ops: docker engine 29.8.2 on esh-ml1, nh3-ml1, nh3-dev, corviduo-dev; docs

docs/runbooks/fleet-guest-patch-2026-10-05.md

@@ -125,8 +125,10 @@ bouncer failing closed on slow LAPI answers. Self-cleared.
   `/health` after boot (startup, not a fault); wings on pfi-pteradactyl restores the running game server by itself.
 - ⚠ For every held engine bump: if `/etc/docker/daemon.json` has `live-restore: true`, an engine upgrade without a reboot leaves
   containers that bind-mount `/var/run/docker.sock` on a stale socket. Restart them afterwards (see servers/ana-docker/README.md).
-- Docker engine bumps held on nh3-docker, esh-docker-vm, corviduo-dev, nh3-dev, esh-ml1, nh3-ml1, pfi-pteradactyl
-  (ana-docker → 29.8.0 and pfi-gx10 → 29.6 done 10-05).
+- Docker engine bumps still held on nh3-docker (29.4.1; the bus + NH3 DNS), esh-docker-vm (29.4.1; ESH DNS + homepage
+  wedge risk) and pfi-pteradactyl (docker.io; the live game server). Done 10-05: ana-docker → 29.8.2 (bookworm repo),
+  pfi-gx10 → 29.6.2, and ~1410–1415 esh-ml1, nh3-ml1, nh3-dev, corviduo-dev → 29.8.2, each verified (same containers,
+  health endpoints, the embedding routes back on their primaries). NVIDIA/CUDA holds on the ml1s were kept.
 - fv-ml1: ZFS 2.3.9, nvidia-container-toolkit 1.20, and a kernel with an NVIDIA driver that builds on it.
 - CrowdSec bouncer `updateMaxFailure` (0 = block on the first failed LAPI pull): infra-ops leans toward ~3.