Skip to content

chore: run amazing-app-containers fixture pod as uid 101 - #121

Merged
alovh merged 1 commit into
mainfrom
fix/containers-fixture-uid
Jun 29, 2026
Merged

chore: run amazing-app-containers fixture pod as uid 101#121
alovh merged 1 commit into
mainfrom
fix/containers-fixture-uid

Conversation

@henrik242

@henrik242 henrik242 commented Jun 12, 2026

Copy link
Copy Markdown
Contributor

The common chart's containers-array path ignores uid in containers[] entries; the pod securityContext only reads container.uid (singular) and fell back to the default 1000, which getting-started:latest cannot run as. This made test-deploy-ok-gcp-containers crash-loop and fail CI on every PR since the container_name feature was released in 1.6.0 (the feature's own PR ran the pre-release helm-deploy@v1 action, so the test never actually exercised the replacement).

This fixes the root cause for the failing #120 (https://github.com/entur/gha-helm/actions/runs/27414274923/job/81022995404)

@henrik242
henrik242 requested a review from a team as a code owner June 12, 2026 13:48
@henrik242
henrik242 force-pushed the fix/containers-fixture-uid branch from 5ac9368 to 7699729 Compare June 15, 2026 07:22
@henrik242
henrik242 force-pushed the fix/containers-fixture-uid branch from 7699729 to 5b188d5 Compare June 16, 2026 10:24
The common chart's containers-array path ignores uid in containers[]
entries; the pod securityContext only reads container.uid (singular)
and fell back to the default 1000, which getting-started:latest cannot
run as. This made test-deploy-ok-gcp-containers crash-loop and fail CI
on every PR since the container_name feature was released in 1.6.0
(the feature's own PR ran the pre-release helm-deploy@v1 action, so
the test never actually exercised the replacement).

Also fix the rollback CI test, which broke once the containers test
passed and CI progressed to it. The test calls the helm-deploy
composite action directly, which bypasses deploy.yml's set-values-file
step, so the values input stayed at its literal default "values" and
the bad-image deploy failed on a missing env/values file instead of on
the bad image. No failed revision was created, so the rollback was
non-deterministic (status 2 vs 3) depending on leftover revisions in
the shared amazing-app release. Resolve the dev values file explicitly
so the deploy fails on the bad image and rolls back to the healthy
previous revision (status 2). Assert status 2 (was a stale status 3
left over from the deleted helm-deploy@negative_test action branch),
restore the test-deploy-gcp-with-rollback job name, and make cleanup
idempotent with --ignore-not-found.
@alovh
alovh merged commit 912edd8 into main Jun 29, 2026
45 checks passed
@alovh
alovh deleted the fix/containers-fixture-uid branch June 29, 2026 13:26
@henrik242

Copy link
Copy Markdown
Contributor Author

@alovh 🙇🏻

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.

2 participants