Week Ending March 17, 2019

Developer News

The project has created the #pr-reviews channel on Slack as a place for you to get help with finding a reviewer for your pull request.

Steering Committee Meeting

The Steering Committee had a public meeting last week. They discussed the audit of WGs and charters (WG-ML may be converted to a UG), and updated the SC charter. They will be contacting SIGs who need to update charters. The cloud providers will all be moved under SIG-Cloud-Provider.

They also did more to clarify SubProjects, where we need roles and contact information for each, and document User Groups, which have been added to sigs.yaml. We need more documentation around funding and resources from the CNCF and the CoC.

Release Schedule

NEXT DEADLINE: Docs complete March 18th. Code Thaw March 19th

Last week was Week 10. This means that RC1 is due out next week, and release the week after that. All enhancements should be done, and your docs are due now.

The Release Team is considering migrating to Go 1.12.1, which would delay Code Thaw and the release; watch kubernetes-dev for warnings.

Featured PRs

A bit of a light week thanks to code freeze, but a few PRs to look over.

community#3180: Scope of Kubernetes

Kubernetes defined. Or at least a formal declaration of what things belong in Kubernetes and which should be affiliated ecosystem projects. This specification has seen a lot of discussion from the community and will doubtless see more, but this is at least where we are today.

We also saw big upgrades in the community guidelines for both YouTube and Slack.

#75413: upgrade to go1.12.1

The upgrade to Go 1.12.1 appears to have stabilized several tests that were broken until 1.12.0. Be on the lookout for any lingering issues given the close proximity of the release.

test-infra#11755: Slack bot to report messages

A great addition to keep our Slack safer now that public registration is enabled again, all messages now have a context menu option to report them to the moderator team. If you’re having any issues in Slack and are unsure how to contact the moderators (you can find us in #slack-admins), please use this new tool.

We also added a welcome bot to help new users as they join our community!

#75416: Don’t follow non-local redirects from HTTP probes

The documentation for HTTP readiness and liveness probes has always said that any HTTP response with a code between 200 and 400 would be considered a success, but until now 301 and 302 redirects have been expanded by the kubelet and queried again. As a compromise between matching the spec and not overly breaking backwards compatibility, redirects to the same hostname as used originally will still be followed, but a redirect to a new hostname will be considered a success and not followed. This may result in a loss of accuracy in some very rare cases such as clustered applications redirecting to each other. But on the plus side it improves the security of Kubernetes and closes a potential traffic amplification vector.

Other Merges

Fix Windows SMB mount and unmount failures
Prevent Dockershim panic when removing images
Make Azure rules and LBs work together
Another shot at verifying static critical pods correctly; hopefully this version does it
Don’t let non-owners delete load balancers
Stop deletion of objects if ResourceVersion has changed since last seen
LoadBalancer Services with no Endpoints should hard-reject packets rather than dropping them

Deprecated

hyperkube’s --make-symlinks is deprecated

Version Updates

CSI to 1.1 (in 1.14)

Last Week In Kubernetes Development (LWKD) is a product of multiple contributors participating in Kubernetes SIG Contributor Experience. All original content is licensed Creative Commons Share-Alike, although linked content and images may be differently licensed. LWKD does collect some information on readers, see our privacy notice for details.

You may contribute to LWKD by submitting pull requests or issues on the LWKD github repo.