Skip to content

Preserve existing pod annotations/labels to fix restarting Temporal deployments#986

Open
bmorton wants to merge 3 commits into
alexandrevilain:mainfrom
bmorton:fix-controller-preventing-restart
Open

Preserve existing pod annotations/labels to fix restarting Temporal deployments#986
bmorton wants to merge 3 commits into
alexandrevilain:mainfrom
bmorton:fix-controller-preventing-restart

Conversation

@bmorton
Copy link
Copy Markdown
Contributor

@bmorton bmorton commented May 7, 2026

This fixes #858 where a rollout restart would cause in the new pod being terminated and the old pod continuing to persist. Now, we merge in the annotations/labels that are added by Kubernetes for rollout restart such that it can complete the rollout.

I think this will end up being related to #746, but I don't have additional details there yet.

As a caveat, this was written with Copilot and Opus 4.6 under my supervision in VSCode, but I tested this in my homelab cluster already.

@bmorton bmorton changed the title Preserve existing pod annotations/labels Preserve existing pod annotations/labels to fix restarting Temporal deployments May 7, 2026
@bmorton
Copy link
Copy Markdown
Contributor Author

bmorton commented May 7, 2026

I see there are some linting and testing issues that I need to resolve, but not quite sure how yet. It looks like the linter is still on v1, which doesn't support the version of Go that we're building with here. I'm trying to understand how to fix this without taking on 300+ lint fixes in the same PR

@sonarqubecloud
Copy link
Copy Markdown

sonarqubecloud Bot commented May 7, 2026

@bmorton
Copy link
Copy Markdown
Contributor Author

bmorton commented May 7, 2026

There's some deeper stuff that needs to get fixed, but I was able to get the test passing and the lint fix applied without pulling on the Go version quite yet.

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.

Controller preventing rollout restart

1 participant