Solo mode¶

A personal, standalone copy of the gateway for a machine that has no access to the live ModernWMS/ Part-DB docker containers — a second Arch box at home, say, rather than the server those containers actually run on. It is not a lighter version of the test instance; it is for a genuinely different situation:

	Test instance	Solo mode
Where it runs	The same box as live, or one with access to the same containers	Any machine, no containers needed at all
Sign-in	Your real ModernWMS/Part-DB account	A local password, set just for this copy
Reachable from	`127.0.0.1` only (SSH tunnel to reach it from elsewhere)	Its own IP and port, like live
LEGO collection	Fresh public catalog, empty collection	Same: fresh public catalog, empty collection
Part-DB parts / ModernWMS-only screens	Work normally (same containers)	Not available — see below

If you do have access to the ModernWMS/Part-DB containers (you're setting up a second copy on the same server, or one that can reach them), use the test instance instead — it's simpler and reuses your real login.

What's different¶

ModernWMS is only ever reachable via docker exec into its container — there is no network path to it. Part-DB's own reads are a local SQLite file; only its writes have a real HTTP API. Neither of those is available on a machine with no access to the containers, so solo mode:

Signs in against a password stored in access.json instead of either backend (internal/access's PasswordHash field, bcrypt, same scheme as the remote-bot PIN). Set one with:
```
wms access user set-password <name> --source local
wms access user set-groups "local:<name>" admin
```
(omit --password to get one generated and printed once). This account is governed by the same permission groups, 2FA policy, channels and persisted login lockout as any other — it just isn't ModernWMS or Part-DB underneath.
Starts even when there's no Part-DB SQLite file to open, rather than aborting — every Part-DB screen/API call then returns a clear "Part-DB not available (solo mode)" error instead of working, so a Part-DB parts lookup fails cleanly rather than taking the whole app down.
ModernWMS-only screens (warehouse ops, stock, orders, user management) behave the same as they would with an unreachable container today: a normal "docker exec failed" error, not a crash. The LEGO side of this app (sets, parts, checks, bagging — internal/lego, its own SQLite file) was always independent of both backends and works exactly as it does live.

Setting it up¶

bash scripts/solo-instance.sh kc      # first run: builds, creates account "kc", seeds the public
                                       # catalog, starts the gateway
bash scripts/solo-instance.sh status
bash scripts/solo-instance.sh stop

No root, no systemd, no containers — it builds into and runs entirely out of ~/.local/share/wms-solo (override with WMS_SOLO_DATA_DIR), never touching /usr/local/bin/wms-go, /root/docker-server/wms, or any systemd unit. Telnet defaults to port 2325 and the web terminal to 7696 (both overridable — WMS_SOLO_TELNET_PORT, WMS_SOLO_WEB_PORT), bound to 0.0.0.0 by default (WMS_SOLO_HOST) rather than the test instance's hardcoded loopback — the point of this one is being reachable from elsewhere on your own network.

The sign-on screen's message-of-the-day says *** PERSONAL SOLO COPY ***, so it's never mistaken for the real one.