Nearly every radio station uses Layer 2 Ethernet switches as part of their systems. Usually, they're just plug and play; but what about that occasion when there appears to be something wrong with an Ethernet connection? How do you go from having a hunch to really figuring out what the problem is?

In these examples I'll refer to Cisco switches, but clearly the same issues can manifest themselves no matter what brand of managed switch you are using.

To see what we want to see with a managed switch, we're going to refer to the console input, which is a serial connection for Cisco, made via an RJ-45 on the unit itself labeled "console." The easiest way to accomplish this is via the serial-to-RJ45 cable that came with the unit. On your computer, configure hyper-terminal (or Putty with Windows 7) for 9600 baud, 8-n-1. (Alternatively if this switch already has an IP address, you can telnet in to it.) If you use the console, when first connecting you'll see something like this:

Dougs_3550 con0 is now available

Press RETURN to get started

I have to assume for purposes of this narrative that either your console requires no password or else you know it. When you get the prompt

: switchname >

you'll be able to look at the statistics that I'll refer to below. As an example:

Dougs_3550>

If you now type in:

<show interfaces fastethernet 0/3 counters errors>

you will get an output that looks like this:

Dougs_3550>sh interfaces da0/3 counters errors
Port     Align-Err   FCS-Err   Xmit-Err   Rcv-Err   UnderSize
Fa0/3            0         0          0         0           0
Port    Single-Col   Multi-Col   Late-Col   Excess-Col   Carri-Sen   Runts   Giants
Fa0/3            0           0          0            0           0       0        0
Dougs_3550>

I used port three because I happened to plug my computer in there. Let's look at a few of these error counters and see what causes them.

Align-Err: These are usually the result of a duplex mismatch or a physical problem (such as cabling, a bad port, or a bad NIC).

FCS-Err: This is typically a physical issue (such as cabling, a bad port, or a bad Network Interface Card (NIC)) but can also indicate a duplex mismatch.

Now if you type in

<show interfaces fastethernet 0/3>

With Cisco you are presented with a large amount of information. You'll have to visually sort through it. One of the common issues between a switch port, and a connected device, is duplex mismatch. A symptom I have seen of this condition was the connected device simply 'freezing' up and needing a reboot. Excessively slow communication is another. When you operate in a half-duplex mode some data link errors are normal -- for example, errors noted in FCS (Frame Check Sequence), alignment, runts, and collisions. According to Cisco, a ratio of one percent errors to total traffic is acceptable. A degradation in performance can be noted if the error ratio exceeds 2 or 3 percent. However, if you are operating in the full-duplex mode, FCS, CRC (Cyclic Redundancy Checks), alignment, and runt counters must be minimal. If those counters are incrementing then check to make sure the connected device expects to communicate in the full-duplex mode (or change it).

A few other counters of interest:

Lost carrier/No carrier: the number of times the carrier (the signal coming in on the receive pair) was lost in transmission. If that count increments, check for a bad cable, and check the physical connection on both sides.

Output buffer failures: this can be a sign that the ports are run at an inferior speed and/or duplex, or there is too much traffic for the port.

---

Irwin is transmission systems supervisor for Clear Channel NYC and chief engineer of WKTU, New York. Contact him at doug@dougirwin.net.

---