Alexa Amundson
78fbe80f2a
bin/ 230 CLI tools (ask-*, br-*, agent-*, roadid, carpool) scripts/ 99 automation scripts fleet/ Node configs and deployment workers/ Cloudflare Worker sources (roadpay, road-search, squad webhooks) roadc/ RoadC programming language roadnet/ Mesh network (5 APs, WireGuard) operator/ Memory system scripts config/ System configs dotfiles/ Shell configs docs/ Documentation BlackRoad OS — Pave Tomorrow. RoadChain-SHA2048: d1a24f55318d338b RoadChain-Identity: alexa@sovereign RoadChain-Full: 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
51 KiB
51 KiB
BlackRoad Device Registry — Master Index
Generated 2026-03-09 (Exhaustive Deep Scan — All Nodes)
Every device gets a unique ID (1-255)
═══════════════════════════════════════════════════════════════════════════════════
ID NAME TYPE LOCATION / IP STATUS
═══════════════════════════════════════════════════════════════════════════════════
────── CORE INFRASTRUCTURE (1-10) ──────
1 eero Mesh Router Gateway 192.168.4.1 ONLINE
MAC: 44:ac:85:94:37:92 Base MAC: 44:ac:85:94:37:80
mDNS: eero-0bp4.local BSSID: 44:AC:85:94:37:87
WiFi: 5GHz CH100 SSID: asdfghjkl Subnet: 192.168.4.0/22
Thread/Matter mesh (_meshcop, _trel), Sleep Proxy
2 Anastasia Cloud VPS 174.138.44.45 ONLINE
DigitalOcean nyc1 | 1 vCPU, 765MB RAM | CentOS 9 Stream
Kernel: 5.14.0-651.el9.x86_64 | Uptime: 71 days
Disk: 25GB (67% used = 17GB) — was 94%, cleaned up
WG hub :51820 | Headscale :8080/:9090 | Ollama (TS only)
Nginx :80 (blackroad.io) | uvicorn :8000 | Redis :6379
WebSockets :8765/:8766 | Node :3000/:3001 | PM2
Cloudflared | fail2ban | lucidia-agent | blackroad-api
Cron: agent-health.sh every 5m
Failed: caddy, cockpit
WG peers: alice(.6)✓ cecilia(.3)✓ octavia(.4)✓ aria(.7)⚠1h gematria(.8)✓
2 unused: 10.8.0.2, 10.8.0.5
3 Gematria Cloud VPS 159.65.43.12 PARTIAL
DigitalOcean nyc3 | 4 vCPU, 8GB RAM | Ubuntu 22.04
Disk: 80GB (43GB used) | SSH DOWN but WG alive (10.8.0.8)
Caddy, Ollama, NATS, Cloudflared, Tailscale
────── COMPUTE NODES (11-20) ──────
11 Alexandria Mac 192.168.4.28 ONLINE
MAC: b0:be:83:66:cc:10 | macOS Darwin 23.5.0
Tailscale: 100.117.200.23 (idle)
12 Alice Pi 400 192.168.4.49 ONLINE
MAC: d8:3a:dd:ff:98:87 | BCM2835 Rev c03130
Serial: 1000000091da3c05 | 4GB RAM (443MB used)
Kernel: 6.1.21-v8+ (2023-04-03) ⚠️ OLD
SD: 15GB (77% = 11GB used) | Temp: 38.9°C | Throttle: 0x0
CPU: ondemand, 1.8GHz (Cortex-A72) | 1926 packages
PCIe: BCM2711 bridge + VIA VL805 USB 3.0 (Pi 400 arch)
Timers: blackroad-watchdog (30s), gitops-sync (5m), prism-pull (2m)
dmesg: brcmf_cfg80211_change_iface errors (WiFi/RoadNet AP conflict)
DNS: 8.8.8.8, 8.8.4.4, 100.100.100.100, fdbc: (IPv6)
/etc/hosts: alice listed twice (127.0.1.1)
Docker: 0 volumes, 0 containers, default networks only
Git repos: road-deploy, road-control, .lucidia (4 repos)
42 running services | 4 failed (caddy, dnsmasq, nginx, prism-pull)
Timers: blackroad-watchdog (30s), prism-pull (2m), gitops-sync (5m)
Cron (pi): stats-push (15m, PUSH_SECRET in plaintext)
Pi-hole: FTL active on :53 (UDP+TCP, IPv4+IPv6), blocking enabled
PostgreSQL 13: "blackroad" DB (0 tables) | Redis: 1 key in db0
Qdrant: 0 collections | task_watchdog.py monitors Redis worker heartbeats
/opt/prism/: EMPTY dir (prism-pull timer pulls from empty git repo)
Git repos: road-deploy (2 dirty), road-control, .lucidia, blackroad-worlds
Disk: 15GB SD, 77% full (3.2GB free) — tight!
PM2: road-control (online), road-deploy (stopped), compliance (stopped)
GitHub Actions Runner: BlackRoad-OS-Inc-blackroad.alice-pi
Pi-hole DNS :53 | PostgreSQL 13 :5432 | Qdrant :6333/:6334
Redis :6379 | Cloudflared | Docker | WireGuard (10.8.0.6)
30 TCP ports | 15 UDP ports | iowait: 15.32% on mmcblk0 (SD stress)
Python services: agent(:8180), operator(:8001), dashboard(:3000),
agents-proxy(:4010), status(:8013), task-queue(:8184),
agents-daemon, blackroad-ops(:8012), world-engine, stats-proxy(:7890)
Node.js: road-control(:8083), model-server(:8787), prism-agent
RoadNet AP: uap0 CH1 10.10.1.0/24 | brcmfmac roamoff=0 (causes iface errors)
max_usb_current=1 in config.txt, sg3-utils installed
Cron: Pi-hole gravity update (weekly), log flush (daily), sysstat (10m)
Docker: 0 images, 0 containers (Docker installed but unused)
⚠️ hostapd crash loop: "Failed to start Access point" (brcmfmac limitation Pi 400)
⚠️ blackroad-status.service missing venv | rclone WebDAV mount failed
13 Cecilia Pi 5 192.168.4.96 ONLINE
MAC: 88:a2:9e:3b:eb:72 | Rev d04171
Serial: 98db8e2ec705f7c4 | 8GB RAM (1.3GB used)
