check layer 1

networkingIT support

The first commandment of troubleshooting: before anything clever, confirm the cable, the port, the power.

The OSI model doubles as a debugging ladder, and an embarrassing share of 'routing problems' end at an unplugged patch lead or a dead SFP. Start at the bottom; the physical layer has no sense of humor.

Check layer one is the reminder that a great many problems presented as complex are a cable, a port, a transceiver, or power. It is said with a certain weariness because it is said after someone has spent an hour on the routing table.

The reason the physical layer gets skipped is that it is boring and invisible from a terminal. An engineer diagnosing remotely sees interface counters and configuration, both of which are abstractions over something physical that may be half-seated, kinked behind a rack, plugged into the neighbouring port, or an optic operating at a wavelength the far end does not use. Error counters incrementing steadily are the tell, and they are frequently on a screen nobody looked at.

The discipline is to check it first rather than last, because it is cheap to check and expensive to skip. Is the link up, is it at the expected speed and duplex, are errors incrementing, does the optic report sensible light levels, is it in the port the documentation claims. Five minutes of that eliminates a large fraction of possibilities, and the alternative is discovering at midnight that the answer was a cable somebody moved during the day.

Also known as: layer 1 problem, is it plugged in

← All glossary entries