Document dotminipc Docker capture node and its 2026-07-27 config-format outage

The upstream image switched from .env.local to TOML config, which silently
broke the dotminipc container; record what happened and where its config
lives now, since it shares the same upstream software as the Pi fleet.

Co-Authored-By: Claude Sonnet 5 <[email protected]>

README.md

@@ -10,6 +10,11 @@ Ansible playbooks for deploying MeshCore monitoring nodes (Raspberry Pi Zero W /
 | dm-ashwell | Pi Zero 2 W (armv7) | zero2_w |
 | dm-edworth | Pi Zero 2 W (armv7) | zero2_w |
 
+There's also a fourth capture node, **dotminipc** (device name `dm-stotfold`) —
+not managed by this repo (it's a Docker container on the shared dotnetwork
+host, not a dedicated Pi), but documented below since it's the same upstream
+software and shares config with the Pi fleet. See "dotminipc" below.
+
 ## Prerequisites
 
 **Local machine:**
@@ -72,7 +77,7 @@ You'll be prompted for a Tailscale auth key — leave blank if the node is alrea
 1. **wifi** — configures NetworkManager connections for `dotnetwork` (home), `dotmobile` (phone hotspot, field troubleshooting fallback), and the host's deployed-location network
 2. **base** — apt upgrade, installs screen/pipx/vnstat/git, sets MOTD, authorizes SSH keys for both laptop partitions, installs Tailscale (always) and authenticates it (only if `tailscale_auth_key` is set — otherwise run `sudo tailscale up` manually once, see checklist above)
 3. **meshcore_cli** — installs `meshcore-cli` via pipx
-4. **meshcore_capture** — runs the agessaman/meshcore-packet-capture install script (skipped once already installed — see "meshcore-packet-capture install is interactive" above, this needs a manual first run), writes `.env.local` config, enables `meshcore-capture.service`, deploys update/log helper scripts
+4. **meshcore_capture** — runs the agessaman/meshcore-packet-capture install script (skipped once already installed — see "meshcore-packet-capture install is interactive" above, this needs a manual first run), writes config (TOML `config.d/99-user.toml` on current "system"-layout nodes, legacy `.env.local` on the two nodes still on the old layout — see `meshcore_capture_layout` in `group_vars/all/vars.yml`), enables the capture service, deploys update/log helper scripts
 5. **scripts** — deploys `voltage.sh` and `bandwidth.sh`
 
 ## Config
@@ -91,3 +96,43 @@ ansible-vault edit --vault-password-file ../.vault_pass host_vars/dm-edworth/vau
 `*/vault.yml.example` shows the expected keys.
 
 `group_vars/meshcore.yml` (MQTT credentials) is still plaintext — consider moving it into the vault too if this repo is shared further.
+
+## dotminipc (Docker node, not managed by this repo)
+
+A fourth capture point, device name `dm-stotfold`, runs as a Docker container
+(`ghcr.io/agessaman/meshcore-packet-capture:latest`) on **dotminipc**
+(172.16.31.92), the shared Docker/HA host documented in the sibling
+`dotnetwork` repo. It isn't part of this repo's inventory — no Ansible role
+here touches it — but it's the same upstream software as the Pi fleet, so
+config drift between the two is worth knowing about.
+
+- Compose file: `dotnetwork/docker/dotminipc/meshcore-packet-capture/docker-compose.yml`
+- Config: `dotnetwork/docker/dotminipc/meshcore-packet-capture/config.d/99-user.toml`
+  — host-only, gitignored (contains MQTT credentials), same TOML shape as
+  this repo's `meshcore_capture` role template
+  (`ansible/roles/meshcore_capture/templates/99-user.toml.j2`); keep the
+  broker list in sync between the two if either changes. The committed
+  `config.d/99-user.toml.example` in the dotnetwork repo is the sanitized
+  reference copy.
+- Deploy/restart: `ssh [email protected]`, then from `/opt/docker`:
+  `docker compose up -d meshcore-capture --force-recreate`. Logs:
+  `docker logs -f meshcore-packet-capture`.
+
+**2026-07-27 outage, for reference**: the `:latest` image was rebuilt
+2026-07-25 and switched its config format from `.env.local` to TOML
+(`/etc/meshcore-packet-capture/config.toml` + `config.d/*.toml` —
+the same "system" layout `meshcore_capture_layout` already models for the Pi
+fleet). It silently stopped reading `.env.local` at all — despite upstream's
+own README still describing that as a supported "legacy" path, in practice it
+was just ignored — so the container fell back to the image's baked-in
+defaults (`connection_type = "ble"`, `serial.ports = ["/dev/ttyUSB0"]`, no
+owner key, only 2 of 5 brokers) and crash-looped on
+`could not open port /dev/ttyUSB0`. Fix was to mount a full TOML override at
+`config.d/99-user.toml` instead (see above) — same pattern as the Pi
+`meshcore_capture` role already uses for `system`-layout nodes. Also found in
+the process: the old `.env.local` had `PACKETCAPTURE_MQTT5_SERVER` set to
+`mqtt.meshmapper.cc`, which fails TLS handshake — the correct domain is
+`mqtt.meshmapper.net` (confirmed via `openssl s_client`; `.net` presents a
+valid Let's Encrypt cert, `.cc` returns a TLS alert). If the Pi fleet's
+`99-user.toml.j2` broker list is ever regenerated from a stale copy of this
+node's old config, watch out for that typo resurfacing.