Kernel: 6.12.62+rpt-rpi-2712 (2025-12-18)
NVMe boot: 457GB (17% = 73GB used) | SD: 238.4GB (backup)
Temp: 46.8°C (CPU) / 39.8°C (NVMe) / 42.1°C (RP1)
hwmon: cpu_thermal, nvme, rp1_adc, pwmfan, rpi_volt
Swap: 2GB (15MB used) | Watchdog: /dev/watchdog0
RTC: /dev/rtc0 | Timezone: America/Chicago ✅ FIXED (was Europe/London)
44 enabled services | Docker v29.2.1 (+buildx, compose, model plugins)
Users (5): root, cecilia, postgres, blackroad, alexa
Sudo NOPASSWD: cecilia, blackroad
SSH keys: blackroad=2
Input: pwr_button, gpio_ir_recv (IR receiver!), 2x HDMI CEC
Audio: 2x HDMI output (vc4-hdmi-0/1) | Video: 17 /dev/video devices
Services: cece-api, cece-heartbeat, cece-net-gateway, blackroad-dashboard,
blackroad-monitor, brnode, stats-proxy, wayvnc, wayvnc-control,
cloudflared, dnsmasq, docker, hailort, hostapd, influxdb,
lightdm, minio, nginx, node_exporter, ollama, pironman5,
postgresql, roadnet, roadnet-failover, rpcbind, nfs-blkmap
CECE API :3100 | TTS :5001 | Monitor :5002 | Dashboard :3001/:3000
Model Server :8787/:8788 | Ollama :11434 (13 models) | Proxy :11435
PostgreSQL 17 :5432 | MinIO :9000/:9001 | InfluxDB :8086/:8088
Nginx :80/:8080 | VNC :5900 | Cloudflared | Docker | Hailo-8
WireGuard (10.8.0.3) | Pironman5 :34001 | CUPS :631 | RPC :111
Port :7890 (stats-proxy) | :8788 (blackroad-api uvicorn)
30 TCP | 12 UDP
Firewall: nftables NAT (Docker + RoadNet 10.10.2.0/24 → wlan0/wg0)
Cron (blackroad): autonomy heartbeat (1m), heal (5m), rclone→gdrive (15m),
brady-bunch SDL display (@reboot), blackroad-server (@reboot),
blackroad-api uvicorn :8788 (@reboot), ollama-proxy (@reboot),
git pull master (5m), stats-push (15m, PUSH_SECRET=34e1aa...),
github-relay Gitea→GitHub (30m) ⚠️ HARDCODED CREDS: blackroad:BlackRoad2026OS
⚠️ Gitea URL in relay still points to .97 (now .100)
Ollama models (15): cece, cece2, deepseek-r1:1.5b, nomic-embed-text,
OpenELM-1B/3B, qwen3:8b, llama3:8b, codellama:7b, deepseek-coder:1.3b,
qwen2.5-coder:3b, llama3.2:1b/3b, tinyllama | CECE actively loaded
rclone remotes: gdrive:, gdrive-blackroad: (Google Drive backup)
Brady-bunch: pygame/curses fleet status grid on SDL framebuffer
✅ REMOVED: obfuscated /tmp/op.py cron entry (dropper pattern)
rclone: 3 concurrent instances eating memory!
RoadNet AP: uap0 CH6 10.10.2.0/24
Desktop: labwc + Xwayland + wf-panel-pi + pcmanfm
eth0: direct cable to BliKVM (1Gbps, 0 rx packets)
Cloudflare Tunnel d67bf4a5 (22 hostnames):
blackroad.io, www.blackroad.io, blackroadai.com, blackroad.me,
blackroad.network, blackroad.systems, aliceqi.com, lucidiaqi.com,
lucidia.earth, lucidia.studio, roadchain.io → nginx :8080
cecilia.blackroad.io → :3000 | api.blackroad.io → :8788
gateway.blackroad.io → :8787 | ollama.blackroad.io → :11434
monitor → :9100 | dashboard → :3001 | agents → :4010
metrics → :9000 (MinIO) | storage → :9001
Docker: 0 images, 0 containers (Docker v29.2.1 installed but unused)
Ollama blobs: 22GB (13 models)
Hailo params: force_desc_page_size=4096, support_soft_reset=Y
Cron (/etc/cron.d): nginx @reboot
✅ dnsmasq FIXED: bind-interfaces→bind-dynamic + listen 192.168.4.96 (was 0.0.0.0)
cece-net.conf: custom DNS zones (.cece, .blackroad, .entity, .soul, .dream) → 192.168.4.22
⚠️ Voltage: 0.8693V after vcgencmd fix (mknod /dev/vcio c 100 0)
Unix sockets: hailort_uds.sock, docker.sock, postgresql, cups, wayvnc
CPU: ondemand, 2.4GHz | 2197 packages | 408 Python packages!
PCIe: Hailo-8 + Micron/Crucial NVMe (5427) behind ASM1182e + RP1
Python: torch 2.10.0+cpu, hailo-tappas-core 5.1.0, hailort 4.23.0,
Flask, FastAPI, celery, redis (full AI/ML stack)
Node global: wrangler 4.68.0
GPU: /dev/dri/card0, card1, renderD128
⚠️ dmesg: Undervoltage detected! (9+ events — CECILIA ALSO UV!)
⚠️ 4 rclone instances now (growing from 3!) — all syncing same gdrive
⚠️ vcgencmd broken for non-root (needs: sudo mknod /dev/vcio c 100 0)
2x wayvnc: desktop + rpi-connect remote access
/tmp: fleet-telemetry.py (22KB), stats-proxy.py (8KB) — NO op.py found
30+ git repos in blackroad-source/repos/ (anthropic, canva, openai, aws, etc.)
Established: 4 rclone→Google Drive (IPv6), SSH→GitHub, InfluxDB
14 Octavia Pi 5 192.168.4.100 (was .97) ONLINE
MAC: 88:a2:9e:10:0a:3a | Rev d04171
Serial: a91e903b3e7bfcc4 | 8GB RAM (1.2GB used)
Kernel: 6.12.62+rpt-rpi-2712 (2025-12-18)
SD boot: 117GB (66% = 74GB used) | NVMe: 916GB (1% = 3GB used)
Temp: 38°C (CPU) / 36.8°C (NVMe) / 36.3°C (RP1)
hwmon: cpu_thermal, nvme, rp1_adc, rpi_volt (NO pwmfan — fan missing/broken?)
Swap: 2GB (0 used) | Watchdog: /dev/watchdog0
RTC: /dev/rtc0 | Timezone: America/North_Dakota/Center
39 enabled services | Docker v29.2.1 (+buildx, compose, model)
Users (5): root, pi, lucidia, blackroad, alexa
Sudo NOPASSWD: (default pi via sudoers.d)
⚠️ SSH keys: pi=52 (excessive — security concern)
Input: pwr_button, gpio_ir_recv (IR receiver!), 2x HDMI CEC
Audio: 2x HDMI playback | Video: 17 /dev/video (pispbe + rpi-hevc-dec)
Services: blackroad-agent (/opt/blackroad/agent.py as root),
blackroad-dashboard, blackroad-monitor, blackroad-worker-hailo,
brnode, cloudflared, docker, github-runner-lucidia, hailort,
hostapd, influxdb, lightdm, octoprint, ollama, pironman5,
roadnet, roadnet-failover, stats-proxy, wayvnc-control,
dnsmasq, cups-browsed, rpcbind, nfs-blkmap, node_exporter
Docker Swarm LEADER :2377 (alice=Down, aria=Down, octavia=Ready)
Containers: gitea(:3100/:2222), NATS(:4222/:8222 via ingress),
ollama, edge-agent(:8082)
OctoPrint :5000 | InfluxDB :8086/:8088 | Ollama :11434 (9 models)
Model Server :8787 | Dashboard :3000 | Agent :8080
Cloudflared | dnsmasq :53 | CUPS :631 | Pironman5 :34001
WireGuard (10.8.0.4) | 29 TCP | 25 UDP
Port :7890 (stats-proxy) | :9100 (node_exporter)
Firewall: nftables Docker Swarm NAT + INGRESS (NATS 4222/8222, edge 8082)
+ RoadNet 10.10.3.0/24 masquerade → wlan0/wg0
Cron (pi): git pull blackroad-parts (5m), stats-push (15m, same PUSH_SECRET)
Cron (blackroad): autonomy heartbeat (1m) + heal (5m) — self-healing service checks
NVMe: /mnt/nvme/blackroad/ (3GB), models/ (26MB), quantum_discoveries/ (2.8MB)
Hailo-8: /dev/hailo0 ✓ FIXED | hailo_pci module loaded | hailort_uds.sock
I2C 0x3c: Pironman5 OLED display
Desktop: labwc + wf-panel-pi + pcmanfm
RoadNet AP: uap0 CH11 10.10.3.0/24
USB: no external devices (4 root hubs only)
CPU: ondemand, 2.6GHz (FASTEST in fleet!) | 2290 packages | 318 Python pkgs
PCIe: Hailo-8 + Micron/Crucial NVMe (5427) behind ASM1182e + RP1
Python: Flask, hailort 4.23.0, nats-py, redis (no torch)
GPU: /dev/dri/card0, card1, renderD128
Throttle: 0x50000 (historical, NOT currently) | Voltage: 0.8451V ✅ (was 0.750V pre-optimize)
Docker volumes: "ollama" | 8 networks (2 overlay swarm, 4 bridge, host, none)
Docker Swarm: 4 nodes (alice=Down v28.5.2, aria=Down v20.10.24, octavia=Leader v29.2.1, old-octavia=Down)
Swarm services: blackroad-nats (1/1 replicas), my-service (0/4 FAILED)
Cloudflare Tunnel b7e9f25e (10 hostnames):
lucidia.blackroad.io → :3000 | api-lucidia → :8000
monitor-lucidia → :9090 | ssh-lucidia → ssh://22
git.blackroad.io → :3100 | code.blackroad.systems → :3100
roadcode.blackroad.systems → :3100
cloud.blackroad.io → :3200 | cloud.blackroad.systems → :3200
Docker images (11): gitea 260MB, NATS 22MB, ollama 8.57GB, nginx:alpine 93MB,
edge-agent 264MB, blackroad.systems 81MB, blackroadai.com 81MB,
auth-gateway 970MB, metaverse 81MB, ultimate 81MB, octoprint 1.77GB
Ollama blobs: 15GB (9 models) | ⚠️ qwen manifest corrupted (EOF errors)
Ollama journal: 500 errors on /api/chat and /api/generate (2min timeouts)
Hailo params: force_desc_page_size=4096, support_soft_reset=Y
Claude Time: 2.8GB on NVMe (/mnt/nvme/blackroad/claude/)
Established: git→GitHub, Runner.Worker→GitHub Actions, cloudflared (100.49.232.196)
Git repos: quantum/blackroad-os-quantum, actions-runner, Claude config (NVMe)
OctoPrint requires API key for access
15 Aria Pi 5 192.168.4.98 OFFLINE ⚠️
MAC: 88:a2:9e:xx:42:07
8GB RAM | 30GB SD (96% full)
Last WG handshake: ~1h ago (went down recently)
Portainer :9443 | Headscale :8090 | Ollama (4 models)
Pironman5 | Magic Keyboard BT | TONOR mic (intermittent)
WireGuard (10.8.0.7) | RoadNet AP: uap0 CH1
NOT PINGABLE — host is down, needs physical check/reboot
16 Lucidia Pi 5 192.168.4.38 ONLINE ⚠️HOT
MAC: 2c:cf:67:cf:fa:17 | Rev d04171
Serial: aa088196e6935b14 | 8GB RAM (4.7GB used = 59%!)
Kernel: 6.12.62+rpt-rpi-2712 (2026-01-19)
SD: 235GB (41% = 90GB used) | Temp: 57.9°C ✅ FIXED (was 73.8°C) (RP1: 52°C)
hwmon: cpu_thermal, rp1_adc, pwmfan, rpi_volt (no NVMe — SD-only boot)
Swap: 8.5GB (437MB used) via dphys-swapfile
Watchdog: /dev/watchdog0 | RTC: /dev/rtc0 | Timezone: America/Chicago
Hostname: "octavia" (MISNAMED — Tailscale name "lucidia")
62 enabled services! (highest in fleet)
Users (9): root, pi, postgres, deploy, lucidia, nova, octavia, blackroad, alexa
Sudo NOPASSWD: pi, octavia
SSH keys: octavia=15
PipeWire running for blackroad + alexa users
Input: pwr_button, 2x HDMI CEC, 2x HDMI Jack (no IR receiver)
Audio: 2x HDMI output | Video: 17 /dev/video devices
Firewall: UFW (nftables + iptables) — INPUT policy DROP!
Tailscale (ts-input/ts-forward chains), Docker NAT, mDNS allowed
Services: actions.runner (GitHub CI), blackroad-agent, blackroad-api,
blackroad-relay, blackroad-salesforce-agent, btc-compose, cloudflared,
docker, dphys-swapfile, fail2ban, hostapd, influxdb, java-hello,
lightdm, llama, lucidia, nginx, node_exporter, ollama, ollama-bridge,
operator, pironman5, pm2-pi, postgresql, roadnet, roadnet-failover,
rpi-display-backlight, simpleweb, tailscaled, ufw, wayvnc-control
Docker: 11 containers running!
road-pdns-admin :9192 (healthy), road-dns-db (postgres:15-alpine),
roadapi :4001, roadauth :4002, blackroad-edge-agent :8082/:9090,
blackroad.systems :3005, blackroadai.com :3006,
blackroad-auth-gateway, blackroad-metaverse :3109,
blackroad-os-carpool :3002, pi-my-agent :8080 (healthy)
Docker images (14): edge-agent 182MB, blackroad.systems 54MB,
blackroadai.com 54MB, auth-gateway 704MB, metaverse 54MB,
ultimate 54MB, carpool 819MB, postgres:15-alpine 270MB,
pi-eps 268MB, pi-my-agent 383MB, node:18-alpine 126MB,
powerdns/pdns-auth-48 186MB, powerdnsadmin/pda-legacy 222MB,
ruimarinho/bitcoin-core 145MB
Tailscale peers: lucidia(self), alexandria(idle), alice(offline 15d),
aria(offline 15d), cecilia(relay "ord"), octavia(offline 15d),
codex-infinity(offline 15d), lucidia-operator(offline 33d),
shellfish(offline 15d)
Lucidia API :8000 | Ollama :11434 (6 models) | Bridge :3109
PostgreSQL 17 :5432 | InfluxDB :8086 | Nginx :80 (530+ apps)
Tailscale :38850 (100.66.235.47) | 2x Cloudflared :20241/:20242
Node Exporter :9100 | Pironman5 | CUPS :631
Java Hello World :8888 | simpleweb | socat relay :8011
Model Server :8787 | stats-proxy | dnsmasq :53
Ports: 60+ TCP! :3000-:3006, :4001/:4002, :5000-:6300 (15+ Python),
:8000, :8011, :8080-:8090, :8180/:8182, :8787, :8888-:8889, :9090/:9192
Cloudflare Tunnel 0447556b (4 hostnames):
octavia.blackroad.io → :3000 | api-octavia → :8000
monitor-octavia → :9090 | ssh-octavia → ssh://22
Note: Lucidia serves "octavia.*" domains — names reflect old hostname confusion
but routing is CORRECT (each tunnel routes to localhost services on the right box)
Ollama blobs: 20GB (6 models) | lucidia:8b loaded = 265% CPU
Ollama-bridge: SSE proxy service
Lucidia API: /home/pi/lucidia/.venv/bin/uvicorn main:app :8000 (FastAPI)
Environment: /etc/opt/lucidia.env
llama.service: /home/pi/llama.cpp/build/bin/llama-server :8080
Model: /home/pi/models/model.gguf | ConditionPathExists checks
btc-compose: ruimarinho/bitcoin-core + EPS (docker compose) — FAILED
/var/www: 6 dirs (blackroad=335 items!, blackroad-cloud, blackroad-console,
blackroad-data, certbot, html, lucidia) — 530+ static apps
Nginx sites-enabled: blackroad-cloud, blackroad-domains, blackroad-multi
blackroad-multi: wildcard map *.blackroad.io → /var/www/blackroad/blackroad-$sub
Swap: 8.5GB via dphys-swapfile (1.3GB used, growing)
Cron (octavia): blackroad-agent-startup (5m) — script is EMPTY
Cron (root): certbot renewal (3am daily) — ⚠️ CERT EXPIRED Oct 2025!
