41 Commits

Author SHA1 Message Date
882cdf4226 docs: 修正 CT102 cloudflared 為 DoH proxy-dns(非 Tunnel)+ 新增 CT103 MediaMTX
- CT102: cloudflared 實為 proxy-dns(DoH,127.0.0.1:5053,上游 1.1.1.1/1.0.0.1),更正表格/架構圖標示
- CT103: MediaMTX v1.19.1 上線(pve-02, .103),更新狀態統計 19→20、節點負載、備份說明

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-21 13:11:23 +08:00
55e5a0d5ad docs: PVE_API.md 新增 PBS 區段(CLI + 備份/PBS API + datastore 擴容步驟)
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-21 07:48:42 +08:00
46b66af7ef docs: CLAUDE.md 新增 PBS 免密 SSH / proxmox-backup-manager 存取說明
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-21 07:47:41 +08:00
6c8396a65f docs: PBS datastore 擴容 80G→200G + 記錄 dedup/GC/驗證結果
- datastore 因全叢集首輪備份近滿(98%/剩2.1G),將 VM111 scsi2 由 80G 線上擴至 200G
- 擴容後使用率 39%(剩~116G);dedup ~17×(邏輯1.27TB→落地~75G)
- 記錄擴容步驟、GC daily、首次 verify 25/25 groups 0 errors
- 同步更新 VM111 規格列

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-21 07:45:44 +08:00
e792ac7a76 docs: PBS 新增每週 verify job + 記錄免密 SSH
- 建立 verify-job v-weekly(每週六 05:00, outdated-after=30)解決備份驗證狀態為「無」
- 記錄 PBS 主機名 pbs、開發機已設免密 SSH、datastore 路徑 /mnt/datastore

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-21 06:56:39 +08:00
b361559ccf docs: 重建備份排程 (2026-01 停用後 5 個月無備份)
- 三個排程於 2026-01-07 後全被停用 (enabled=0),導致長期無新備份
- 重建為 3 個分組 job:Prod(115/117)、OpenWRT(110)、catch-all(all=1)
- catch-all 自動含未來新建 guest;排除 PBS(111) 與模板(9000)
- 統一保留策略 7d/4w/3m;刪除指向已不存在 CT103/CT106 的舊 job
- 手動測試備份 CT116 確認 PBS pipeline 正常

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-18 19:01:06 +08:00
70f3fcd32c docs: 新增 CT124 LiteLLM AI Gateway 並記錄 healthcheck 假警報修正
CT124 表格與拓撲圖補上 LiteLLM AI Gateway (:4000, /opt/ai-proxy-hub)。
litellm 容器長期顯示 unhealthy 為假警報:healthcheck 用 curl 但 image 內無 curl,
已於宿主改用 python3 urllib(compose 檔在 CT124,不在本 repo)。

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-04 13:05:24 +08:00
bd62f397a1 docs: CT118 改 static .118 + adminer restart=always;SOP 新增 IP/Podman 復原檢查
2026-06-04 停電後 CT118(ip=dhcp)漂到 .207 導致以 .118 連 DB/Redis 的服務中斷,
改為 static 192.168.42.118;adminer 容器停在 Created(無 restart 策略)已重建為
--restart=always。SOP 新增步驟 5(IP 回 VMID pattern)與步驟 6(Podman restart 策略)。

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-04 12:34:54 +08:00
79d50ef6ed docs: add power-outage recovery SOP + onboot/startup tiering + K3s VM notes
停電事件處理收尾:三節點 2026-05-29 14:34 同時斷電復電。
- 新增「停電/不正常斷電復電 SOP」:判斷停電、losetup+e2fsck 離線修復根碟、pct start 卡 SSH 的坑
- 22 台全設 onboot=1 + startup 三層(T1網路/T2基礎/T3應用)對照表
- 新增「K3s VM 登入與操作」:ubuntu 帳號、kubectl 需 sudo、host key
- 紀錄停電造成 K3s VM(120/121/122)根碟 ext4 損壞,已全數 e2fsck 修復