ansible/group_vars/all/vars.yml

@@ -1,6 +1,10 @@
 remote_user: david
 ansible_user: "{{ remote_user }}"
-ansible_become: true
+# become is set at the play level in site.yml, NOT here — ansible_become as a
+# connection variable takes precedence over a task's own `become: false`
+# keyword, which silently broke every become:false task in this repo (they
+# all ran as root regardless). Play-level `become:` correctly loses to a
+# task-level override.
 
 meshcore_description: "meshcore management"
 

ansible/host_vars/dm-ashwell/vars.yml

@@ -3,6 +3,3 @@ serial_port: /dev/serial/by-id/usb-Espressif_USB_JTAG_serial_debug_unit_1C:DB:D4
 # Deployed-location wifi
 wifi_ssid_location: "awesome"
 wifi_password_location: "{{ vault_wifi_location }}"
-
-# Predates the current upstream installer's system-service layout
-meshcore_capture_layout: "legacy"

ansible/roles/base/tasks/main.yml

@@ -17,6 +17,15 @@
     state: present
   become: true
 
+- name: Install libffi headers (armv6 has no prebuilt cffi wheel, needs to compile it)
+  apt:
+    name:
+      - libffi-dev
+      - pkg-config
+    state: present
+  become: true
+  when: ansible_architecture == "armv6l"
+
 - name: Set MOTD
   template:
     src: motd.j2

ansible/roles/meshcore_capture/tasks/main.yml

@@ -29,7 +29,9 @@
 - name: Write .env.local config (legacy layout)
   template:
     src: env.local.j2
-    dest: "{{ ansible_env.HOME }}/.meshcore-packet-capture/.env.local"
+    dest: "/home/{{ remote_user }}/.meshcore-packet-capture/.env.local"
+    owner: "{{ remote_user }}"
+    group: "{{ remote_user }}"
     mode: "0640"
   become: false
   when: meshcore_capture_layout == 'legacy'
@@ -56,11 +58,12 @@
 
 - name: Deploy meshcore-capture-update script
   copy:
-    content: >-
+    content: |
       #!/usr/bin/env bash
-
       bash <(curl -fsSL https://raw.githubusercontent.com/agessaman/meshcore-packet-capture/main/install.sh){{ ' --update' if meshcore_capture_layout != 'legacy' else '' }}
-    dest: "{{ ansible_env.HOME }}/meshcore-capture-update.sh"
+    dest: "/home/{{ remote_user }}/meshcore-capture-update.sh"
+    owner: "{{ remote_user }}"
+    group: "{{ remote_user }}"
     mode: "0755"
   become: false
 
