fix(recovery): surface 110 local package blocker
Some checks failed
CD Pipeline / workflow-shape (push) Has been cancelled
CD Pipeline / cancel-stale-cd (push) Has been cancelled
CD Pipeline / tests (push) Has been cancelled
CD Pipeline / build-and-deploy (push) Has been cancelled
CD Pipeline / post-deploy-checks (push) Has been cancelled
AI 技術雷達監控 / ai-technology-watch (push) Successful in 34s

This commit is contained in:
Your Name
2026-07-01 13:58:22 +08:00
parent d7ae71debc
commit e580954e82
3 changed files with 24 additions and 0 deletions

View File

@@ -27,6 +27,22 @@
**邊界**:只改 metadata-only LOG tagging contract、AI Loop current blocker evidence 投影、tests 與 LOGBOOK未使用 GitHub / `gh` / GitHub API未 workflow_dispatch未讀 secret / token / `.env` / raw sessions / SQLite / auth未讀 authorized_keys 內容或 `.runner` 內容;未執行 KM/RAG/PlayBook/MCP live write未執行 110 runtime apply。
## 2026-07-01 — 14:50 110 control path / recovery package blocker readback
**照主線處理的問題**
- 使用最新 Gitea main 後重跑 110 bounded diagnosis`NODE_LOAD1=7.69``NODE_LOAD5=13.03``NODE_LOAD1_PER_CPU=0.64``NODE_LOAD_CLASSIFIER=load_not_high`,表示 110 CPU/load 已從高壓狀態下降;但 `wooo` publickey 仍是 `publickey_offer_timeout`legacy runner units 仍回 `systemctl_show_timeout`,所以問題已收斂為 110 SSH/control path blocker不是繼續泛稱 CPU 過高。
- public route / queue readback 仍 blocked`https://registry.wooo.work/v2/` 502、`https://harbor.wooo.work/api/v2.0/health` 502Gitea public queue 回 `status=blocked_harbor_110_remote_ssh_publickey_auth_stalled`current CD `#4247 Waiting`Harbor repair `#4237 Running` 但 jobs API stale/cross-workflow mismatch。
- 13:46 cold-start artifact `/tmp/awoooi-full-stack-cold-start-after-110-console-20260701-134632.log``PASS=62 WARN=8 BLOCKED=6`。仍不可宣稱全服務 10 分鐘自動恢復hard blockers 含 110 registry `/v2`、110 SSH read-only check、K3s registry pull refused by `110:5000`、SignOz TLS / 502、Stock route 502、188 MOMO daily sales stale。
- 99 VMware console 已可進入 110 local shell現場只讀確認 110 home 有舊版 `full-stack-cold-start-check.sh``harbor-watchdog.sh``systemd-units-textfile-exporter.py` 等腳本,但沒有 `recover-110-control-path-and-harbor-local.sh``/home/wooo/awoooi` 也不存在。代表目前 scorecard 指向的 local recovery package 未在 110 live console 可用,這是自動恢復閉環缺口。
- `scripts/reboot-recovery/full-stack-cold-start-check.sh` 已補 machine-readable blockerSSH 110 blocked 時輸出 `/usr/local/bin/recover-110-control-path-and-harbor-local.sh``/usr/local/bin/repair-110-ssh-publickey-auth-local.sh``/usr/local/bin/check-awoooi-110-controlled-cd-lane-readiness.sh` expected paths並固定 `SSH_110_RECOVERY_PACKAGE_BLOCKER=package_presence_not_verifiable_while_remote_control_channel_unavailable` / `SSH_110_RECOVERY_PACKAGE_NEXT_ACTION=verify_or_preinstall_local_recovery_package_from_console_before_harbor_repair_retry`
**仍維持**
- 未讀 secret / token / `.env` / raw sessions / SQLite / auth未讀 authorized_keys 內容或 `.runner` 內容;未使用 GitHub / `gh` / GitHub API未 workflow_dispatch。
- 未重啟主機,未 restart Docker / Nginx / K3s / DB / firewallconsole 曾觸發 `sudo` password prompt 但已中斷,未輸入密碼。
**下一步**
- P0 不切支線:先讓 110 local recovery package 可在 110 console 或已恢復 SSH control path 執行;接著跑 `recover-110-control-path-and-harbor-local.sh --check`,只有 check 指向 metadata drift 時才做 `--apply-ssh-control-path`,再驗證 Harbor `/v2` 200/401、Gitea queue、cold-start、Stock freshness 與 SLO scorecard。
## 2026-07-01 — 14:15 Harbor receipt 納入 stale repair running blocker
**照主線修正的問題**