Skip to content

chore(release): prepare v1.41.2 and kates-operator 0.3.1 - #684

Merged
bmscomp merged 1 commit into
mainfrom
chore/release-v1.41.2
Nov 16, 2026
Merged

bmscomp merged 1 commit into
mainfrom
chore/release-v1.41.2

Conversation

@bmscomp

@bmscomp bmscomp commented Nov 16, 2026

Copy link
Copy Markdown
Owner

Why

Ships #683. With v1.41.1, kates deploy on kind ran a stale local kates:native-local built before the quorum. The operator then reported NoLeader indefinitely, and deploy waited out 10 minutes. #683 fixes the CLI and the operator, so both need a release:

  • kates-operator 0.3.1: NoQuorum for a backend older than the quorum, and a roll that replaces a stale member running no quorum first.
  • v1.41.2: the CLI runs the published backend from published charts, and stops at once on NoQuorum. Its ChartVersions pin kates-operator 0.3.1.

What changes

  • Every pin moves to 1.41.2: the kates chart's appVersion and image tags (0.20.3), kates/k8s/deployment.yaml, the kates-tester pins (strimzi-operator 0.5.6, kafka-cluster 1.1.6, connect-cluster 2.2.6, mirror-maker2 0.12.6, kates-chaos 2.6.3), the Litmus experiments, server.json, the registry docs, the runbook and recipes, appendix D, and kates-platform 0.10.3.
  • operator/chart moves to 0.3.1, along with the README's install, verify and pull commands. The CronJob example pins kates-tester 1.41.2.
  • The backend and tester images are unchanged since 1.41.0; nothing under kates/src or tester/ moved.

Release order

  1. Squash-merge this PR.
  2. Tag operator-v0.3.1 on the merge commit, then wait for publish-operator.yml to push the image and chart, signed and attested.
  3. Tag v1.41.2 on the same commit. The CLI pins kates-operator 0.3.1, which has to be published before deploy pulls it.

Tests

  • Local: scripts/check-versions.sh, python3 scripts/chart_bumps.py ("every changed chart carries a new version"), scripts/gen-version-matrix.sh --check, scripts/check-book-style.sh, gen-book-index.py --check, and yamllint on the Litmus experiments and chart values.
  • helm lint on all 16 charts plus operator/chart, with dependencies built in a scratch copy.
  • The pin script asserts each file's count. The counts match chore(release): prepare v1.41.1 #682's: 12 Litmus pins, 5 in docs/registry/kates.md, 3 in server.json.

Moves every pin to 1.41.2, the tag this release publishes: the kates
chart's appVersion and image tags (0.20.3), kates/k8s/deployment.yaml,
the kates-tester pins in strimzi-operator 0.5.6, kafka-cluster 1.1.6,
connect-cluster 2.2.6, mirror-maker2 0.12.6 and kates-chaos 2.6.3, the
Litmus experiment pins, server.json, the registry docs, the runbook and
recipes, the version matrix in appendix D, and kates-platform 0.10.3.

kates-operator 0.3.1 (chart and appVersion, and the README's install,
verify and pull commands) carries #683's operator side: NoQuorum for a
backend older than the quorum, and a roll that replaces a stale member
running no quorum first. The CronJob example pins kates-tester 1.41.2.

The release ships #683: on kind, a release's published charts run the
published backend instead of a stale local kates:native-local, and
deploy stops at once on NoQuorum. Nothing under kates/src or tester/
moved since v1.41.0.
@bmscomp
bmscomp merged commit f566a19 into main Nov 16, 2026
75 checks passed
bmscomp added a commit that referenced this pull request Nov 16, 2026
…e copied (#685)

* fix(scripts): wait for KafkaUser Secrets to settle before they are copied

Strimzi's User Operator reads a user's Secret from its informer cache. The
reconcile its own status update triggers can run before the cache holds the
Secret the first reconcile created: it issues a new password, sets it in
Kafka and overwrites the Secret (create, 409, update). A copy taken at the
first Ready, as e2e-stack.sh and integration.yml do, can hold the password
that was just replaced, and every Kafka call then fails authentication.

That failed CI · Operator twice on main at f566a19, the tree #684's PR run
passed with: both runs copied within a second of Ready. After the Ready
wait, deploy-kafka-generic.sh now waits until the user Secrets' resource
versions hold for 10 s (60 s at most).

* fix(cli): kates deploy copies KafkaUser Secrets only once they hold

The kafka-users wait now ends with the same settle as
deploy-kafka-generic.sh: the KafkaUser Secrets' resource versions must hold
for 10 s (a minute at most). Kates, Connect, Kafka UI and MirrorMaker 2
copy their Secrets after it, so none copies a password the User Operator
replaces a moment later.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant