Core Committer Weekly Interlock - October 19th 2017

Owned by Amy Mullins
Last updated: Oct 19, 2017
3 min read

Attendees

Thomas Sullivan

Michael Hepfer

Former user (Deleted)

Former user (Deleted)

Amy Mullins

Former user (Deleted)

Leo Zhang

Bishoy youssef

Former user (Deleted)


Agenda

  1. RackHD Documentation A
    1. Per the Slack channel:  3lm0 [11:42 AM] 

    @maithri indeed, the blockers for now are the UI/UX for our end users which aren't technical one, and the consistency among the documentation leading to a difficult installation process with various pitfalls to be aware of plus the fact that as operator you have to deal with too many configuration and manual operations. 

     Michael Hepfer sent stories per his AI last week:

    https://rackhd.atlassian.net/browse/RAC-6240 - Documentation Updates (Mustang team backlog)

    https://rackhd.atlassian.net/browse/RAC-6241 - Inconsistency in code placement/organization (Leo Zhang to be assigned to the Maglev board)

  2. On-web-ui long term plan

    1. Maintain as-is until Kastalist UI is required.  but will it be open source?  No.  Plan to support on-web-ui as previously discussed (and listed below)
      1. Critical bug fix and API alignment only
        1. Former user (Deleted) to review list below to see if there are any centurylink issues that are already in the list.  Plan to create PRs for the known issues worked on with Centurylink.  
        2.  list of outstanding Jira bugs to determine if they are "critical" or can be set to p3, can the teams please take a look to see if these are still valid.
          1. Stories around CI can be closed , icebox'd , etc... 

          2. type key summary assignee reporter priority status resolution created updated due
            Refresh

      2. Continue to deploy as part of RackHD, but do not build up CI test capabilities (Jenkins or Concourse). 
  3. RackHD CI/CD update - 
    1. 10/27 is still the target date to bring the Concourse based PR quality gates and Post Merge online.  
      1. Dockerhub updates:
        1. devel will reflect the latest merged PR (always overwritten)
        2. latest will reflect the latest Sprint Release
        3. nightly will be deleted, covered by devel
        4. preserve the releases (2.5.0, 2.10.0, ....)
        5. remove everything else. 
          1. There was a concern for Traceability: we can recreate image via the digest of the built docker image that references the git commit hash.  If, in the rare case, an external contributor is looking for an engineering build, then they can make a request for it or rebuild it themselves.  If we find there is an excessive amount of requests for rebuilding engineering drops, we can explore how best to preserve them.  The effort to support this in our CI pipeline is believed to be greater than effort to recreate the image on the rare occasion.
          2. All the builds are also saved locally.   
      2. Debian updates: 
        1. NA, will plan to re-use what is already there.
    2. Sprint Release cadence will still be maintained on a weekly basis and will change from running virtually to running on physical HW (regression-baremetal job).
      1. Regression-baremetal quanta stack has few issues with the SEL tests, Dell stack is green.
      2. Can start looking in to creating the Concourse pipeline around the Regression-baremetal.  Plan to support testing of both stacks for Sprint release, expecting both stacks to pass to make the release.
      3. Until we're ready to move over to Concourse Regression-Baremetal for our sprint release, MasterCI will still be the release vehicle.  MAnual review of the Regression-Baremetal job will occur and will not hold up the Sprint Release unless a critical issue is escalated.
    3. Add post meeting: Regression-baremetal job will be set up to run Sun-Thurs nights to mitigate risk and minimize surprises for the Friday release.
      1. Epic created to track the Sprint Release/Regression test efforts: RAC-6326 - RackHD Sprint Release Regression-Baremetal Transition BACKLOG
  4. Review slides from Former user (Deleted) for CI Security
    1. AI: CC team to review the slides,  compare and discuss with the ongoing dialog leveraging Virtustream between Felouka and Veryon.
    2. Since this has been a longstanding agenda item, Amy Mullins will move to email and re-kick off the discussion again.
  5. Update on the Workflow Engine as a standalone service
    1. Will have an inception week of 11/3 and include stakeholders.  Multiple projects interested in consuming this service.