Repository navigation
Conversation
Two failures unrelated to the PR under test were seen recurring across
multiple unrelated runs on main and PRs:
- helm/kind-action's own installer downloads the kind binary/checksum
without curl's -f/--fail flag, so a transient HTTP error response is
saved as if it were the real file instead of triggering a retry.
sha256sum then fails with "no properly formatted checksum lines
found" or a checksum mismatch, aborting the job before any test
runs. Pre-populate kind-action's cache directory ourselves with a
properly -f/retry/checksum-verified binary so it skips its own
install step.
- velero/velero-plugin-for-aws:latest was left to be pulled live from
Docker Hub inside the kind node at pod start, with no retry, which
intermittently failed ("trying and failing to pull image") and made
BeforeSuite abort. Pre-pull it on the runner (with retry) and
`kind load docker-image` it before the suite starts. This is safe
because pkg/install/deployment.go sets ImagePullPolicy:
IfNotPresent for the plugin container whenever VELERO_IMAGE isn't
tagged "latest" (it never is in this workflow), so kubelet uses the
preloaded image instead of re-pulling.
A third flake (bitnami/containers reorganizing its MinIO Dockerfile
path) was found on release-1.18 but is already fixed on main via the
BITNAMI_CONTAINERS_COMMIT pin (8850862); no change needed here.
Signed-off-by: Tiger Kaovilai <tkaovila@redhat.com>
✅ Deploy Preview for velero canceled.
|
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
auto_label_prs.yml applies kind/changelog-not-required (and other exempt labels) from a separate workflow triggered by the same PR-open event. That workflow checks out the repo and diffs file globs, so it routinely takes a few seconds longer to land the label than this job's checkout-and-run-one-script path takes to fail. This was reproduced on PR velero-io#10555 itself: the label landed at the same second the changelog-check job completed and failed. Poll for the exempt label briefly (up to 30s) before concluding the changelog is missing. This only adds latency on the path that was already about to fail without it (no changelog file and, so far, no exempt label); PRs that already carry a changelog file or already have the label return immediately as before. Signed-off-by: Tiger Kaovilai <tkaovila@redhat.com>
ywk253100
reviewed
Sep 24, 2026
| run: | | ||
| docker run -d --rm -p 9000:9000 -e "MINIO_ROOT_USER=minio" -e "MINIO_ROOT_PASSWORD=minio123" -e "MINIO_DEFAULT_BUCKETS=bucket,additional-bucket" bitnami/minio:local | ||
| # kind-action's own kind installer (kind.sh's call_curl) downloads without | ||
| # curl's -f/--fail flag, so a transient HTTP error page from GitHub's |
Contributor
There was a problem hiding this comment.
Why wouldn't contribute the -f/--fail flag part to the upstream(helm/kind-action)?
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Note
Responses generated with Claude
Thank you for contributing to Velero!
Please add a summary of your change
Scanned recent
Run the E2E test on kindruns onmainand open PRs and found two CI-only flakes that were failing runs unrelated to the change under test, plus a third flake surfaced by this PR itself:helm/kind-action's own installer (kind.sh'scall_curl) downloads the kind binary/checksum without curl's-f/--failflag, so a transient HTTP error response from GitHub's release CDN gets saved as if it were the real file instead of triggering a retry.sha256sumthen fails withno properly formatted checksum lines foundor a checksum mismatch, aborting the job before any test runs. This repo pre-populates kind-action's cache directory itself with a properly-f/retry/checksum-verified binary, so kind-action's own (flaky) install step is skipped.velero/velero-plugin-for-aws:latestwas left to be pulled live from Docker Hub inside the kind node at pod start, with no retry, which intermittently failed withtrying and failing to pull imageand madeBeforeSuiteabort. This pre-pulls it on the runner (with retry) andkind load docker-images it before the suite starts. This is safe becausepkg/install/deployment.gosetsImagePullPolicy: IfNotPresentfor the plugin container wheneverVELERO_IMAGEisn't taggedlatest(it never is in this workflow), so kubelet uses the preloaded image instead of re-pulling.auto_label_prs.yml's file-glob labeler (.github/labeler.yml) auto-applieskind/changelog-not-requiredto PRs that only touch.github/**(like this one), but that workflow checks out the repo and diffs file globs, so it routinely lands the label a few seconds afterpr-changelog-check.yml's much faster job has already evaluated (and failed) the check — reproduced on this PR itself, where the label landed the same second the changelog-check job completed.hack/changelog-check.shnow polls briefly (up to 30s) for the exempt label before concluding the changelog is missing, only adding latency on the path that was already about to fail without it.A fourth flake (bitnami/containers reorganizing its MinIO Dockerfile path, hit on the
release-1.18build) is already fixed onmainvia theBITNAMI_CONTAINERS_COMMITpin (8850862e9), so no change is needed here for that one.No behavior change to the product itself — CI workflow/tooling only.
Does your change fix a particular issue?
N/A — CI flake fix, not tied to a tracked issue.
Please indicate you've done the following:
-s)./kind changelog-not-required— CI-only change, no user-facing behavior.🤖 Generated with Claude Code