Skip to content

ci(track-binaries): refresh amd64 and arm64 pins together, with tests - #288

Merged
jeefy merged 1 commit into
projectbluefin:mainfrom
jeefy:ci/track-binaries-arm64
Sep 29, 2026
Merged

jeefy merged 1 commit into
projectbluefin:mainfrom
jeefy:ci/track-binaries-arm64

Conversation

@jeefy

@jeefy jeefy commented Sep 29, 2026

Copy link
Copy Markdown
Collaborator

Follow-up to #272 (binary tracker) and #279 (arm64 pins under arch == "aarch64" in k0s-bin.bst / kubeadm-bin.bst).

Finding: the tracker already refreshes the arm64 refs. BstComponent.pins() treats every url: line carrying the component's marker as a pin, including those under (?):, and _candidates() requires every pin's asset and checksum file. Nothing tested it, though, and the docstring and docs only described amd64. This PR pins that behaviour down. There is no logic change.

Changes

  • tests/unit/test_track_binaries.py: behavioural tests (mocked urlopen, no network) on a fixture with the real (?): arch == "aarch64" layout:
    • runc bump and k0s bump: both the amd64 and arm64 refs are rewritten, and both assets are downloaded and hashed
    • a release that lacks runc.arm64 is not a candidate
    • a missing arm64 checksum entry, or a missing (404) arm64 asset, fails the component (exit 1) and writes nothing
    • a no-op run leaves the files byte-identical
    • the real elements pin an arm64 source for every amd64 source, for each component
  • The docstring of track-binaries.py, ci-tooling.md and kubeadm-sysext.md now say all pinned architectures move together. ORAS stays amd64-only because it is CI tooling.

Evidence

  • Mutation check: a tracker that drops arm64 pins fails 11 of the new tests.
  • pytest: 579 passed, in the worktree and again in a detached worktree without files/boot-keys. bats: 74/74. docs-checks.py passes. Workflow unchanged, so no actionlint run.
  • Live check: kubernetes 1.34.3→1.34.12, containerd 2.1.5→2.1.9, runc 1.3.3→1.3.6, k0s 1.36.4+k0s.0→+k0s.1 pending; cri-tools, cni-plugins, oras current.
  • Live apply runc in a scratch copy, which was then discarded (no version bump committed). It wrote amd64 3f3921db…023c and arm64 6be27a06…b747. Both match sha256sum of the downloaded runc.amd64 / runc.arm64 and upstream's runc.sha256sum.

projectbluefin#279 added arm64 sources under `arch == "aarch64"` in k0s-bin.bst and
kubeadm-bin.bst after projectbluefin#272's tracker landed. The tracker already treats
every `url:` line carrying a component's marker as a pin, so it rewrites
the arm64 refs too, but nothing held it to that. Behavioural tests (mocked
HTTP) now do: a runc and a k0s bump rewrite both architectures' refs; a
release missing the arm64 asset is not a candidate; a missing arm64 asset
or checksum fails the component and writes nothing; a no-op run leaves
the files byte-identical; the real elements pin arm64 for every amd64
source. The docstring and the tracker docs say so.

Assisted-by: Claude Opus 5.5 via GitHub Copilot
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@jeefy
jeefy merged commit 212cd97 into projectbluefin:main Sep 29, 2026
11 of 12 checks passed
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