Appearance
Clear past-retention orphans — keep live streams
Demo video (~44s) · Specter finds old orphan video chunks on Shadow 1 and Shadow 2 (past retention, never live streams or active pulls), plans a safe purge, deploys prune commands to both devices, and verifies free-space ACK. Audio-balanced re-render (Ara VO + Ghost Protocol US Anthem). Cloudflare Stream signed iframe.
Loading stream…
Watch for: retention scan · KEEP live · dual-device deploy · free space ACK.
Story beats
| # | Scene | Story |
|---|---|---|
| 01 | Hero | Disk pressure from stale orphans on S1/S2 |
| 02 | Detect | Orphan candidate table · retention scan · KEEP live |
| 03 | Plan | Confirm → Hermes queue → deploy → verify |
| 04 | Deploy | Purge commands stream to S1 and S2 · free space ACK |
| 05 | CTA | ghostprotocol.us |
Operator takeaways
- Orphans = past retention, inactive justified chunks — not live RTSP/VAPIX pulls.
- KEEP live is explicit in the plan before any delete.
- Same plan can land on both shadows with verify.
- Production disk hygiene still uses prune / soak gates and proxy
get_recording_status.
When to run this drill
| Context | Use |
|---|---|
| Soak prep | Free space before long gates |
| Training | Teach retention vs live media |
| Stakeholder demo | Show multi-device safe purge |
Related production commands
See Prune, storage & soak for proxy tool calls and shadow-prune-recordings flows. Deploy unlock also clears orphaned containers/sockets (process orphans)—different problem from video-chunk orphans: Deploy.