Disk: /home/blackroad/ = 33GB, /home/alexa/ = 1.8GB
SuperDrive A1379 on USB (sr0) — CANNOT load disc (OC)
RoadNet AP: uap0 CH11 10.10.5.0/24
Desktop: labwc + wayvnc + wf-panel-pi
Named pipes: .NET debug pipes for GitHub Actions Runner (CLR)
CPU: ondemand, 2.4GHz | 2208 packages | 339 Python pkgs
PCIe: RP1 South Bridge ONLY (no NVMe, no Hailo — SD boot)
Python: Flask, gunicorn, redis (no torch, no hailo)
Node global: npm, pm2@6.0.14
PM2 (pi): road-registry-api (40.8MB, online, port :8090)
GPU: /dev/dri/card0, card1, renderD128
✅ Ollama runner FIXED: world-engine.py (alexa user service) was calling /api/generate
blackroad-world.service disabled, world-engine killed, temp 73.8°C → 57.9°C
Microservices (14, all Python, under /home/blackroad/):
:5100 load-balancer/app.py (18MB) | :5200 fleet-monitor/app.py (24MB)
:5300 notifications/app.py (22MB) | :5400 metrics/app.py (24MB)
:5500 analytics/app.py (25MB) | :5600 grafana/app.py (16MB)
:5700 alert-manager/app.py (16MB) | :5800 log-aggregator/app.py (18MB)
:5900 backup-system/app.py (14MB) | :6000 perf-cache/app.py (17MB)
:6100 resource-optimizer/app.py (17MB) | :6200 compression-middleware/app.py (17MB)
:6300 connection-pool/app.py (17MB) | :5002 monitoring/monitor-api.py (24MB)
Also: :5001 tts-api/app.py (24MB) | :5000 simpleweb gunicorn (12MB x2)
Total: ~350MB RAM on 21 Python processes + Docker containers
/var/www/blackroad: 334 static app directories (nginx wildcard multi-site)
Samples: ai-agent-framework, blockchain-explorer, carbon-tracker, chaos-engineering...
Alexa user services: blackroad-api (active), blackroad-status (active),
blackroad-world (disabled), blackroad-git-worker (disabled)
⚠️ Leaked GitHub PAT in alexa's blackroad-git-worker.service (gho_Gfu...)
⚠️ dmesg: "mmc0: Card stuck being busy!" — SD card degrading!
⚠️ dmesg: sr0 "Can't lookup blockdev" + USB error -71 (SuperDrive)
⚠️ DNS: Tailscale MagicDNS ONLY (100.100.100.100) — no fallback!
⚠️ cmdline: cgroup params duplicated 4x (messy config.txt)
⚠️ vcgencmd broken (needs: sudo mknod /dev/vcio c 100 0)
UFW: deny incoming, allow outgoing, deny routed
Allow: SSH anywhere, 8000 from LAN, 8080+50001 on tailscale0, 8180 anywhere
Docker: 3 volumes (caddy_config, caddy_data, pdns-db), 9 networks, 12 containers
Git repos (10): blackroad-agents, actions_github_pages_* (4x timestamped),
.nvm, .lucidia, blackroad, untitled-folder
Established: Alice SSH tunnel, GitHub Actions, Google Drive rclone,
cloudflared, InfluxDB, Tailscale, Microsoft Azure (Actions infra)
Last logins: pi autoloads tty1, rebooted 2x on Mar 8, alexa SSH Mar 3
⚠️ Failed logins: blikvm, admin, kvmd from Mac (192.168.4.28)
btc-compose = "Bitcoin Core + EPS (docker compose)" — FAILED
fail2ban FAILED | certbot FAILED — SSL EXPIRED Oct 2025, no intrusion protection
Nginx sites: blackroad-cloud, blackroad-domains, blackroad-multi
SSL: blackroad.io cert (domains: blackroad.io, blackroadinc.us + www variants) EXPIRED
Timers: brnode-heartbeat (5m), certbot (12h), tls-watch (daily), pins-refresh (weekly)
GitHub Actions: 21 runner dirs under /home/blackroad/runners/ (19GB total!)
Repos: blackroad-{agents,api,cli,core,docs,gateway,hardware,infra,math,operator,
sdk,sf,web}, blackroad-os-{agents,api-gateway,core,docs,helper,mesh,web}, demo-repository
2 runner processes: blackroad (Runner.Listener) + pi (Runner.Listener via systemd)
/etc/hosts: 127.0.1.1 → lucidia (correct despite hostname "octavia")
17 raspberrypi Pi 5 192.168.50.63 OFFLINE
MAC: 88:a2:9e:3b:eb:70 | On Alice eth0 secondary subnet
Powered off
────── AI ACCELERATORS (21-25) ──────
21 Hailo-8 #1 AI PCIe Cecilia M.2 ONLINE
Serial: HLLWM2B233704667 | Part: HM218B1C2FAE
FW: 4.23.0 (release,app,extended context switch buffer)
Product: HAILO-8 AI ACC M.2 M KEY MODULE EXT TEMP
/dev/hailo0 ✓ | hailort.service active
9 LIVE models: YOLO5Seg, YOLO5Face, YOLO6, YOLO8, etc.
26 TOPS | PCIe Gen3 x1 via ASM1182e switch
22 Hailo-8 #2 AI PCIe Octavia M.2 ONLINE ✓ FIXED!
Serial: HLLWM2B233704606 | Part: HM218B1C2FAE
FW: 4.23.0 (release,app,extended context switch buffer)
Product: HAILO-8 AI ACC M.2 M KEY MODULE EXT TEMP
/dev/hailo0 ✓ | hailo_pci module loaded | persistent via /etc/modules-load.d/
Pre-compiled models: resnet_v1_50.hef, yolov5s.hef
26 TOPS | PCIe Gen2 x1 via ASM1182e switch
⚠️ Octavia has undervoltage (0x50005) — PSU may be insufficient for Hailo+NVMe
────── STORAGE DEVICES (26-35) ──────
26 Alice SD microSD Alice 14.8 GB (77%)
27 Cecilia NVMe NVMe SSD Cecilia PCIe 465.8 GB (17% boot)
Micron/Crucial Device 5427 | PCIe Gen2 x1 via ASM1182e
28 Cecilia SD microSD Cecilia 238.4 GB (backup)
29 Octavia NVMe NVMe SSD Octavia PCIe 931.5 GB (1% = 3GB)
Micron/Crucial Device 5427 | PCIe Gen2 x1 via ASM1182e
30 Octavia SD microSD Octavia 119.1 GB (66% boot)
31 Aria SD microSD Aria ~30 GB (96% full!)
32 Lucidia SD microSD Lucidia 238.8 GB (41%)
33 Anastasia Disk VPS Disk Anastasia 25 GB (67%)
34 Gematria Disk VPS Disk Gematria 80 GB (54%)
────── USB DEVICES (36-50) ──────
36 Alice Keyboard USB HID Alice USB 1-1.4 04d9:0007 Holtek ONLINE
37 Alice SD Reader USB Storage Alice USB 1-1.2 14cd:1212 Super Top ONLINE (no card)
38 Alice USB Hub USB Hub Alice USB 1-1 2109:3431 VIA Labs ONLINE
39 Cecilia CP2102 USB Serial Cecilia USB 3-2 10c4:ea60 SiLabs ONLINE (SILENT)
/dev/ttyUSB0 | 9600 8N1 | Timeout errors (-110)
40 Cecilia Pixart Mouse USB HID Cecilia USB 1-1 — MISSING
41 Cecilia Sipeed M1s USB Storage Cecilia USB 3-2 BL808 RISC-V MISSING
Triple-core SoC (C906+E907+E902) | WiFi+BLE+Zigbee
8MB flash | firmware.bin = BL IoT SDK WiFi stack
42 Lucidia SuperDrive USB Optical Lucidia USB 3-2 05ac:1500 Apple DEGRADED
/dev/sr0 | Serial: KZAZ9NI1741 | FW: 2.03
24x/24x writer cd/rw xa/form2 cdda caddy
USB 2.0 (480M) | bMaxPower: 500mA (actually needs ~1.1A)
77+ over-current events | Cannot load disc
43 Aria TONOR Mic USB Audio Aria USB TC-777 MISSING
44 Aria Magic Keyboard Bluetooth Aria BT Apple A1644 OFFLINE
45 Cecilia IR Receiver GPIO Input Cecilia gpio_ir_recv rc0 ONLINE
IR remote control receiver on GPIO pin — can receive IR commands
46 Octavia IR Receiver GPIO Input Octavia gpio_ir_recv rc0 ONLINE
IR remote control receiver on GPIO pin — can receive IR commands
────── NETWORK PERIPHERALS (51-55) ──────
51 BliKVM KVM-over-IP Cecilia eth0 DEAD
Direct 1Gbps cable | 0 rx packets | No IP either end
Scanned 169.254.x.x + 192.168.1.x — no response
HDMI+USB passthrough to Cecilia
52 Sipeed BL808 SoC RISC-V MCU Cecilia USB MISSING
(see #41 — same device, listed both as USB and network peripheral)
────── I2C / SPI DEVICES (56-65) ──────
56 Octavia OLED I2C Display Octavia i2c-1 @ 0x3c ONLINE (Pironman5)
57 Alice I2C-1 I2C Bus Alice ONLINE
58 Alice I2C-20 I2C Bus Alice ONLINE
59 Alice I2C-21 I2C Bus Alice ONLINE
60 Alice SPI 0.0 SPI Alice ONLINE
61 Alice SPI 0.1 SPI Alice ONLINE
────── BLUETOOTH RADIOS (66-70) ──────
66 Alice BT BT 5.0 Alice hci0 Broadcom ONLINE
67 Cecilia BT BT 5.0 Cecilia hci0 Cypress ONLINE
68 Octavia BT BT 5.0 Octavia hci0 Cypress ONLINE
69 Aria BT BT 5.0 Aria hci0 Cypress OFFLINE
70 Lucidia BT BT 5.0 Lucidia hci0 Cypress ONLINE
────── NETWORK INTERFACES — PHYSICAL (71-80) ──────
71 Alice wlan0 WiFi 5GHz 192.168.4.49 -58 dBm ONLINE
72 Alice eth0 Ethernet 192.168.50.1/24 UP (secondary)
73 Cecilia wlan0 WiFi 5GHz 192.168.4.96 -64 dBm ONLINE
74 Cecilia eth0 Ethernet (→ BliKVM) 1Gbps UP (silent)
75 Octavia wlan0 WiFi 5GHz 192.168.4.100 (was .97) -63 dBm ONLINE
76 Octavia eth0 Ethernet — DOWN
77 Aria wlan0 WiFi 5GHz 192.168.4.98 OFFLINE
78 Aria eth0 Ethernet — DOWN
79 Lucidia wlan0 WiFi 5GHz 192.168.4.38 -62 dBm ONLINE
80 Lucidia eth0 Ethernet — DOWN
────── NETWORK INTERFACES — VIRTUAL (86-110) ──────
86 Alice uap0 RoadNet AP 10.10.1.1 CH1 ONLINE
87 Cecilia uap0 RoadNet AP 10.10.2.1 CH6 ONLINE
88 Octavia uap0 RoadNet AP 10.10.3.1 CH11 ONLINE
89 Aria uap0 RoadNet AP 10.10.4.1 CH1 OFFLINE
90 Lucidia uap0 RoadNet AP 10.10.5.1 CH11 ONLINE
91 Alice wg0 WireGuard 10.8.0.6 ONLINE
92 Cecilia wg0 WireGuard 10.8.0.3 ONLINE
93 Octavia wg0 WireGuard 10.8.0.4 ONLINE
94 Aria wg0 WireGuard 10.8.0.7 OFFLINE (~1h)
95 Anastasia wg0 WireGuard 10.8.0.1 (hub) ONLINE
96 Gematria wg0 WireGuard 10.8.0.8 ONLINE
97 Lucidia tailscale0 Tailscale 100.66.235.47 ONLINE
98 Alice docker0 Docker 172.17.0.1 ONLINE (linkdown)
99 Octavia docker0 Docker Swarm :2377 (leader) ONLINE
100 Aria docker0 Docker 172.17.0.1 OFFLINE
101 Lucidia docker0 Docker 172.17.0.1 ONLINE (12 containers)
102 Cecilia docker0 Docker 172.17.0.1 ONLINE (linkdown)
────── CLOUDFLARE TUNNELS (111-115) ──────
111 Alice Tunnel CF Tunnel 52915859 ONLINE 65+ hostnames
112 Cecilia Tunnel CF Tunnel d67bf4a5 ONLINE 22 hostnames
Routes: 11 domains→nginx:8080, cecilia→:3000, api→:8788,
gateway→:8787, ollama→:11434, monitor→:9100, dashboard→:3001,
agents→:4010, metrics→:9000, storage→:9001
113 Aria Tunnel CF Tunnel 93a03772 OFFLINE
114 Octavia Tunnel CF Tunnel b7e9f25e ONLINE 10 hostnames
Routes: lucidia→:3000, api-lucidia→:8000, monitor-lucidia→:9090,
ssh-lucidia→:22, git/code/roadcode→:3100, cloud(x2)→:3200
115 Lucidia Tunnel CF Tunnel 0447556b ONLINE 4 hostnames
Routes: octavia→:3000, api-octavia→:8000, monitor-octavia→:9090,
ssh-octavia→:22 ⚠️ NAMES SWAPPED with Octavia!
116 Anastasia Tunnel CF Tunnel (shellfish) ONLINE
────── LAN DEVICES — NON-BLACKROAD (121-140) ──────
121 Roku Stick Streaming 192.168.4.21 cc:08:fa:a4:49:c2 ONLINE
AirPlay-enabled (_airplay._tcp), Spotify Connect
122 AltoBeam Device IoT/TV 192.168.4.22 30:be:29:5b:24:5f ONLINE
AltoBeam Inc. (Chinese semiconductor — smart TV chip)
123 iPhone Phone 192.168.4.27 6c:4a:85:32:ae:72 ONLINE
124 Alexa's Apple TV Apple TV 192.168.4.33 60:92:c8:11:cf:7c ONLINE
AirPlay + AirTunes/377.40.00 :7000 + Companion Link
125 Funai DVD Player AV 192.168.4.44 98:17:3c:38:db:78 ONLINE
126 Private MAC Device IoT 192.168.4.45 d0:c9:07:50:51:ca ONLINE
Randomized MAC, all ports closed — phone or IoT
127 Unknown Device IoT 192.168.4.53 98:41:5c:aa:13:f2 DEAD
128 Unknown Device IoT 192.168.4.90 a0:4a:5e:2a:db:d2 DEAD
129 Phone Phone 192.168.4.95 f2:ff:5d:9a:f6:63 STALE
130 Phone Phone 192.168.4.99 2e:24:91:6a:af:a3 ONLINE
────── GPIO CONTROLLERS (141-155) ──────
141 Alice gpiochip0 GPIO Alice ONLINE
142 Alice gpiochip1 GPIO Alice ONLINE
143 Cecilia gpiochip0 GPIO Cecilia ONLINE
144 Cecilia gpiochip4 GPIO Cecilia ONLINE
145 Cecilia gpiochip10-13 GPIO (4) Cecilia ONLINE
146 Octavia gpiochip0 GPIO Octavia ONLINE
147 Octavia gpiochip4 GPIO Octavia ONLINE
148 Octavia gpiochip10-13 GPIO (4) Octavia ONLINE
149 Aria gpiochip0 GPIO Aria OFFLINE
150 Aria gpiochip4 GPIO Aria OFFLINE
151 Aria gpiochip10-13 GPIO (4) Aria OFFLINE
152 Lucidia gpiochip0 GPIO Lucidia ONLINE
153 Lucidia gpiochip4 GPIO Lucidia ONLINE
154 Lucidia gpiochip10-13 GPIO (4) Lucidia ONLINE
────── SERIAL PORTS (161-170) ──────
161 Alice ttyAMA0 Serial Alice ONLINE
162 Alice ttyS0 Serial Alice ONLINE
163 Cecilia ttyAMA0 Serial Cecilia ONLINE
164 Cecilia ttyAMA10 Serial Cecilia ONLINE
165 Cecilia ttyUSB0 Serial USB Cecilia (CP2102) ONLINE (SILENT)
166 Octavia ttyAMA0 Serial Octavia ONLINE
167 Octavia ttyAMA10 Serial Octavia ONLINE
168 Lucidia ttyAMA0 Serial Lucidia ONLINE
169 Lucidia ttyAMA10 Serial Lucidia ONLINE
────── PIRONMAN5 CASES (171-175) ──────
171 Octavia Pironman5 Case/OLED Octavia :34001 ONLINE (I2C 0x3c)
172 Aria Pironman5 Case/OLED Aria :34001 OFFLINE
173 Lucidia Pironman5 Case/OLED Lucidia :34001 ONLINE
174 Cecilia Pironman5 Case/OLED Cecilia :34001 ONLINE
────── OLLAMA MODELS (176-199 — logical devices) ──────
176 Cecilia Ollama AI Runtime :11434 (13 models) ONLINE
cece, cece2, qwen3:8b, llama3:8b, codellama:7b,
OpenELM-1B/3B, deepseek-coder:1.3b, qwen2.5-coder:3b,
llama3.2:1b/3b, tinyllama
177 Octavia Ollama AI Runtime :11434 (9 models) ONLINE
codellama:7b, apple-openelm-1B/3B, phi3.5, gemma2:2b,
llama3.2:1b/3b, qwen2.5:1.5b, tinyllama
178 Lucidia Ollama AI Runtime :11434 (6 models) ONLINE ⚠️ HOT
lucidia:8b (LOADED, 265% CPU!), qwen2.5:3b,
nomic-embed-text, llama3.2:1b, tinyllama, qwen2.5:1.5b
179 Anastasia Ollama AI Runtime :11434 (TS only) ONLINE
────── TAILSCALE GHOST NODES (181-185) ──────
181 codex-infinity Tailscale 100.108.132.8 OFFLINE 15d
Registered to amundsonalexa@ | linux | Unknown physical host
182 shellfish Tailscale 100.94.33.37 OFFLINE 15d
Registered to amundsonalexa@ | linux | Unknown physical host
183 lucidia-operator Tailscale 100.91.90.68 OFFLINE 33d
Registered to amundsonalexa@ | macOS | Stale — likely old Mac config
184 alice (TS) Tailscale 100.77.210.18 OFFLINE 15d
Should be Alice Pi — Tailscale not running on Alice
185 aria (TS) Tailscale 100.109.14.17 OFFLINE 15d
Should be Aria Pi — node is down
────── RESERVED (200-255) ──────
200-254 (reserved for future / dynamic DHCP)
255 Broadcast 192.168.4.255 ff:ff:ff:ff:ff:ff
═══════════════════════════════════════════════════════════════════════════════════
FLEET SUMMARY
═══════════════════════════════════════════════════════════════════════════════════
Total Registered: ~125 devices (IDs 1-199 allocated)
ONLINE: ~80
OFFLINE/MISSING: ~25
DEGRADED: 2 (SuperDrive, Gematria)
STALE: 5
┌───────────────────────────────────────────────────────────────────┐
│ NODE KERNEL TEMP RAM DISK LOAD │
├───────────────────────────────────────────────────────────────────┤
│ Alice 6.1.21 ⚠️OLD 40°C 443M/4G 77% 2.44 │
│ Cecilia 6.12.62 40°C 1.3G/8G 17% 0.99 │
│ Octavia 6.12.62 36°C 1.2G/8G 66% 0.40 │
│ Aria — — — 96%⚠️ — │
│ Lucidia 6.12.62 60°C✅ 3.9G/8G 41% <1.0 │
│ Anastasia 5.14.0 — 411M/765M 67% 0.02 │
└───────────────────────────────────────────────────────────────────┘
Enabled Services: 42 (Alice) + 44 (Cecilia) + 39 (Octavia) + 62 (Lucidia) + 28 (Anastasia) = 215 total
Docker Containers: 0 (Alice) + 0 (Cecilia) + 4 (Octavia Swarm) + 11 (Lucidia) + 0 (Anastasia) = 15
Docker Images: 0 (Alice) + 0 (Cecilia) + 11 (Octavia=11.2GB) + 14 (Lucidia=3.3GB) = 25 images
Ollama Models: 15 + 9 + 6 + ? = 30+ across fleet
Ollama Storage: 22GB (Cecilia) + 15GB (Octavia) + 20GB (Lucidia) = 57GB fleet
GitHub Actions: 4 runners (Alice, Octavia, Lucidia x2)
TCP Ports Open: 25 + 28 + 22 + 52 = 127 fleet-wide (198 with UDP)
Port Map: ~/roadnet/PORT-MAP.md (every port identified)
WireGuard Peers: 5 active (alice, cecilia, octavia, gematria) + 1 stale (aria)
Tailscale Nodes: 2 active (lucidia, alexandria) + 7 offline
Swap Total: 2G (Cecilia) + 2G (Octavia) + 8.5G (Lucidia) = 12.5GB fleet
Packages: 1926 (Alice) + 2197 (Cecilia) + 2290 (Octavia) + 2208 (Lucidia) = 8621
Python Packages: ? (Alice) + 408 (Cecilia) + 318 (Octavia) + 339 (Lucidia) = 1065+
Git Repos: 4 (Alice) + 30+ (Cecilia) + 4 (Octavia) + 10 (Lucidia) = 48+
All nodes have: /dev/watchdog0, /dev/rtc0, NTP synced, Docker v29.2.1
All Pi 5s have: DRM card0+card1+renderD128, conservative governor (optimized)
Connection Topology (who talks to who):
Alice → Mac (.28), Lucidia (.38:22 SSH)
Cecilia → Mac (.28), Google Drive (IPv6 rclone)
Octavia → Mac (.28), GitHub Actions (20.85.130.105), Cloudflare
Lucidia → Mac (.28), Alice (.49:44090), GitHub Actions, Cloudflare
CPU speeds: Alice 1.8GHz, Cecilia 2.4GHz, Octavia 2.6GHz, Lucidia 2.4GHz
NVMe: Cecilia + Octavia (both Micron/Crucial 5427 behind ASM1182e switch)
Undervoltage: Cecilia (9+ events) + Octavia (5+ events) — BOTH Hailo nodes!
