SiloOS Operations Skill

Environment facts (read first)

• Pi = Raspberry Pi 4, hostname SiloOS, reachable at siloos.local (Wi-Fi wlan0; keep ethernet unplugged — see siloos-pi-recovery).
• Two systemd services, both enabled + Restart=always:
  • scale_bridge — Python bridge, WebSocket :8765.
  • silo-dashboard — Vite dev server (vite --host, dual-stack :5173).
• Access the dashboard at http://siloos.local:5173/.

Commands

Build Dashboard

• Command: cd dashboard && npm run build
• Purpose: Compile TypeScript and build the Vite production bundle.

Lint Dashboard

• Command: cd dashboard && npm run lint
• Purpose: Check for code style and logic errors in the frontend.

Sync to Raspberry Pi

• Command: ./sync-project.ps1
• Purpose: Synchronize local changes to the Raspberry Pi and restart services.

Manual Code Deployment (Pi)

1. Copy Bridge: scp -i siloos_key ble_bridge.py siloos@siloos.local:/home/siloos/ble_bridge.py
2. Copy Config: scp -i siloos_key config.json siloos@siloos.local:/home/siloos/config.json
3. Restart Service: ssh -i siloos_key siloos@siloos.local "sudo systemctl restart scale_bridge"
4. Verify Logs: ssh -i siloos_key siloos@siloos.local "sudo journalctl -u scale_bridge -n 20 --no-pager"

Run Dashboard (Dev)

• Command: cd dashboard && npm run dev -- --host
• Purpose: Start the development server with external access enabled.

Run Python Bridge (Local Debug)

• Command: python3 ble_bridge.py
• Purpose: Start the hardware bridge for local testing (requires BLE hardware).

Managing the services

ssh -i siloos_key siloos@siloos.local "systemctl status scale_bridge silo-dashboard --no-pager"
ssh -i siloos_key siloos@siloos.local "sudo systemctl restart silo-dashboard"   # after dashboard changes
ssh -i siloos_key siloos@siloos.local "sudo journalctl -u silo-dashboard -n 30 --no-pager"

If a service is ever missing (is-enabled → not-found), reinstall it with tools/install_bridge_service.sh (bridge) or re-run tools/pi_full_fix.sh (both).

Deploying over a flaky Pi link

Do NOT hand-run multi-step SSH commands when the link drops — they corrupt mid-stream. Push a script and launch it detached so it survives disconnects, then poll its log:

scp -i siloos_key tools/<script>.sh siloos@siloos.local:/home/siloos/
ssh -i siloos_key siloos@siloos.local "sudo systemd-run --unit=job --collect bash /home/siloos/<script>.sh"
# then: ssh ... "cat /home/siloos/<script>.log"   # poll for the completion MARKER

Tip: if siloos.local is flapping, SSH to the raw Wi-Fi IP (10.0.124.199) directly — flapping usually means the name is resolving to the dead .5 (eth0).

Deployment Procedure

1. Build local assets: npm run build
2. Sync to Pi using sync-project.ps1 or manual scp.
3. Restart the affected service (scale_bridge and/or silo-dashboard).
4. Verify: systemctl status scale_bridge silo-dashboard and load http://siloos.local:5173/.