Core Commiter Weekly Interlock - Feb 13th,2017
0) 1.2.0 Release:
Anomaly: time change There was some manual testing to expedite.
QRB gives the thumbs up.
QRB looking into why nightly build is red.
AI: TagBasedManifest job needs to be run to set the 1.2.0 official branch, push up the artifacts. Former user (Deleted) volunteered.
1) Semantic Version. When we deprecate 1.1 API do we up the major version to 2.0? How do we notify community of this deprecation? Who will do it?
Yes, we should bump to 2.0 API. AI: write up in confluence the definition of what is our public API
Added warning to the API header. Former user (Deleted): should we also change the status code? Former user (Deleted) That could be a breaking change to the clients. That is why we decided to go with the header warning.
Also announce on the groups, slack, and possibly the portal page: AI: Former user (Deleted)
The 1.1 API will be deprecated at the end of the sprint ending on 3/10.
2) https://rackhd.atlassian.net/browse/RAC-4167 - problem closed out not an issue
Former user (Deleted) Counter-intuitive behavior because it is "anyOf". We should consider supporting "allOf" as well. Proposed workaround to add another noop terminal task that waits on the parallel tasks.
Current behavior is not documented. AI: Create a story to update the documentation.
Proposal to add "allOf" behavior to the workflow engine. What is the next step? Create a spike to draft a high-level proposal in confluence and review with the community. AI: Former user (Deleted) will talk to Amy Mullins. Will communicate with Paul Scharlach regarding related issues?
3) Workflow annotation - We have been working with community members to update FW on the dell r730xd. information would be extremely useful on how-to use or string together multiple workflows/tasks in order to perform firmware update in addition to providing info on known limitations / workarounds. Devs teams need to make sure this information is available to a broader audience. For example: Have documentation for quanta in the readme (https://github.com/RackHD/on-skupack/tree/master/quanta-d51-2u) but missing for dell r730xd (https://github.com/RackHD/on-skupack/tree/master/dell-r730xd).
Amy Mullins: How do we best document what we have available for workflows? Tips and tricks to get workflows working. Move some of these thing from internal site to external.
AI: Open a story to move some of this information to an external workflow. Amy Mullins
AI: Story/RI How much of this can be automated? swagger-ish Amy Mullins
About 1.5 years ago, this effort began but was ultimately never completed, tied up in details for how-to programmatically collect this information. What we are finding is that RackHD consumers would find this very useful.
4. Revisit release deployment strategy. Are we ready for continuous deployment? Discussions at the Dev Con had a long term vision of continuous deployment. Short term it was decided not to move in this direction due to time and resources. Resources have more recently been made available to support test/package/deployment per PR(PP: should it be per-commit ? )
AM: No, what will be presented during the meeting not be per commit as discussions with QRB, RackHD devs, and dojo team members progressed over the course of the week.
Amy Mullins will send out slides. CC's look these over. Any comments on slack channel. Add to the top next weeks agenda.
Updated slide from Peter
5. REQUEST to grant Peter Pan (experts in CI/CD domain) CC like role (as cc +1 and merge privilege) to on-build-config repo in order to increase development efficiency.
Currently, merges being done by other CCs on request.
Amy Mullins What are the long-term plans for on-build-config? Former user (Deleted) Yes, this repo includes script other than the pipeline code
Former user (Deleted) Andre should be involved in the discussions.
Leo Zhang In the short term, the need is the still valid. Needs domain expert to have rights to merge.
AI: meeting to get alignment on the CD/CI strategy. Tee this up at the OLT. Please review these slides before this meeting.
6- RackHD defects- Mustang team has compiled list of defects , get a broader consensus on these defects for resolution.
Former user (Deleted): One defect during discovery regarding redundant enclosure/chassis data. There is a JIRA ticket open. Former user (Deleted) has started to take a look. There is definitely an issue here.
Tim Larson Each sprint end, dispatch defects to the team backlogs. Autogenerate email with what is in the defect queue, rankings, etc.
AI: Tim Larson, Thomas Sullivan, dev managers: work on a clear process to triage the single backlogs
7- Node attributes task. Trying to come to conclusion on older github issue and outstanding work in the backlog (Paul S)
Paul Scharlach and Former user (Deleted) to take this offline.