Skip to content

ci(e2e): run the VM jobs on ubuntu-24.04 - #1131

Merged
Danathar merged 1 commit into
mainfrom
ci/pin-ubuntu-2404-vm
Nov 10, 2026
Merged

Danathar merged 1 commit into
mainfrom
ci/pin-ubuntu-2404-vm

Conversation

@Danathar

Copy link
Copy Markdown
Contributor

Summary: every testsuite VM boot has failed since GitHub switched ubuntu-latest to Ubuntu 26.04 on 2026-11-09. On 26.04, bootc's loopback install stops with "root partition 3 has type " before it writes anything; the same image installs on 24.04. That has kept v1 frozen since 2026-11-09 09:41Z (#1123) and turns the startup checks red on every PR that runs them (#1127). This PR runs the two VM jobs on ubuntu-24.04, as iso-validation.yml already does. This PR's own startup checks are the test.

Evidence

Run Runner Image Install
42479593163 (gate, 11-08 12:18Z) ubuntu-24.04 sha256:c10a8b18… ✅ deployed, v1 moved
42594102294 (gate, 11-09 09:31Z, first after the switch) ubuntu-26.04 sha256:dc3f6aee… ❌ root partition 3 has type
42908422162 (#1130, 11-10 21:47Z) ubuntu-26.04 sha256:c10a8b18… (the 11-08 image) ❌ same error

The last row settles it: the image that installed on 24.04 fails on 26.04. Also ruled out: a timing race (#1125's 3 retries fail identically every time) and a missing /run/udev in the install container (#1129, no change).

What changes

  • e2e.yml: the test job (the one that boots the VM) runs on ubuntu-24.04. The matrix and compose jobs stay on ubuntu-latest.
  • startup-qualification.yml: runs on ubuntu-24.04.
  • docs/skills/ci-ops/e2e-workflow/references/troubleshooting.md: the "root partition 3" entry now gives the real cause.

Each pinned line has a comment saying why, and to lift the pin once the install works on 26.04. That fix is left for a follow-up issue.

Merging

e2e.yml is a human-merge path, so a maintainer merges this once the startup checks are green. After merge, the next v1 gate run should boot again and can move v1. Image repos that call e2e.yml get the fix when Renovate bumps their v1 pin.

Checks run locally

actionlint on both workflows: same findings as main (nothing new).

Fixes #1123

🤖 Generated with Claude Code

ubuntu-latest became ubuntu-26.04 on 2026-11-09. On 26.04 bootc's
loopback install stops with "root partition 3 has type <none>" before
writing anything, so every VM boot fails: the v1 gate has held v1 since
2026-11-09 09:41Z and the startup checks fail on every PR that runs
them (#1123). The same image (sha256:c10a8b18) installed on 24.04 on
11-08 (run 42479593163) and fails on 26.04 today (run 42908422162).

Run the e2e test job and the startup lane on ubuntu-24.04, as
iso-validation.yml already does, until the install works on 26.04.
Correct the troubleshooting entry, which called this a timing race.

Fixes #1123

Assisted-by: Claude Opus 5.5 via Claude Code
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Danathar <Danathar@users.noreply.github.com>
@hivecommons-hive hivecommons-hive Bot added author/clanker A contributor's own hive authored this PR under their login dco All commits are DCO signed-off role/trusted Author is trusted for unattended merge (trusted-author role or registry/claim tier in require_tiers) stage/050-review-pending Review swarm requested; awaiting a verdict for the current head labels Nov 10, 2026
@codecov-commenter

Copy link
Copy Markdown

⚠️ Please install the 'codecov app svg image' to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

@Danathar
Danathar merged commit 92c6b56 into main Nov 10, 2026
9 checks passed
hivecommons-hive Bot pushed a commit that referenced this pull request Nov 10, 2026
…ns (#1133)

- Third-party GitHub Actions automerge only once their release is three
  days old, the same decision as finpilot#645. First-party
  projectbluefin/** actions are not delayed.
- Action bumps in the three release workflows (e2e.yml, build-runner.yml,
  build-kde-runner.yml) get their own PR. The hive's pin-only exception
  only applies when every file in a PR is a release workflow, so while
  these bumps were grouped with other workflows (#1127) they always
  waited for a person.
- Keep the VM jobs on Ubuntu 24.04: no major runner bump (#1131).
- Correct the first rule's description: it said digest, patch and minor
  updates need human review, but the inherited org rules automerge them,
  and the hive's reviewer cited that line when it parked #1120.