⚠️ ALERTS (Layer 3 Deep Scan):
- Lucidia (#16) Ollama runner RESPAWNED at 250% CPU / 2.1GB RAM — needs permanent fix
- Lucidia SD card: dmesg "mmc0: Card stuck being busy!" — SD degradation warning!
- Lucidia DNS = Tailscale MagicDNS ONLY (100.100.100.100) — no fallback!
- Lucidia fail2ban FAILED — no intrusion protection despite UFW
- Lucidia certbot FAILED — no valid SSL certs
- Lucidia cmdline has cgroup params duplicated 4x — messy config.txt
- Cecilia ALSO has undervoltage! (9+ dmesg events — both Pi 5s with Hailo!)
- Cecilia now has 4 rclone instances (was 3!) — growing problem
- Cecilia vcgencmd broken (needs: sudo mknod /dev/vcio c 100 0)
- Octavia throttle improved: 0x50000 (was 0x50005) but voltage 0.9923V (low)
- Alice dmesg: brcmf_cfg80211_change_iface errors — WiFi/RoadNet AP conflict
- Alice kernel 6.1.21 from 2023 — 3 years old, needs update
- Aria (#15) DOWN — not pingable, needs physical reboot
- Aria SD (#31) at 96% — critical storage
- Cecilia timezone Europe/London — wrong, should be US
- Cecilia obfuscated /tmp/op.py cron ✅ REMOVED (was dropper pattern: exec from /tmp)
- Octavia pi user has 52 SSH keys — audit needed
- SuperDrive (#42) — USB over-current on Pi 5
- btc-compose.service = "Bitcoin Core + EPS" (failed docker compose)
- Failed logins on Lucidia: blikvm, admin, kvmd attempted from Mac
- Tailscale ghosts: codex-infinity, shellfish, lucidia-operator (offline)
- Octavia Ollama: qwen manifest corrupted (EOF), /api/chat returning 500s (2min timeouts)
- Octavia/Lucidia CF tunnel hostnames SWAPPED (octavia routes serve lucidia, vice versa)
- Cecilia dnsmasq crash loop: --bind-interfaces vs --bind-dynamic config conflict
- Lucidia swap usage growing: 1.3GB/8.5GB (was 437MB earlier in session)
- Lucidia llama.service: ConditionPathExists on /home/pi/models/model.gguf
- Alice iowait 15.32% on SD card — I/O bottleneck
- Lucidia has 14 skeleton Python microservices (:5000-:6300) wasting ~280MB RAM
- Lucidia Java HelloWorld on :8888 — test service left running
- Lucidia simpleweb Flask on :5000 — test service left running
- Octavia OctoPrint broadcasts mDNS on ALL 7 interfaces (Docker bridge leak)
- Alice Ollama is SSH-tunneled (:11434 → remote), not local
- Octavia STILL getting undervoltage after OC removal — PSU is the problem, needs 5V/5A
- Cecilia rpi-connect-wayvnc crash loop ✅ FIXED (masked both system + user service)
- Cecilia has Wolfram Engine 14.3 installed (4.9GB!)
- Cecilia has Bitcoin blockchain sync (6.2GB in alexandria-sync/.bitcoin-main)
- Cecilia Ollama now has 15 models (added deepseek-r1:1.5b + nomic-embed-text post-reboot)
- Cecilia PostgreSQL: empty (no user DBs, just templates)
- Alice PostgreSQL "blackroad" DB exists but has 0 tables
- Alice Qdrant: 0 collections (vector DB installed but unused)
- Alice Redis: 1 key in db0 (basically empty)
- Lucidia PostgreSQL: "operator" DB owned by pi
- Lucidia has 334 static apps in /var/www/blackroad/
- Lucidia /home/blackroad = 33GB (19GB runners, 4.4GB models, 2GB blackroad)
- Lucidia Docker stats: PowerDNS Admin 193MB, edge-agent 30MB, auth-gateway 22MB
- Lucidia 14 microservices (:5100-:6300) are HTML dashboard apps, not APIs
- Octavia Gitea: 207 repos across 7 orgs, 26 EMPTY repos (all in blackroad-os/roadchain)
blackroad-os: 127 repos (61MB, 20 empty) | lucidia: 14 (18MB, 1 empty)
platform: 29 (7MB) | roadchain: 21 (1.2MB, 4 empty) | infrastructure: 8 (1MB, 1 empty)
tools: 5 (0.6MB) | agents: 3 (0.4MB)
Largest: blackroad-operator 48MB, lucidia-metaverse 16MB
All updated 2026-03-09 (github-relay keeps them synced)
- Octavia NVMe Claude Time: config/ 2.1GB + time/ 800MB = 2.9GB
- Octavia InfluxDB: _internal + pironman5 (fan/temp data)
- All nodes InfluxDB: only pironman5 DB (identical)
- Lucidia 2 cloudflared instances: ports 20241 + 20242
- Lucidia SSL cert EXPIRED: blackroad.io expired 2025-10-04 (5 months ago!)
