From 411ef9e36d89b975922e1217ef10604a16c0f9e6 Mon Sep 17 00:00:00 2001 From: kisfenyo Date: Thu, 17 Sep 2026 11:56:03 +0200 Subject: [PATCH] operator decisions: floor raised to 0.246.0 (delivered), agent vouch refused by R-120 Decision 1: global controller floor 0.244.0 -> 0.246.0 with declared MinAgent 0.131.0. Served for demo-felhom; the N100 ran 0.246.0 within seconds; demo-hp already did. Peti's box is DOWN on the hub and cannot receive it until it reports. Decision 2: vouching agent 0.132.0 was refused by the hub's R-120 gate - the vouched golden 0.245.0 is older than the newest controller the fleet reports (0.246.0). Nothing stored. A golden >= 0.246.0 is needed first; put to the operator. Co-Authored-By: Claude Opus 5 Claude-Session: https://claude.ai/code/session_0159rPz1ZhFKsS53msqPYxtS --- .../decision1-floor-0246.txt | 21 +++++++++++++++++++ .../decision2-vouch-agent-0132.txt | 11 ++++++++++ 2 files changed, 32 insertions(+) create mode 100644 documentation/audits/evidence-chaos-fixes-2026-09-17/decision1-floor-0246.txt create mode 100644 documentation/audits/evidence-chaos-fixes-2026-09-17/decision2-vouch-agent-0132.txt diff --git a/documentation/audits/evidence-chaos-fixes-2026-09-17/decision1-floor-0246.txt b/documentation/audits/evidence-chaos-fixes-2026-09-17/decision1-floor-0246.txt new file mode 100644 index 00000000..80cdb4c3 --- /dev/null +++ b/documentation/audits/evidence-chaos-fixes-2026-09-17/decision1-floor-0246.txt @@ -0,0 +1,21 @@ +# Operator decision 1 (2026-09-17, "yes"): raise the global controller floor to 0.246.0 +before: global floor min_controller_version=0.244.0 min_agent=0.131.0; vouched golden 0.245.0 +2026-09-17T09:53:22Z POST /configuration/global-floor min_controller_version=0.246.0 min_agent=0.131.0 + HTTP/1.1 303 See Other + Location: /configuration?flash=floor_set +--- hub log (the served/held verdicts) --- + 2026/09/17 11:53:22 [INFO] Global controller-version floor set to "0.246.0" (declared MinAgent "0.131.0") + 2026/09/17 11:53:24 [INFO] managed floor SERVED for demo-felhom: floor 0.246.0, agent requirement "0.131.0" from declared (golden 0.245.0) +--- 2026-09-17T09:55:05Z the boxes themselves --- + N100 (demo-felhom) guest 9201: gitea.dooplex.hu/admin/felhom-controller:0.246.0 Up About a minute (healthy) + demo-hp guest 9201: gitea.dooplex.hu/admin/felhom-controller:0.246.0 Up 32 minutes (healthy) + N100 SetFloor line: 2026/09/17 09:53:30 [INFO] [offsite-apply] settle-gate: awaiting floor knowledge (first report ACK) before offsite apply 2026/09/17 09:53:40 [INFO] [offsite-apply] settle-gate: GO — at/above floor 0.246.0 (we are 0.246.0), no managed update running +--- hub verdicts for every customer since the floor was set --- + 1 [INFO] managed floor SERVED for demo-felhom: floor 0.246.0, agent requirement "0.131.0" from declared (golden 0.245.0) + +## Fleet view from the hub (/configs, 09:55Z) + demo-felhom OK controller 0.246.0 floor 0.246.0 (global) + demo-hp OK controller 0.246.0 floor 0.243.0 (per-customer OVERRIDE; runs 0.246.0 from the B.4 deploy) + peti-felhom DOWN controller 0.115.0 floor 0.246.0 (global) — not reporting, so the floor cannot reach it + tester-1 DOWN controller 0.245.0 (tonight's destroyed box; no host) + drill-r50 BLOCKED diff --git a/documentation/audits/evidence-chaos-fixes-2026-09-17/decision2-vouch-agent-0132.txt b/documentation/audits/evidence-chaos-fixes-2026-09-17/decision2-vouch-agent-0132.txt new file mode 100644 index 00000000..35defc0b --- /dev/null +++ b/documentation/audits/evidence-chaos-fixes-2026-09-17/decision2-vouch-agent-0132.txt @@ -0,0 +1,11 @@ +# Operator decision 2 (2026-09-17, "yes"): vouch agent 0.132.0 for Day-0 installs +before: agent 0.131.0 (1118b552…), golden 0.245.0 (7a08aa1a…), min_agent 0.131.0, wrapper 104db0a4… +2026-09-17T09:55:22Z POST /configuration/artifacts agent_version=0.132.0 golden_version=0.245.0 min_agent=0.131.0 wrapper_sha256= + HTTP/1.1 303 See Other + Location: /configuration?flash=golden_behind_fleet + hub log: 2026/09/17 11:55:41 [WARN] artifact vouch REFUSED: golden 0.245.0 is older than the newest controller the fleet reports (0.246.0) — a fresh install would land on stale application code (R-120) + +RESULT: REFUSED by the hub's R-120 gate, as the handler's source predicted (hub/internal/web/configs.go +handleSetArtifacts: the golden may not be older than the newest controller any box reports). Nothing was +stored — the check runs before SetArtifactManifest. The Day-0 manifest is still agent 0.131.0 / golden 0.245.0. +Vouching agent 0.132.0 therefore needs a golden >= 0.246.0 first. Put to the operator.