Assisted-by: Claude Opus 5.5 via Claude Code

Signed-off-by: Danathar <Danathar@users.noreply.github.com>
Co-authored-by: Danathar <Danathar@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
hivecommons-hive Bot pushed a commit that referenced this pull request Nov 11, 2026
* fix(gnome-e2e): check out testsuite from the right repository

The action read github.action_repository inside its checkout step,
where it names actions/checkout, so the GNOME OS canary fetched
testsuite's commit from actions/checkout and failed with "not our ref"
(run 42919533096). The extension-validation "script missing" error that
followed came from that empty checkout. Resolve the repository and ref
in a run step first, falling back to the caller's repository for a
local ./ call, and add a unit test that keeps the checkout step from
reading them directly.

Also run the canary on ubuntu-24.04: the action does a bootc loopback
install, which fails on 26.04 runners (#1123, #1131).

Assisted-by: Claude Opus 5.5 via Claude Code
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WcWm7qQa538Zjcuvv8jBtn
Signed-off-by: Danathar <Danathar@users.noreply.github.com>

* fix(gnome-e2e): run bootc install with --pid=host

With the checkout fixed, the canary reached the install and bootc
refused: "This command must be run with the podman --pid=host flag"
(run 42923546804). e2e.yml's install step already passes --pid=host.

Assisted-by: Claude Opus 5.5 via Claude Code
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WcWm7qQa538Zjcuvv8jBtn
Signed-off-by: Danathar <Danathar@users.noreply.github.com>

* fix(gnome-e2e): reach the BLS fallback when no loader entries exist

After bootc's expected late bootloader failure (bootupd) the ESP has no
loader/entries directory. find then exits non-zero, and under the
step's pipefail the command substitution ended the step silently right
after the guest preflight, before the fallback that builds the kernel
arguments without a BLS entry (canary run 42929881520).

Assisted-by: Claude Opus 5.5 via Claude Code
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WcWm7qQa538Zjcuvv8jBtn
Signed-off-by: Danathar <Danathar@users.noreply.github.com>

* fix(gnome-e2e): boot the deployment's real ostree entry without BLS

With no BLS entry the action guessed ostree=/ostree/boot.1/default/<hash>/0,
which does not exist on the installed disk, so ostree-prepare-root failed
("Couldn't find specified OSTree root") and SSH never came up (canary run
42932368882). Find the deployment's boot entry the way e2e.yml and
startup-qualification.yml do: use the existing boot.N/default/<hash>/0
link (mapping a versioned boot.N.M to boot.N for ostree-system-generator)
or create boot.0/default/<hash>/0 pointing at the deployment.

Assisted-by: Claude Opus 5.5 via Claude Code
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WcWm7qQa538Zjcuvv8jBtn
Signed-off-by: Danathar <Danathar@users.noreply.github.com>

---------

Signed-off-by: Danathar <Danathar@users.noreply.github.com>
Co-authored-by: Danathar <Danathar@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

author/clanker A contributor's own hive authored this PR under their login dco All commits are DCO signed-off role/trusted Author is trusted for unattended merge (trusted-author role or registry/claim tier in require_tiers) stage/050-review-pending Review swarm requested; awaiting a verdict for the current head

Projects

None yet

Development

Successfully merging this pull request may close these issues.

v1 tag held by the post-merge VM gate

2 participants