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-05-29 21:01:45 +08:00
47d35990f1 docs: embed animated SVG cluster topology diagram in README
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-21 17:34:36 +08:00
d238644592 docs: CT100 acts as NetBird Subnet Router for 192.168.42.0/24 2026-04-18 22:27:29 +08:00
997516daa2 docs: CT124 adds NetBird client (100.71.40.35) + Tailscale iptables fix
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-18 16:28:44 +08:00
a44ed64122 docs: CT101 adds NetBird client (100.71.200.94)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-18 15:14:21 +08:00
54b31e4673 docs: CT100 adds NetBird client (100.71.236.202)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-18 14:50:15 +08:00
dd95dbe8c2 docs: add CT109 Hermes Agent v0.9.0 (NousResearch)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-18 14:32:31 +08:00
cb9f872e6a docs: VM111 PBS disk shrink 300G→80G, syno-lvm 89%→64%
Migrated PBS datastore from 300G to 80G disk, freeing 220G on syno-lvm.
Updated storage usage figures across all docs.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 17:07:33 +08:00
b5cd1e8bf5 docs: update VM9000 details (Ubuntu 24.04 cloud-init UEFI template)
VM9000 is a UEFI/q35 cloud-init template with Ubuntu 24.04, 2C/2G,
3.5G disk, user=ubuntu, SSH key preset, DHCP. Added clone usage
instructions to PVE description. Updated tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 15:15:57 +08:00
7aba5a7e6f docs: update CT104 details (Caddy static file server + Tailscale)
CT104 runs Caddy on port 80 (static file browsing with gzip) and
Tailscale client (Headscale node 100.64.0.14). Docker-compose at
/opt/caddy/ and /opt/tailscale/. Updated PVE description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 14:55:20 +08:00
4f3c9a3eb8 docs: optimize cluster — delete CT103/CT119, reduce mem/cpu allocation
Deleted:
- CT103 (N8N, stopped, 16G) and CT119 (Gotify, stopped, 8G)

Memory reduced (pending restart):
- CT108: 4096→2048, VM110: 1536→512, CT114: 512→128, CT118: 2048→1024

CPU reduced (pending restart):
- CT115: 4→2C, CT117: 4→2C, CT112: 2→1C, CT118: 2→1C