Cert was issued with manual DNS-01 challenge — can't auto-renew
python3-certbot-dns-cloudflare installed but NOT configured (cloudflare.ini is wrong)
LOW PRIORITY: Cloudflare tunnels handle TLS termination, nginx has no SSL listeners
- Cecilia github-relay.sh has HARDCODED Gitea creds (blackroad:BlackRoad2026OS)
- Cecilia github-relay.sh still points to Octavia .97 (now .100) — relay broken
- PUSH_SECRET exposed in plaintext crontabs on Alice, Cecilia, Octavia
- Lucidia has 21 GitHub Actions runner dirs (19GB!) — blackroad + pi users both running
- Octavia blackroad user has autonomy scripts (heartbeat 1m + heal 5m cron)
- Cecilia autonomy scripts: same heartbeat+heal pattern, checks stats-proxy+ollama
- Alice has task_watchdog.py (Python, monitors Redis worker heartbeats)
- Alice /opt/prism/ is EMPTY (git repo with no files, prism-pull timer pulls nothing)
- Alice blackroad-watchdog.timer runs every 30s (task queue monitoring)
🔒 SECURITY SUMMARY:
┌───────────────────────────────────────────────────────────────┐
│ NODE USERS SSH_KEYS NOPASSWD_SUDO FIREWALL │
├───────────────────────────────────────────────────────────────┤
│ Alice 6 53(pi) alice,pi nftables NAT │
│ Cecilia 5 2(br) cecilia,br nftables NAT │
│ Octavia 5 52(pi)⚠️ pi(default) nftables NAT │
│ Lucidia 9 15(oct) pi,octavia UFW+nftables │
│ Anastasia ? ? ? ? │
└───────────────────────────────────────────────────────────────┘
Notes: Lucidia is ONLY node with UFW (INPUT DROP policy)
Alice+Octavia have 50+ SSH keys on pi user — audit needed
Same PUSH_SECRET used across multiple nodes in plaintext crontabs
Cecilia github-relay.sh has hardcoded Gitea creds (blackroad:BlackRoad2026OS)
✅ Obfuscated cron on Cecilia REMOVED (was exec from /tmp dropper pattern)
✅ Cecilia rpi-connect-wayvnc crash loop MASKED
⚠️ Lucidia SSL cert expired Oct 2025 (renewal in progress)
✅ FIXED THIS SESSION:
- Octavia Hailo-8 (#22) — copied module+firmware from Cecilia, /dev/hailo0 now LIVE
- Lucidia temp — killed runaway Ollama runner (72°C → 66°C, now back to 74°C)
- Identified all LAN devices: eero router, Apple TV, Roku Stick, AltoBeam chip
- .53 and .90 confirmed DEAD (removed from active inventory)
- Cecilia timezone fixed: Europe/London → America/Chicago
- Cecilia dnsmasq fixed: bind-interfaces→bind-dynamic + listen-address→192.168.4.96 (NOW RUNNING)
- Cecilia vcgencmd fixed (sudo mknod /dev/vcio c 100 0)
- Mapped all 4 Cloudflare tunnel configs (Cecilia=22 hostnames, Octavia=10, Lucidia=4)
- Discovered Octavia/Lucidia tunnel hostname swap (lucidia routes on Octavia, vice versa)
- Cataloged all Docker images: 25 across fleet (Octavia 11.2GB, Lucidia 3.3GB)
- Found Octavia Ollama qwen manifest corruption + 500 errors on /api/chat
- Found Lucidia llama.service (llama.cpp server :8080) + lucidia.service (FastAPI :8000)
- Mapped Lucidia /var/www: 335 blackroad-* apps via nginx wildcard multi-site config
- Found Hailo kernel params identical on both nodes: force_desc_page_size=4096, support_soft_reset=Y
- Lucidia overheating FIXED: killed world-engine.py (alexa user service calling Ollama), disabled blackroad-world.service, 73.8°C → 57.9°C
- Octavia qwen manifest cleaned (already removed, remnant dir cleared)
- Found leaked GitHub PAT in Lucidia's alexa user blackroad-git-worker.service (gho_Gfu...)
- Found Lucidia alexa has 4 user services: blackroad-status, blackroad-api, blackroad-git-worker, blackroad-world (disabled)
- Tunnel hostname swap confirmed INTENTIONAL — routing is correct despite name confusion
⚡ POWER OPTIMIZATION (applied to all 4 nodes):
- CPU governor: ondemand → conservative (all nodes) — slower ramp = less power spikes
- Octavia: REMOVED overclock (over_voltage=6 + arm_freq=2600) + gpu_mem 256→16
- Cecilia: CPU capped 2.4GHz → 2.0GHz (Hailo headroom) + gpu_mem=16
- Octavia: CPU capped 2.6GHz → 2.0GHz (Hailo headroom) + gpu_mem=16
- vm.swappiness: 60 → 10 (all nodes) — reduces SD writes
- vm.dirty_ratio/background: tuned for batch writes (less I/O thrashing)
- Disabled unused services: lightdm, cups, cups-browsed, rpcbind, nfs-blkmap
- WiFi power management enabled (saves ~40mA per node)
- Power monitor deployed: /opt/blackroad/power-monitor.sh (cron */5)
- vcgencmd persistent fix: udev rule on Cecilia for /dev/vcio
- Leaked GitHub PAT removed from Lucidia blackroad-git-worker.service (was full repo+delete scope!)
- blackroad-git-worker.service disabled on Lucidia
- Cecilia rpi-connect-wayvnc crash loop FIXED (masked system + user service)
- Cecilia obfuscated /tmp/op.py cron REMOVED (dropper pattern executing from /tmp)
- Lucidia SSL cert renewal attempted (failed — needs Cloudflare DNS plugin config, LOW PRIORITY)
- Lucidia 16 skeleton microservices DISABLED (alert-manager, analytics, backup, compression,
connection-pool, fleet-monitor, grafana, load-balancer, log-aggregator, metrics,
monitor-api, notifications, perf-cache, resource-optimizer, tts-api, blackroad-web)
~350MB RAM freed, 16 fewer Python processes
- Lucidia Java HelloWorld + simpleweb DISABLED
- Cecilia github-relay.sh: IP fixed .97→.100, creds moved to ~/.github-relay.env (chmod 600)
- PUSH_SECRET removed from plaintext crontabs on Alice+Cecilia+Octavia, moved to /opt/blackroad/stats-push.env
- Gitea: 207 repos analyzed — 181 active, 26 empty (mostly blackroad-os placeholders)
- ⚠️ xmrig.service reference found on Lucidia (crypto miner unit — file gone but was configured)
═══════════════════════════════════════════════════════════════════════════════════
ID ALLOCATION PLAN
═══════════════════════════════════════════════════════════════════════════════════
1-10 Core infrastructure (router, cloud VPS)
11-20 Compute nodes (Pis, Mac, future nodes)
21-25 AI accelerators (Hailo-8, future TPU/GPU)
26-35 Storage devices (NVMe, SD, VPS disks)
36-50 USB peripherals (keyboards, mice, drives, serial)
51-55 Network peripherals (KVM, MCU)
56-65 I2C / SPI devices and buses
66-70 Bluetooth radios
71-85 Physical network interfaces (WiFi, Ethernet)
86-110 Virtual network interfaces (RoadNet AP, WG, Docker, Tailscale)
111-120 Cloudflare tunnels
121-140 LAN devices (non-BlackRoad)
141-160 GPIO controllers
161-170 Serial ports
171-175 Pironman5 cases
176-180 AI runtimes / Ollama instances
181-185 Tailscale ghost nodes
186-199 Reserved (future AI/TS)
200-254 Reserved (dynamic/DHCP/future)
255 Broadcast