The additional_contexts approach required cc-platform-core to exist on
the Docker host at a relative path. This fails on servers where the
repo layout differs. Instead, use a multi-stage build: stage 1 clones
cc-platform-core from Gitea (depth 1), stage 2 copies server/ into the
app and rewrites the file: path. Fully self-contained — no host deps.
Use additional_contexts to copy platform server package into the Docker
build context. Rewrites the file: dependency path and removes the
lockfile so npm install can resolve the local package correctly.
- Entrypoint: chown -R /data (not just the directory) so existing
volume files owned by root become writable by node user
- Entrypoint: add echo logging so startup progress is visible
- db.js: verify /data is writable before opening SQLite
- db.js: wrap Database() constructor in try-catch with clear error
message instead of crashing silently at ESM import time
The Opus package manager deletes the entire package directory on
update (fs.delete(packageDir)) before re-downloading files. This
wiped all config files (.manager_config, .client_config, etc.) that
were stored inside packages/inventory-manager/.
Fix: add _configPath() helper to every program that resolves config
file paths to usr/config/inventory-manager/ when running under Opus,
which lives outside the package directory and survives updates.
Falls back to the local _path() for standalone (non-Opus) use.
Updated files:
- inventoryManager.lua, manager/config.lua, inventoryClient.lua,
inventoryWebBridge.lua, dropperController.lua, craftingTurtle.lua
- .package install script: saves configs to usr/config/inventory-manager/
- autorun/startup.lua: checks both persistent and package dirs
- startup/client.lua: uses persistent dir for .client_setup/.dropper_config
- web/server/Dockerfile: switch health check to wget (from prior fix)
Spawning a full Node process for health checks is slow on Alpine and
can exceed the 3s timeout, causing the container to be marked unhealthy.
Use wget (built into Alpine) instead, and increase start_period to 15s.
- Add entrypoint script that ensures /data is owned by node user
before dropping privileges with su-exec
- Remove USER node from Dockerfile (entrypoint handles it)
- Change client depends_on to service_healthy so nginx waits for
the server to pass its healthcheck before starting
