Weekly Interlock - August 21st 2017

Attendees

Agenda

2.19.0 Release and MasterCI status
1. 2.19.0 Sprint release failed on 8/18/17. MasterCI continues to fail. EOS Artifactory went down ~1pm EST on 8/18/17.
2. Peter was in to migrating from EOS Artifactory to MN based Artifactory as a workaround. Per Peter:
  1. Master CI was in red because EOS Artifactory met problem on upload/download deb.
    I was trying to migrate CI to MN Artifactory, but taken it requires certification verification and I can’t simply skip verifying in code due to security concern, so I decide to wait for EOS team’s fix, which will be faster.
    EOS team : Anastasia.Zinner@emc.com is working on the fix.
    Once she fixed it in US day time , US team can help to re-kick off the Master CI.
  2. AI: Former user (Deleted) to follow-up with Anastasia for expediting a fix during US daytime.
3. Once MasterCI is back to a passing state, P1s to be updated, and sprint release to be re-started to publish the 2.19.0 release.
From QRB meeting notes

Agreement at OLT that we will be going fully wsman-based and eliminate racadm from workflow support.
1. AI: Amy Mullins Draft proposal created and socialized via email. Will ping hte team one last time for technical accuracy and Leo Zhang / Maglev team will work with Thomas Sullivan on generating the official KI
2. Test plan: the idrac simulation tool will be used for virtualized testing (PR quality gates/MasterCI) and introduce more Dell physical hardware to the Regression-Baremetal job.
  1. Former user (Deleted) has downloaded the idrac simulation tool, currently under evaluation.
  2. Approval received to order 14 g hardware and add to a hopkinton based Dell stack.
  3. Michael Hepfer , Former user (Deleted), and Amy Mullins to sync up via email on if RR gear needs to be shipped or if we could re-purpose what was used for Veyron /RackHD feature dev.

Ease of use of the Vagrant based demo
1. Week of 8/7 comment on the slack channel: "At this point I've given up on rackhd. If even the demo requires an old version of ubuntu to run an old version of virtualbox to get it working, I will stick with something simpler."
  1. priority must be on keeping it simple and easy to use
  2. create specific requirements that the team agrees upon, use that to come to consensus on the next steps
2. Suggestion 1 - easier to create script to re-use the nightly built docker images + infrasim.
  1. disadvantage is using docker compose yml missing the network setup, linux specific demo only
  2. What are the RackHD OS requirements? If linux only, this solution would fit. See 3.a.ii above, requirements to be generated.
  3. POC - Former user (Deleted) + requirements generation
3. Suggestion 2 - use vagrant script to deploy rackhd (docker) + infrasim (similar to Suggestion #1, but using the vagrant script for deployment to meet an OS agnostic requirement)
  1. advantage - OS agnostic, virtual switch deployment can be done in a ruby script
  2. disadvantage - multiple vagrant versions , also thought to be an advantage as the vagrant file version can be bumped.
  3. Move to Ubuntu 16.04 with docker installed to be the base vagrant instance
    1. Currently available in Concourse based env
    2. Will make available on 8/23 for eval to all teams
    3. Former user (Deleted) to provide details on what's available as POC.
4. Suggestion 3 - run docker in docker, leveraging existing docker images
  1. privileges, networking, and performance concerns with this solution
RackHD Tooling Updates (did not complete discussion during the meeting, will add to the agenda for the next meeting)
1. Ubuntu to be upgraded to 16.04, 18 to be released ~ April 2018
  1. Docker distributed on 16.04 is a little old, Michael Hepfer tends to upgrade to a newer version. KI should include testing of pulling down the latest version.
  2. KI needs to include that on-imagebuilder (overlayfs images, microkernel) can be built
  3. KI needs to include testing the latest MongoDB version (Mongo recommending using 3.X + versions only, not supporting anything in the 2.X version family)
2. Node
3. what else?
Plugfest demo
1. Amy pinged Jon to see if the video demo is available
2. Remaining stories to polish-up and merge into master (currently on a fork) have been transitioned to Corsair
BareMetal Regression Pipeline now created/monitored.
1. Plan is to monitor for a few weeks, should it then be a gate?
2. BareMetal OS install on real hardware currently runs every 2 hours on the nightly docker images. Will need to kick off BareMetal at same time as CI
3. Do we then continue to run BareMetal every 2 hours
4. Should this be part of the Master CI pipeline, if so then we would need a modification to the Merge Freeze tool to freeze on failure of BareMetal regression tests.
All SMI Services have been published.
1. What documentation is needed, what kind of communication is needed for the open source community?
Moved to a bi-weekly meeting series on Thursdays. Next meeting will be Thursday September 7.