Skip to content

[release-v1.42] Run kube-controllers on the host network during a datastore migration - #5319

Merged
caseydavenport merged 1 commit into
tigera:release-v1.42from
caseydavenport:casey-backport-13871-v142
Sep 16, 2026
Merged

caseydavenport merged 1 commit into
tigera:release-v1.42from
caseydavenport:casey-backport-13871-v142

Conversation

@caseydavenport

Copy link
Copy Markdown
Member

Description

Backport of projectcalico/calico#13871, a bug fix. Losing the kube-controllers pod while a datastore migration is in progress wedges the migration: the CNI plugin refuses requests while the migration holds the v1 datastore lock, and the controller that would release that lock runs in the pod that can no longer be scheduled. Recovery needs a manual patch today.

The deployment runs on the host network while a migration is active, so the pod never calls CNI and the controller's existing resume path picks the migration back up.

The monorepo commit applies here with the path prefix stripped; the only change is where the new config field sits in this branch's kube-controllers config. go test ./pkg/render/kubecontrollers/... passes.

AI assistance: Claude Code did the port and ran the tests. I reviewed it and am responsible for it.

Release Note

Fixes a datastore migration becoming permanently stuck if the calico-kube-controllers pod is lost while the migration is in progress.

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

@caseydavenport
caseydavenport merged commit 105917c into tigera:release-v1.42 Sep 16, 2026
5 checks passed
@danudey danudey added the kind/bug Something isn't working label Oct 27, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants