Diagnostics and troubleshooting

Software 3.4 · reviewed 2026-08-21

The order to work in

Four questions, in this order, before touching anything. Most calls are answered by the second.

Is the unit running? Power LED on, status LED steady green. If not, go to LED patterns and stop reading here.
Is it reachable? From a normal workstation, not from the service port. If the service port works and the LAN does not, the problem is addressing or cabling, not the unit.
What does it think is wrong? mluctl status then mluctl alert list --open. The unit usually knows.
Did anything change? mluctl log grep -i change --since 72h. Firmware on the printers, a VLAN change, a new firewall rule, a DHCP scope edit. Something changed; it always has.

LED patterns

Power	Status	State	Do
green	steady green	Running normally	Nothing.
green	blinking blue	Booting, or an update in progress	Wait. Up to 6 minutes during an update.
green	slow amber	Running, with an open critical alert	Read the console.
green	steady amber	Started into recovery — the main software did not come up	Recover from USB.
green	fast red	Storage fault	Take a bundle if you can, then replace the unit.
green	off	Powered but not booting	Serial console. If there is no output at all, the unit is dead.
off	off	No power	Supply, barrel jack, or the PoE switch port on an MLU-1P.

Symptom table

Symptom	Most likely	Check
Console unreachable on the LAN, fine on the service port	Address conflict, or the LAN patched into eth1	mluctl net show; look at which interface has a link.
All devices unreachable at once	Firewall rule or VLAN change	mluctl net test <printer> from the unit.
One device unreachable, prints fine	SNMP disabled on that device, often after a firmware update	The device’s own web page, SNMP section.
Meters stopped at a date, devices reachable	Collection service stuck	mluctl log tail --service collect, then restart it.
New printers never appear	Address outside every scan range	mluctl scan range list.
Jobs held but never print when released	Outbound 9100 blocked, or the device is in an error state	Alert MLU-402; try one job to one device.
Nothing reaches the queue at all	Workstations pointing at an old address	Print a test from one workstation and watch --service queue.
Reports missing a day	Unit was off, or the clock moved	mluctl meter history <device> --flags.
Supplies all show unknown after an update	Profile match dropped to generic	mluctl device list --profile generic.
Console slow, collection late	Storage nearly full	mluctl status; look at held jobs first.

The three faults we actually see

SNMP turned off by a printer firmware update. By far the most common. One device goes quiet, everything else is fine, and nobody connects it to the update that ran at the weekend. Re-enable SNMP on the device and, where the platform allows, record the community again — some updates reset it to default.
The two Ethernet sockets swapped. Someone tidied the cabinet. Symptoms are total: no console from the LAN, no discovery, no collection, and the unit looks dead from every direction except the service port — which now needs the cable that is in the wrong socket. Label the cables at install and this fault disappears.
A full queue pretending to be a broken unit. Storage fills with held jobs, the queue stops accepting, users say “printing is down”, and meter collection carries on perfectly. MLU-303 and MLU-304 were both raised days earlier and went to an unread mailbox.

Diagnostic bundle

Collect the bundle before rebooting. A reboot clears the state that explains the fault, and the second occurrence is often weeks away.

mluctl diag bundle --out /tmp/mlu-bundle.tgz
# or straight to a stick, when the unit has no route anywhere
mluctl diag bundle --to usb

Contains	Service logs (7 days), configuration, device inventory, interface and route state, a self-test, the last 200 alerts
Does not contain	Held job data, passwords, SNMP credentials, card numbers
Typical size	4–20 MB
Send to	the documentation desk, with the site reference and what you were doing

Recovery from USB

For a unit that boots to steady amber: the main software did not start, and the recovery image did. The unit is alive and its data is intact.

Write the recovery image for the installed version to a FAT32 stick, in the root, keeping its filename.
Insert it into the USB-A socket on a powered unit. The status LED blinks blue within ten seconds.
Wait. Five to nine minutes, no power interruption, no removing the stick.
Steady green means it worked. Remove the stick and sign in from the service port.
Check that the configuration survived: mluctl status and mluctl device list. If the inventory is empty, restore the backup.

Recovery keeps data, a factory reset does not. The reset pinhole next to the USB socket, held for 10 seconds, erases device records, meter history and held jobs. It is for a unit leaving a customer site, and for nothing else.