Issue 335: a build that passed its gate was failed and put back by the first machine's later reports #227

Merged

mesh-admin merged 3 commits from issues/335-a-passed-gate-was-failed-by-later-reports into main 2026-10-08 19:51:27 +00:00

Conversation 1 Commits 3 Files Changed 2

+118

mesh-admin commented 2026-10-08 18:27:44 +00:00

Contributor

Copy Link

Copy Source

On 2026-10-08 dunst's new build passed its gate on the laptop (healthy 3 times over 2m5s). Twenty-eight minutes later the same walk marked it failed at its gate, with those same words as the reason, and put it back.

• Cause: the controller judged the first machine's report again on every step of the walk, also after the gate had passed.
  • The send to the rest waited on another walk.
  • That walk then sent the laptop a declaration the laptop did not report on.
  • At the 30-minute bound for a first machine's report, the walk failed.
  • failFirstSend then put back a module whose gate had passed.
• Fix: merged in mesh-controller#165 (8170fc5).
  • A passed gate sends the rest.
  • A carried module takes over its lead's pass before its step is read.
  • A guard, not reached with those two changes, keeps a later path from putting back a passed build.
• Replay: R335 in mesh-lab#72, which points at 8170fc5 and is proved to fail on the merge's first parent and pass on the merge.
• Status: resolved, with fixed-by: mesh-controller#165, mesh-lab#72 (playbook 03, step 3).
• No design changes: ADR 0236 already says a pass is a verdict. The code did not follow it.
• Still open, not part of this issue: why the laptop's node-engine did not report on a send between 16:49 and 18:01 UTC.

merge-check.sh (records, index, cycle, words) passes locally.

On 2026-10-08 dunst's new build passed its gate on the laptop (healthy 3 times over 2m5s). Twenty-eight minutes later the same walk marked it failed at its gate, with those same words as the reason, and put it back. - **Cause:** the controller judged the first machine's report again on every step of the walk, also after the gate had passed. - The send to the rest waited on another walk. - That walk then sent the laptop a declaration the laptop did not report on. - At the 30-minute bound for a first machine's report, the walk failed. - `failFirstSend` then put back a module whose gate had passed. - **Fix:** merged in mesh-controller#165 (8170fc5). - A passed gate sends the rest. - A carried module takes over its lead's pass before its step is read. - A guard, not reached with those two changes, keeps a later path from putting back a passed build. - **Replay:** R335 in mesh-lab#72, which points at 8170fc5 and is proved to fail on the merge's first parent and pass on the merge. - **Status:** `resolved`, with `fixed-by: mesh-controller#165, mesh-lab#72` (playbook 03, step 3). - No design changes: ADR 0236 already says a pass is a verdict. The code did not follow it. - **Still open, not part of this issue:** why the laptop's node-engine did not report on a send between 16:49 and 18:01 UTC. `merge-check.sh` (records, index, cycle, words) passes locally.

mesh-admin added 1 commit 2026-10-08 18:27:44 +00:00

Issue 335: a build that passed its gate was failed and put back by the first machine's later reports ... b409537dbc

Recorded so the gate's verdict is known to be final, and the replay R335
holds it.

mesh-admin commented 2026-10-08 18:27:50 +00:00

Author

Contributor

Copy Link

Copy Source

Delivery novox/hq@bbbaf943e805 — delivered since 2026-10-08T19:51:54Z

Delivery plan — builds nothing: the change touches no module of the mesh's graph

Transitions

• 2026-10-08 19:49 (new) → proposed (announced): novox/hq#227's head announced
• 2026-10-08 19:49 proposed → checked (checked): the verdict names this commit
• 2026-10-08 19:49 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
• 2026-10-08 19:51 ready → published (merged): on the trunk its modules follow: merged there, and its walk opened — or nothing for a walk to move
• 2026-10-08 19:51 published → delivered (done): nothing for a walk to move

The commit's note under refs/notes/mesh-plan keeps every transition: git log --notes=mesh-plan.

<!-- mesh-delivery:view --> **Delivery** `novox/hq@bbbaf943e805` — **delivered** since 2026-10-08T19:51:54Z **Delivery plan** — builds nothing: the change touches no module of the mesh's graph **Transitions** - 2026-10-08 19:49 (new) → proposed (announced): novox/hq#227's head announced - 2026-10-08 19:49 proposed → checked (checked): the verdict names this commit - 2026-10-08 19:49 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail - 2026-10-08 19:51 ready → published (merged): on the trunk its modules follow: merged there, and its walk opened — or nothing for a walk to move - 2026-10-08 19:51 published → delivered (done): nothing for a walk to move The commit's note under `refs/notes/mesh-plan` keeps every transition: `git log --notes=mesh-plan`.

jschoubben added 1 commit 2026-10-08 18:49:42 +00:00

Issue 335: leave fixed-by for the merge, and call the put-back refusal a guard f346efb938

jschoubben added 1 commit 2026-10-08 19:49:04 +00:00

Issue 335: resolved by mesh-controller#165 (8170fc5) with replay R335 bbbaf943e8

mesh-admin merged commit 4955a663d3 into main 2026-10-08 19:51:27 +00:00

mesh-admin deleted branch issues/335-a-passed-gate-was-failed-by-later-reports 2026-10-08 19:51:28 +00:00

Sign in to join this conversation.

Reviewers

No Reviewers

Labels

Clear labels

mesh/ready

Reviewed and ready: mesh-delivery merges at the marked head once its checks pass (hq ADR 0364)

No labels

Milestone

No items

No Milestone

Assignees

Clear assignees

jschoubben mesh-tools miel.truyen

No Assignees

2 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: novox/hq#227