wiki: nightly audit 2026-11-05 window — network-map re-verified by live probe (9377=Camofox not Stratos, 8790/5174 gone, +9120/9222/7777/6767/8085/9900, full Tailscale Serve table, two dead routes 8769/9123); commit pending 11-04 stocktake re-verifications; log + skills-index

This commit is contained in:
hermes
2026-11-05 01:22:45 +08:00
parent e4086ec18e
commit 3b5daccb4d
15 changed files with 942 additions and 552 deletions
+6 -3
View File
@@ -3,8 +3,8 @@ title: Nanobot
type: system
status: degraded
created: 2026-07-22
updated: 2026-10-12
verified_on: 2026-10-12
updated: 2026-11-04
verified_on: 2026-11-04
confidence: high
tags: [system, nanobot, network]
sources: [live-probe, omniroute-investigation]
@@ -18,7 +18,7 @@ Role in Anthony's setup.
## Current state
- **Status: DEGRADED (re-verified 2026-10-12).** The Nanobot web UI on CT333 is **UP again**: HTTP `200` on `http://100.122.189.81:8766/` serving the nanobot web UI (verified this audit). However port `8900` — the endpoint the OmniRoute `nanobot` provider connection (`58debd59`, prefix `nanobot`) uses — is still **closed**, so the OmniRoute integration remains disabled. Contrast with 2026-08-07, when connection was refused on 8900 and no service was confirmed listening anywhere.
- **Prior state:** `nanobot` process listening on port `8766` (verified via `ss -tlnp` on ct333, 2026-07-22); DOWN 2026-08-07 (connection refused on 8900 — OmniRoute event-loop stalls from model sync + credential health-check retries; provider connection disabled in OmniRoute).
- `whatsapp-bridge.service` (systemd, active) — "WhatsApp Bridge for Nanobot" — also runs on ct333. Status not re-verified on 2026-10-12 (SSH to CT333 not reachable from CT460: connection timed out on 22/tcp; verification above is TCP/HTTP probing only).
- `whatsapp-bridge.service` — "WhatsApp Bridge for Nanobot" — also runs on ct333. ⚠ **Corrected 2026-11-04: this unit is NOT active — it is in a crash-restart loop.** Live `systemctl status whatsapp-bridge.service` on CT333 shows `Active: activating (auto-restart) (Result: exit-code)` with `NRestarts=75514`. The `ExecStart` is `/bin/bash -c cd /home/nano/nanobot/bridge && BRIDGE_TOKEN=$(cat /home/nano/.nanobot/whatsapp-auth/bridge-token) exec node dist/index.js`, and it exits **`status=200/CHDIR`** — the working directory **`/home/nano/nanobot/bridge` does not exist** on CT333 (nor does `/home/nano/nanobot`; only `/home/nano/.nanobot/whatsapp-auth/bridge-token` is present). The unit is `enabled`, so systemd keeps restarting it every few seconds, burning a restart every ~3 s. **The old "systemd, active" claim was wrong and had been carried unverified since 2026-10-12.** Verified 2026-11-04 from inside the guest via `pct exec 333 -- systemctl status whatsapp-bridge.service` — CT333 is reachable from the PVE host even though it was not directly SSH-reachable from CT460 on 2026-10-12 (CT460 → CT333 port 22 now answers `TCP22_OPEN`).
- Publicly reachable via the same Cloudflare Tunnel as Hermes: `nanobot.martinwa.org` → `100.122.189.81:8766`, `nano-ssh.martinwa.org` → SSH on the same host. See [[infrastructure/domains-and-tunnels]].
## Current state as of 2026-07-22 (superseded — see below)
@@ -42,7 +42,10 @@ Role in Anthony's setup.
## Known issues
- Overlap/purpose vs. Hermes not documented anywhere — see [[comparisons/hermes-vs-nanobot]] for the open question this raises.
- **`whatsapp-bridge.service` restart loop (open since at least 2026-10-12, confirmed 2026-11-04).** `NRestarts=75514` and climbing; the cause is a missing `/home/nano/nanobot/bridge` working directory. Two ways to clear it, both needing a decision rather than a blind fix: **(a)** recreate the directory and redeploy the bridge build, or **(b)** `systemctl disable --now whatsapp-bridge.service` if the bridge is no longer wanted. Until one is chosen, the unit generates a restart every few seconds indefinitely. The bridge's own auth token (`/home/nano/.nanobot/whatsapp-auth/bridge-token`) is still present, so the deployment is not merely half-removed.
- **No monitoring covers either fault.** The OmniRoute `8900` port being closed and the bridge's restart loop both produce **no cron failure and no incident row** — the job reads green while the integration is down. That is the same class as the backup-task stall in [[incidents/2026-11-04-vzdump-freeze-ct299-sftpgo-outage]]: a service that fails *quietly* is invisible to a green-row fleet.
## Change history
- **2026-11-04: `whatsapp-bridge.service` claim corrected — it is crash-looping, not active.** The unit is in `activating (auto-restart)` with `NRestarts=75514`, failing `status=200/CHDIR` because `/home/nano/nanobot/bridge` does not exist. The page had carried "systemd, active" without verification since 2026-10-12 (that audit could not reach CT333 and said so). Status stays **DEGRADED** for the same reason as before — OmniRoute's port `8900` is still closed (live `ss -ltn` on CT333 shows no `8900` listener) — and now **also** because the WhatsApp bridge is down. The web UI on `:8766` remains up (HTTP 200). **Root cause of the loop is a missing deployment directory, not a service bug** — see [[systems/nanobot]] open items.
- 2026-07-22: audit corrected "not running" claim after direct verification on ct333.
- 2026-08-07: status changed to DEGRADED — nanobot confirmed DOWN (connection refused on 8900). OmniRoute provider connection disabled. CT333 service status unknown.
- 2026-10-12: web UI confirmed back up (HTTP 200 on `100.122.189.81:8766`, tailnet IP `100.122.189.81` = tailscale node `clawtest`). Status stays DEGRADED because the OmniRoute port 8900 endpoint is still closed. `whatsapp-bridge.service` not re-verified (SSH to CT333 unreachable from CT460).