Closed blackliner closed 5 months ago
Hello! Thank you for filing an issue.
The maintainers will triage your issue shortly.
In the meantime, please take a look at the troubleshooting guide for bug reports.
If this is a feature request, please review our contribution guidelines.
Hey @blackliner
Can you please write down exact steps how you updated the secret? Did you update it using helm, or by changing it directly?
The secret was created manually, not hardcoded in the helm chart (gitops). Then the secret (the object iteslf) was updated to reflect the new PAT.
We did not change to a new secret, the existing one was updated.
Hey @blackliner,
Changing secret in-place cannot trigger the rollout. We are not watching the secret, and we don't hash its content in order to check for changes. The safest way for you to roll out a new secret is to create a new one, modify the values.yaml file to point to a new secret, update ARC installation and then remove the old secret once it is ready.
Closing this one, please let me know if you need more information :relaxed:
Checks
Controller Version
Chart: 0.9.1 Controller image: ghcr.io/actions/gha-runner-scale-set-controller:0.9.1
Deployment Method
Helm
Checks
To Reproduce
Describe the bug
When changing the secret that is used in the scalesets
githubConfigSecret
, the in-between configs that got generated by the operator do not get updated, and the listener pod keeps failing.Describe the expected behavior
When the content of a relevant secret is changed, the operator should also regenerate the in-between configmaps and secrets that contain aformentioned secret, and recycle the relevant (listener) PODs
Additional Context
Controller Logs
Runner Pod Logs