/sig release

Is this part of the release process documented somewhere?

cc @spiffxp
I was referencing #53004 and made one for 1.10

There's two CRI changes which break backwards compatibility for some users:

#59475

#58973

adding kind

/kind design

Adjusting labels to keep this tracker in the milestone.

/kind cleanup
/remove-kind design
/priority critical-urgent
/remove-priority important-soon

(seems this issue is not as useful and not documented as part of release process, I'll just close it, but feel free to reopen if necessary)

#60764
Known issue, downgrade extra steps with PVC

Clarification, please? Do y'all want a relnote for #60933?

Add #60764 plus related doc kubernetes/website#7731

Mount propagation manual steps: #61126

Clarification, please? Do y'all want a relnote for #60933?

My opinion: yes.

ACK. In progress
ETA: 29/03/2018
Risks: The bot nagging will drive us slowly mad

From release burndown as of March 19:

Flaky timeouts while waiting for RC pods to be running in density test
Controller-manager sees higher mem-usage when load test runs before density
HostPath mounts failing with "Path is not a shared or slave mount"
kubeadm: etcd certs missing in self-hosted deploy (will be fixed in point release)

@nickchase in case it's helpful here: think you don't need to worry about #61126 bc the relnote content is in the PR (such as it is ...). BUT it might not be in the right place in the generated relnotes. It's an action required thing.

We need to doc downgrading and PVC protection. TL;DR: if you have PVCs, and need to downgrade, you need to downgrade to 1.9.6, which will be released next Wednesday, and not to an earlier version of 1.9.

Yup. @nickchase see also comment from liggitt in kubernetes/website#7731

#61446
#61456

@nickchase let's coordinate with @msau42 about relnotes/docs. Slack discussion also.

We also need to add those 2 known issues to 1.9, 1.8, 1.7 patch releases as well.

The workaround for #61446 and #61456 is to not use subPath with hostPath volumes, and instead just specify the whole path in the hostPath volume.

Adding:

• Use of subPath module with hostPath volumes can cause issues during reconstruction (#61446) and with containerized kubelets (#61456). The workaround for this issue is to specify the complete path in the hostPath volume.

Flaky timeouts while waiting for RC pods to be running in density test -- Appears to be fixed?
Controller-manager sees higher mem-usage when load test runs before density -- Does this need a Relnote?
HostPath mounts failing with "Path is not a shared or slave mount" -- Resolved with addition to "Before Upgrading" section
kubeadm: etcd certs missing in self-hosted deploy (will be fixed in point release) -- Added

Added * In large clusters (~2K nodes) scheduling logs can explode and crash the master node. (#60933)

@nickchase

Yes on Controller-manager. Suggested text:

"Some users, especially those with very large clusters, may see higher memory usage by the kube-controller-manager in 1.10."

[MILESTONENOTIFIER] Milestone Issue: Up-to-date for process

@krzyzacy

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

/close