Skip to content

ci(lab): quarantine the apr-stub race test to guards-nightly (PMAT-5873) - #5874

Closed
noahgift wants to merge 2 commits into
mainfrom
lab/apr-stub-race
Closed

noahgift wants to merge 2 commits into
mainfrom
lab/apr-stub-race

Conversation

@noahgift

Copy link
Copy Markdown
Member

Refs #5873. Noah 15:04Z item 4 (verbatim via arbiter-cop): "apr stub" race: quarantine the test to nightly LAB, then fix. Never delete.

  • check_harness_server_cleanup.sh (ci/explicit-test-commands.d/817) moved to guards-nightly.yml + guards_nightly_manifest.txt (manifest check green). One commit; LAB owner: aprender-w5382.
  • Root-cause fix follows in a second PR. Gate-content (workflows, ci/): needs quorum 3/3 + 2 non-author seats; not armed.

🤖 Generated with Claude Code

… the merge path to guards-nightly (PMAT-5873)

Moved, not deleted: 817 leaves ci/explicit-test-commands.d and runs nightly.
Refs #5873

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

@noahgift noahgift left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seat (aprender-triage-worker, Sonnet) @2c47dc47f6: REQUEST-CHANGES, one blocker. scripts/check_roadmap_fragment_required.sh exits 1 on this head: 'DRIFT: docs/roadmaps/roadmap.yaml is NOT aggregate(entries/)'. The PMAT-5873 fragment was added without make roadmap-aggregate; guard-tree runs that guard, so the required job will go red. Fix: run make roadmap-aggregate and commit docs/roadmaps/roadmap.yaml. The GATE->LAB move itself is clean (817 removed, guard-tree skips the guard as wired-elsewhere, no continue-on-error, nightly is schedule/dispatch only; manifest/explicit-test/wired/isolated guards rc 0). Non-blocking: add timeout-minutes (~15) to the nightly step, and consider id + if !cancelled() + a lab_ticket opener so a red does not skip later steps. Quorum: Opus (executed) RC, agy APPROVE (missed the drift). I will re-seat on the new head. Receipt handoff/triage-quorum-5874-2c47dc47f6.md. Not armed.

@noahgift noahgift left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seat (triage-worker, Sonnet 5.5) REQUEST-CHANGES @2c47dc47f6. Quorum AGREED 3/3 REQUEST-CHANGES on one mechanical blocker (Opus executed, agy, me): docs/roadmaps/roadmap.yaml not regenerated, check_roadmap_fragment_required rc=1 on the PR guard path. Fix = make roadmap-aggregate + commit roadmap.yaml (Opus verified the diff is exactly the 20-line PMAT-5873 entry). The GATE->LAB move is clean 3/3 (only invocation left is guards-nightly.yml, 817 deleted, guard_tree skips it, no continue-on-error, schedule-only). Non-blocking: no lab_ticket opener, add if: !cancelled() or move step last. I will re-seat APPROVE when only roadmap.yaml changes. Receipt: handoff/triage-quorum-5874-2c47dc47f6.md

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

@noahgift noahgift left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seat triage-worker APPROVE @6040e299b4. quorum 3/3 AGREED APPROVE (carried from the 2c47dc4 round; delta = roadmap.yaml +20 only, check_roadmap_fragment_required rc=0 executed on this head). Receipt: handoff/triage-quorum-5874-6040e299b4.md. Non-blocking: lab ticket opener, if: !cancelled().

@noahgift noahgift left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seat 2 (aprender-w5032, non-author): APPROVE @ 6040e29. No w5032 trailer on either commit. Read the diff: the 817 fragment is removed from ci/explicit-test-commands.d, and the same check is added to guards-nightly.yml and scripts/guards_nightly_manifest.txt under one identical step name. check_guards_nightly_manifest.sh at this head: rc 0, 14 manifest entries, every one a step in the workflow. The PMAT-5873 entry and roadmap.yaml carry the same 20 lines. No third state on the PR path: the check simply leaves it. Not run: the nightly itself, and the second PR that fixes the apr-stub race.

@noahgift

Copy link
Copy Markdown
Member Author

Gate-accept sign-off (arbiter-cop, non-author): APPROVE @ 6040e29. This moves check_harness_server_cleanup.sh from ci/explicit-test-commands.d into guards-nightly (LAB), as containment for the #5873 stub-server race. The root cause is runner thread overload; that fix is tracked separately. The quorum's earlier blocker was the roadmap regeneration. The delta at this head is roadmap.yaml only, check_roadmap_fragment_required gives rc=0, and github_issue: 5873 appears once. Seat 1: triage-worker. Seat 2: aprender-w5032. Per the operator's ruling of 16:28Z, this PR shares one merge-queue run with #5871 and #5869 as a one-time exception, and keeps its own commit. If that run goes red, the three are split.

@github-actions

Copy link
Copy Markdown

§13.11 rung 1 — quorum shadow verdict

S13-SHADOW pr=5874 head=6040e299b45991e318491e57bf1a3ef69a1ba175 verdict=REFUSE class=Q1 arm_rc=1

Shadow mode: this records a verdict and merges nothing. A refusal
to arm is not a block (§13 adds zero rows to §7) — the pull request is
exactly as green as it was.

@noahgift

Copy link
Copy Markdown
Member Author

superseded by #5927 (EMG-001 epic build-speed; this ticket's commit is in its range)

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