Last Week In Kubernetes Development: Week Ending September 16, 2018

lwkd
development

#1

You can also browse or subscribe to LWKD on its website.

Community Meeting Summary

This week’s community meeting began with an explanation and demo of the Foqal Slack bot that has been added to some of the Kubernetes community Slack channels. This bot helps provide automated assistance to users asking questions, to hopefully reduce the workload of common or repetitive issues from users. If you’re interested in trying out the bot on your channel, you can invite it with /invite @Foqal. Check out the recording for more details about setting up and using Foqal and you can contact @Vlad Shlosberg for any questions about it.

Tim Pepper had an update on the 1.12 release. Things are trending slightly better, but the release date is still very much at risk. Anyone that has some time to help out with CI signal would be greatly appreciated.

Michael Michael checked in from SIG-Windows. The Windows team has been making great progress but it does look like Windows support moving to GA is going to be delayed until 1.13. The team is looking to freeze features in the near future and spend the 1.13 cycle on conformance and perf.

Dawn Chen gave the update for SIG-Node for Q3. The initial SIG charter has been approved and merged. The SIG has been working on a bunch of new node features, including better pod sandboxing via the RuntimeClass feature, Windows container support (with SIG-Windows), and an improved e2e testing framework.

And finally we had an elections update from Paris Pittman. Nominations for the steering committee election are now closed, as is the voter eligibility process. Ballots will be emailed out on Wednesday the 19th. If you don’t get an email from CIVS by Thursday and are on the voter list, please contact community@kubernetes.io for assistance.

Release Schedule

Next Deadline: Doc Complete and RC1, September 18th

The last full week of code freeze is now behind us, PRs will reopen on Thursday. As mentioned above, concerns around stability and CI signal are at risk of slipping the release by a few days. If the release team needs your help with anything, please give them whatever time and energy you can so we can hopefully get 1.12 out uneventfully.

Featured PRs

It was a fairly quiet week for merges in general due to code freeze, and the release team focusing hard on stabilization. So just a few quick PRs to highlight here:

#68613: Core-dns pods crashing in large-cluster performance tests

Not actually a PR, but an issue to be aware of. It looks like the switch to Core-DNS as the default DNS provider is likely to slip to 1.13 if the stability concerns can’t be addressed. The situation is still very much in progress, so check the issue for the latest updates.

#67899: Graduate Schedule daemon set pods by default scheduler to beta

Great to see this get in for the 1.12 release, and will hopefully improve the stability and predictability of DaemonSets.

#68642: remove submit-queue from readme

I mostly raise this one just to remind everyone that kubernetes/kubernetes has been switched over to Tide for merge management. So far it seems to be going great, an awesome job well done by the whole Kubernetes community and the Tide and Release teams in particular.

#68431: Update cAdvisor godeps to v0.31.0

And finally a nice version update for 1.12, bumping CAdvisor and upgrading the interfaces where they changed.