From cbe43334ee71151ad235cc9b72d8b462e08cee1e Mon Sep 17 00:00:00 2001 From: jochen Date: Sat, 21 Nov 2026 19:50:06 +0100 Subject: [PATCH] Design 47 H: H-A-300, an act that fails while a part is away is owed again --- 03-DESIGN/01-to-be/47-milestones/H-lead-answers.md | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/03-DESIGN/01-to-be/47-milestones/H-lead-answers.md b/03-DESIGN/01-to-be/47-milestones/H-lead-answers.md index db750d64..eb793f00 100644 --- a/03-DESIGN/01-to-be/47-milestones/H-lead-answers.md +++ b/03-DESIGN/01-to-be/47-milestones/H-lead-answers.md @@ -1745,3 +1745,12 @@ Where two preparers disagreed, the chosen one is named. not a source repository of the gate, so an hq pull request gets no delivery and no check from the mesh; it merges on hq's own checks run before every commit (ADR 0527), as the day's earlier hq pull requests did; whether hq should be a watched source with mesh/check is a question for the operator (it changes how they work). P3, P-failure-loud. +- **H-A-300 (an act that fails while a part is away is owed again, 19:50)** Met live: while the store was replaced + (19:41:04 to 19:41:32, a move the operator approved) the forge answered 401 to the mesh's admin account for a + minute; the forge module said "the account is refused" and stopped watching until a fresh token worked (19:42:11); + the gate's views owed on two merged pull requests failed in that minute and ended "could not be performed", and + nothing performs them again. Owed, H2: (1) an owed act whose failure is a part being away (401 with the store down, + no responders, connection reset) stays owed and is performed again with a back-off until a bound, not ended; (2) + the forge module names the cause it can read (the store down, the token refused) and never tells the operator to + create an account that exists; (3) one process mints the forge token at a time (a lease), the others reuse it; (4) + a bulk tool performs every owed act of one record again. P-failure-loud, P-NATS-native, H-A-296.