@@ -69,6 +72,8 @@
     content: |
       #!/usr/bin/env bash
       sudo journalctl -u {{ 'meshcore-capture' if meshcore_capture_layout == 'legacy' else 'meshcore-packet-capture' }} -f
-    dest: "{{ ansible_env.HOME }}/meshcore-capture-logs.sh"
+    dest: "/home/{{ remote_user }}/meshcore-capture-logs.sh"
+    owner: "{{ remote_user }}"
+    group: "{{ remote_user }}"
     mode: "0755"
   become: false

ansible/roles/meshcore_capture/templates/99-user.toml.j2

@@ -1,9 +1,10 @@
 # MeshCore Packet Capture - user overrides
 # Managed by Ansible - local changes will be overwritten on next playbook run
 #
-# [[broker]] tables merge by name across config.d/*.toml, so the letsmesh-us
+# Defines every broker fully (not relying on config.d/10-letsmesh.toml being
-# and letsmesh-eu entries below only add owner identity to the connection
+# auto-installed) — whether that preset exists depends on which path the
-# details already installed as config.d/10-letsmesh.toml.
+# installer took (fresh install vs migrate), so a partial override here would
+# silently no-op on nodes where it wasn't created.
 
 [general]
 iata = "{{ packetcapture_iata }}"
@@ -22,17 +23,45 @@ advert_interval_hours = {{ packetcapture_advert_interval_hours }}
 owner_public_key = "{{ packetcapture_owner_public_key }}"
 owner_email = "{{ packetcapture_owner_email }}"
 
+# LetsMesh.net Packet Analyzer (US)
 [[broker]]
 name = "letsmesh-us"
+enabled = true
+server = "mqtt-us-v1.letsmesh.net"
+port = 443
+transport = "websockets"
+keepalive = 120
+qos = 0
+retain = true
+
+[broker.tls]
+enabled = true
+verify = true
 
 [broker.auth]
+method = "token"
+audience = "mqtt-us-v1.letsmesh.net"
 owner = "{{ packetcapture_owner_public_key }}"
 email = "{{ packetcapture_owner_email }}"
 
+# LetsMesh.net Packet Analyzer (EU)
 [[broker]]
 name = "letsmesh-eu"
+enabled = true
+server = "mqtt-eu-v1.letsmesh.net"
+port = 443
+transport = "websockets"
+keepalive = 120
+qos = 0
+retain = true
+
+[broker.tls]
+enabled = true
+verify = true
 
 [broker.auth]
+method = "token"
+audience = "mqtt-eu-v1.letsmesh.net"
 owner = "{{ packetcapture_owner_public_key }}"
 email = "{{ packetcapture_owner_email }}"
 

ansible/roles/meshcore_cli/tasks/main.yml

@@ -11,5 +11,5 @@
   shell: pipx install meshcore-cli
   args:
     executable: /bin/bash
-    creates: "{{ ansible_env.HOME }}/.local/bin/meshcore-cli"
+    creates: "/home/{{ remote_user }}/.local/bin/meshcore-cli"
   become: false

ansible/roles/scripts/tasks/main.yml

@@ -2,13 +2,17 @@
 - name: Deploy voltage.sh
   copy:
     src: voltage.sh
-    dest: "{{ ansible_env.HOME }}/voltage.sh"
+    dest: "/home/{{ remote_user }}/voltage.sh"
+    owner: "{{ remote_user }}"
+    group: "{{ remote_user }}"
     mode: "0755"
   become: false
 
 - name: Deploy bandwidth.sh
   copy:
     src: bandwidth.sh
-    dest: "{{ ansible_env.HOME }}/bandwidth.sh"
+    dest: "/home/{{ remote_user }}/bandwidth.sh"
+    owner: "{{ remote_user }}"
+    group: "{{ remote_user }}"
     mode: "0755"
   become: false

ansible/site.yml

@@ -1,6 +1,7 @@
 ---
 - name: Deploy MeshCore monitoring nodes
   hosts: meshcore
+  become: true
   vars_prompt:
     - name: tailscale_auth_key
       prompt: "Tailscale auth key (leave blank to skip)"