Diagnostics

Before searching, read what the system already tells you: it almost always names the machine, the register or the step at fault. If that is not enough, send the publisher two files: the venue PC’s logs and, for a device, its trace.

1. The signs to read first

Where	What you read there
Back office → Alertes (Alerts)	Each current problem, in a sentence: device offline, silent meters, uncertain transfer, failed backup or update… See Alerts and the Alerts catalogue.
The device’s screen	The missing step (PAS DE RESEAU, CHERCHE SERVEUR, SERVEUR REFUSE, MACHINE MUETTE…). See Devices and firmware.
Exploitation (Operations) → Boîtiers (Devices)	Firmware, addressing, last contact of each device; Faire sonner (Ring) to find the right one.
Paramètres (Settings) → Mises à jour (Updates)	The version, the latest result and the update service’s log.
Paramètres (Settings) → Sauvegardes (Backups)	The backup status, the last failure, the log.
Paramètres (Settings) → Plateforme (Platform)	The link with the platform: uploads waiting, last refusal and its message.
The register	Its banners: server in maintenance, server unreachable, register needs updating. See When the link drops.
Journal d’audit (Audit log)	Who did what, when. See Audit log.

The most frequent cases

Symptom	Look first at
A device does not appear	Its screen. SERVEUR REFUSE → Venue network. A ADOPTER → Exploitation (Operations) → Boîtiers en attente (Pending devices).
“boîtier hors ligne” (device offline)	Network cable, the device’s power supply; its screen.
“machine SAS muette (câble ?)” (SAS machine silent (cable?)) / MACHINE MUETTE	The cable between the device and the machine.
“compteurs muets” (silent meters)	The device talks to the server but no longer reports meters: link with the machine.
No register connects	Is the venue PC on? In maintenance (update, restore)? On the same network?
Nothing is saved, everything is greyed out	Paramètres (Settings) → Mises à jour (Updates): an update or a restore may be in progress.

2. The venue PC’s log file

On the venue PC, one command gathers into one file everything needed to understand a problem: the logs of the server’s services and extracts from the database for the last 7 days. It contains no password and no secret setting.

📍 Type on: Venue PC (Ubuntu)

sudo bash /opt/lucky-touch/current/deploy/collect-logs.sh

By default, the logs cover the last 48 hours; for another period, add for example --since 24h. At the end, the command shows the path of the file created, lucky-logs-<date>.tar.gz (in the folder of the version in service). Send it to the publisher. This file contains the venue’s operations: only pass it on to the publisher.

The update and backup service

To read the status of the service that installs versions and makes backups (current step, last error): sudo lt-updater status. If it no longer responds, restart the venue PC; otherwise call the publisher.

3. A device’s trace

When a problem affects a specific machine (transfer refused, pulses, the device’s network), the trace records everything the device does. The device behaves exactly as usual during this time.

1. Connect the device over USB to a Windows PC that has the flash package (see Devices and firmware).
2. In the package folder, start diagnostic mode, with the device’s port:

   📍 Type on: Windows PC (flash package folder)

   diagnostic.cmd COM5

   It installs the trace firmware, then opens the console, which records everything in the logs folder (serial-COM5-<date>.log).
3. Reproduce the problem on the machine.
4. Stop recording (Ctrl+C) and send this file, together with the venue PC’s log file (§2).
5. Go back to the normal firmware: flash.cmd COM5. The device keeps its settings and what it has learned.

On Linux, ./flash.sh /dev/ttyACM0 trace installs the trace firmware; the command then shows how to record the console.

Reading the trace (for a first idea): the lan:… and mqtt:… lines are about the network — mqtt:UP = the device is connected to the server; lan:server refused = another server answered in its place (see Venue network). The pulse:… lines concern coin machines: pulse:front = one pulse, pulse:xfer_refus followed by a reason = a refusal.

What to tell the publisher

• the venue, the machine (its number) or the register concerned;
• the time the problem happened, to the second if possible (for example 14:32:05);
• what was displayed (exact wording of the alert or refusal, device screen);
• the two files above.