Commit Graph
2 Commits
Author SHA1 Message Date
jschoubben 991e33f749 tautulli: reach plex through the mesh, written before Tautulli starts
Tautulli reached plex at 172.18.0.1, the gateway of a HAL network that goes
away with HAL, and the plan was to retype it by hand in the window. Tautulli
now requires plex-api, and where plex is comes from the binding.

Tautulli keeps the connection only in config.ini, reads it at start and
writes its whole config back on every shutdown; its API cannot set it and
its settings form needs an admin login. So a step after start would be
overwritten the moment the container is recreated. The write is made where
nothing can overwrite it: the linuxserver image's custom-init runs
plex/mesh-plex.py as root before Tautulli starts, and the server restarts on
its binding and credential, so a moved plex or an accepted token lands.

It writes only [PMS] keys, only when they differ, every other line byte for
byte: pms_ip, pms_port, pms_ssl and pms_url from the binding; pms_identifier
from plex's /identity; pms_token only when plex takes it. A minted value -
before the operator accepts the server's X-Plex-Token for this pair - is
never written, while the address still is, so Tautulli's own working token
keeps working at plex's new address.

A failure in custom-init is a log line nobody reads, so a run-once `plex`
step, declared last so it gates nothing (ADR 0136), checks what the mesh can
report: plex takes the credential (else it names the secret accept), Tautulli
holds the bound URL, and Tautulli says it is connected. It writes nothing.

The script is kept as plex/mesh-plex.py and plex/50-mesh-plex; module.json
carries copies, and a test fails when they differ. Tests run the script with
python3 against a fake plex (skipped where there is none) and the step
against fakes; `npm test` builds first.
2026-09-30 13:09:43 +02:00
jschoubben e362fb951c The rest of the catalogue becomes mesh-buildable (issue 060, batch 2)
The 21 media/home modules, the SaaS tool modules (cloudflare-dns,
confluence, gitlab, jira), model-usage, and the model-access trio get
the same Dockerfile + build section as batch 1. Scheduled-only modules
(anthropic-consumer, anthropic-manager, openai-consumer) deliberately
declare no MESH_TOOL_MODULES — every container of theirs names its
command. mosquitto's run-once bootstrap container builds from the same
artifact. Modules with third-party deps (model-usage: pg;
anthropic-manager: tweetnacl) install them beside their compiled code.

Also fixes cloudflare-dns's package.json, unparseable since its
description lost a closing quote.

Deliberately still without build sections: builder and mesh-controller
(the foundation builds them by its own path), distribution (provides
the artifact store — building it through itself is refused by design),
route-proxy (cross-repo build context, deferred), and the
upstream-image-only modules, which have no code to build.
2026-09-18 02:14:09 +02:00