Core Committer Weekly Interlock - December 6th 2017

Attendees

Agenda

Is there a concept of merge freeze in CI/CD ?
1. If an issue has slipped through the quality gates, an immediate revert required. Plus a quality gate should be added to catch this "slip" in the future.
2. If there is an issue with the 3rd party services/libraries/software or Concourse Infrastructure/Tooling/HW, merge freeze is required and the communication mechanism is to be an automatic email to all RackHD committers (info pulled from github). Manual email is required if Concourse crashes.
RackHD metrics collection status
1. Bishoy youssef demo'd total failures/success releases in a single stat, will plan to send snapshots of the admin UI outside of the meeting
2. Thomas Sullivan to send the previous metrics that were previously reported on
RackHD static analysis
1. Currently using Hound jshint for static analysis. Running on every PR and PR merge.
2. Former RI-137 to be reviewed to determine what is applicable for RackHD Security
RackHD SLack channel post from Kenny:
1. Tim Larson, any thoughts on this?
  Maintainers… need to pose a question to you all. We got a few PRs last night in our RackHD repos and want to know if you all want to take ownership. At this time, we do not have the cycles to do anything other than put them in maintenance mode. Please respond back and I can make you admins for them
  https://github.com/codedellemc/docker-machine-rackhd
  https://github.com/codedellemc/gorackhd-redfish
  https://github.com/codedellemc/gorackhd
  https://github.com/codedellemc/rackhdcli

Follow-up

Bintray publication
1. limiting to 6 months of sprint releases and 1 week of per-merge builds. Stories listed below, currently a WIP.
2. Efforts include cleaning up what we have published today and maintaining that clean-up policy going forward.
  1. RAC-6536 - Getting issue details... STATUS
3. Before deleting the older sprint releases from binary, we should copy them to a local store in case someone comes looking (as has happened in the past).
  1. RAC-6537 - Getting issue details... STATUS
Concourse Based Sprint release status
1. December 1st Sprint Release was made (2.34.0)
Should Force merge to master be allowed in the new CD model
1. Team agreed this should be disabled. Former user (Deleted) updated all repos (including on-network, on-topology, smi, ...)
What is considered a 3rd party service? and what are testing / continuous testing requirements for these 3rd party services?
1. What is a core RackHD Service, what is a 3rd party service?
  1. Definition of 3rd party service: services developed outside of Dell EMC org. (ie mongodb)
  2. Definition of non-core services: services developed by Dell EMC orgs, not released with RackHD (ie UCS and SMI services).
    1. Continuous functional testing for these non-core services is not required by the RackHD CI/CD process. They should be tested by their "owners".
    2. Continuous integration test is required
      1. previous discussions had SMI integration testing with RackHD, currently a WIP migrating to Concourse - Michael Hepfer Was this completed?
      2. Needs to be done for UCS - Currently a WIP by the Maglev team
      3. No requirement at this time where the physical resources are for CI.
    3. Separate release pipeline is required
      1. SMI services currently has a release pipeline, to be migrated to Concourse - Michael Hepfer Was this completed?
      2. Needs to be done for UCS - Currently a WIP by the Maglev team