Updated node load distribution and status counts.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 14:45:09 +08:00
6b0ca3031e docs: update CT124 details (20+ Docker services hub)
CT124 is the main Docker service hub with 20+ services including:
NPM, Gitea (31337), Karakeep, Vaultwarden, Jellyfin, Redash, NocoDB,
OpenClaw, Prompt Optimizer, Browserless, Picsur, Docker Registry,
JS Bin, MySpeed, Drawnix, OHIF/Orthanc, Prometheus, Tailscale,
Watchtower. Multiple PostgreSQL instances. GPU passthrough (/dev/dri).
Updated PVE description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 14:29:43 +08:00
974c3c71d5 docs: update CT118 details (PostgreSQL 16, Redis, 10x Tor proxy pool)
CT118 runs Podman 4.9.3 with PostgreSQL 16 (5432), Redis Alpine (6379),
and a 10-node Tor proxy pool (SOCKS 9050-9059, Privoxy 8118-8127,
proxy.py frontend 8899/backend 9000). IP via DHCP. Updated PVE
description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 14:28:12 +08:00
771a37453d docs: update CT114 details (Podman 4.9.3 + Nginx Quadlet)
CT114 runs Podman 4.9.3 with nginx:alpine via Quadlet systemd
integration on port 8080. Static content in /srv/web/. LXC uses
apparmor=unconfined for Podman compatibility. Updated PVE description
and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 14:27:19 +08:00
4cf7235ee8 docs: update CT112 details (OpenClaw Gateway v2026.3.13, Node.js)
CT112 runs OpenClaw Gateway v2026.3.13 on Node.js v22.22.0 as a
systemd user service (port 18789). Also has agent-browser, clawhub,
and utility scripts for AQI and Synology API. Stored on local-lvm.
Updated PVE description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 14:25:59 +08:00
01b4c12628 docs: update CT108 details (12 Python microservices, supervisord+uv)
CT108 runs 12 Python microservices via supervisord with uv package
manager: phone-ip-inspector, holiday-checker, pokedex-service,
pokemon-radar-api, person, hoarder-automation, url-cleaning-tool,
lunar-converter, pogo-viewer, pogo-code-share (streamlit),
cloudflare-error-page, oidc-login-demo. All RUNNING. Updated PVE
description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 14:24:03 +08:00
2ca0a32073 docs: update VM111 details (PBS, 294GB datastore, 41 backups)
VM111 runs Proxmox Backup Server at 192.168.42.30:8007. Datastore
data-store has 294GB (5% used), 41 backups covering CT103/106/108/110/
115/117. Guest agent not running, SSH key auth unavailable. Updated
PVE description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:58:16 +08:00
5d22e04d36 docs: update CT117 details (Lydia test: Python:5000, MSSQL Express)
CT117 runs Lydia test env: Python 3.12 app on port 5000 via supervisord
(RUNNING), MSSQL 2022 Express on port 1433. Caddy installed but inactive.
SSH exposed externally via OpenWRT DNAT port 50022. Updated PVE
description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:55:34 +08:00
71705296e8 docs: remove CT109 (Stirling PDF) from cluster inventory
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:50:34 +08:00
ec3bc864b6 docs: update CT115 details (Lydia prod: Python, Caddy, MSSQL Express)
CT115 runs Lydia production: Python 3.12 app via supervisord (currently
FATAL), Caddy v2.6.2 web server, MSSQL 2022 Express on port 1433.
SSH exposed externally via OpenWRT DNAT port 30022. Updated PVE
description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:49:30 +08:00
f2cb3326e9 docs: update CT107 details (MailPit email testing server)
CT107 runs axllent/mailpit with Web UI on port 8025 and SMTP on
port 1025. Max 5000 messages, accepts any SMTP auth. Docker-compose
at /opt/mailpit/. Updated PVE description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:46:34 +08:00
cb8aedb7fa docs: update CT101 details (RustDesk Server, Headscale backup)
CT101 runs RustDesk hbbs/hbbr (ports 21115-21119, host network) as
self-hosted remote desktop server, plus a backup Headscale instance
(primary is CT126). Also Watchtower on Ubuntu 22.04. Updated PVE
description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:44:59 +08:00
2183e8c107 docs: update CT100 details (SSH Jump, Bore Server, Tailscale)
CT100 runs SSH jump box, Bore tunnel server (port 7835, host network),
Tailscale client (Headscale node 100.64.0.13), and Watchtower on
Ubuntu 22.04. Taiwan IP only access via OpenWRT DNAT. Updated PVE
description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:41:29 +08:00
5d5a0559c5 docs: update CT126 details (Headscale v0.28.0 control server, 17 nodes)
CT126 runs Headscale v0.28.0 as self-hosted Tailscale control server
with 17 devices (Mac, iPhone, Android, Apple TV, Linux servers).
Ports 8080/9090/50443 forwarded via OpenWRT DNAT. Also runs Tailscale
client and Watchtower. Updated PVE description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:37:10 +08:00
48d33e1c39 docs: update CT116 details (Bark push notification server)
CT116 runs finab/bark-server on port 8080 for iOS push notifications,
plus Watchtower. Both managed via docker-compose in /opt/. Updated PVE
description and tags.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:35:38 +08:00
f0411d9910 docs: update CT105 details (EchoIP, not MyIP)
CT105 actually runs mpolden/echoip on port 8080 (not MyIP on 18966
as previously documented). Also runs Watchtower and Postfix (localhost).
Updated PVE description.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:33:03 +08:00
5740c5c536 docs: update CT102 details (Tailscale, Cloudflare Tunnel, Proxy Pool)
CT102 runs Tailscale VPN (17 devices), Cloudflare Tunnel, Tor proxy,
proxy.py aggregator/backend, Telegram MTProto proxy, and Watchtower.
Also updated PVE description.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:31:10 +08:00
bbe9aa0ac2 docs: update VM110 OpenWRT details (24.10.4, PPPoE, WireGuard, DNAT rules)
OpenWRT 24.10.4 serving as main gateway with PPPoE (HiNet), WireGuard
VPN, and port forwarding for SSH Jump, Bore, NPM, RustDesk, Headscale,
SQL Server, and SSH to CT115/CT117. Also updated PVE description.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:29:51 +08:00
7610d55811 docs: update K3s cluster details (VM120 server, VM121/122 agents)
VM120 is the K3s control-plane (server), VM121/122 are agents (workers).
K3s v1.33.6 with Longhorn storage. Running services: Traefik, CodiMD,
Adminer, HTTPbin, OpenGist, PostgreSQL, Vaultwarden.
Also updated PVE descriptions for VM120/121/122.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:27:30 +08:00
272dbb9747 docs: remove CT125 from cluster inventory
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 13:24:42 +08:00
2a661dfccc docs: remove deleted CTs, add CT125, document Ubuntu 25.04 LXC console fix
- Remove CT106, CT113, CT123 (deleted)
- Add CT125 (Ubuntu 25.04, pve, 192.168.42.125)
- Document Ubuntu 25.04 console-getty ImportCredential issue and fix
- Update node load distribution and status counts

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 12:23:44 +08:00
1990d5a011 docs: add README, API reference, quickstart guide, and cluster summary
- README.md: project landing page with cluster overview and doc index
- API.md: comprehensive Proxmox VE API endpoint reference
- QUICKSTART.md: 5-minute guide with copy-paste examples
- SUMMARY.md: cluster health, resource usage, and service inventory
- CLAUDE.md: Claude Code project context

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 11:12:15 +08:00
d30bf61a76 docs: add PVECluster architecture and API reference
Include cluster architecture overview (nodes, VMs, storage, networking,
backup strategy) and Proxmox VE API usage guide with token credentials.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-10 11:08:10 +08:00