[TriLUG] Slightly OT: network switch question

[Cristóbal Palmer via TriLUG]

On 5/15/18 3:00 PM, Matt Flyer via TriLUG wrote:
> I was reviewing some documentation on the previous failure and noticed
> that it got 'fixed' when someone pulled the Ethernet cable between the
> building switch and the POE injector.

I have on multiple occasions wasted more than an hour before testing the
physical layer. Always assume the hardware (switch, PoE injectors,
cables) is broken until you can prove it's not. This saves time in the
long run. Given the rate at which I bill for on-site work, buying cable
testing equipment and switches that can run self-diagnostic reports (or
switches so cheap/dumb that I can swap them out in under two minutes)
works out cheaper for the client, too.

If you're only testing RJ45/ethernet, there are multiple options from
multiple vendors under $15. I *highly* recommend testing your cables. I
solved a "flaky wifi" problem at a client site under a month ago by
testing the cable, finding it was failing, and re-crimping it at the end
where a WAP had been replaced that week. Imagine how frustrated they
would have been had I replaced the ~$200 WAP instead and not fixed the
problem.

In summary: cables are more likely to fail than you might guess, and you
should test them routinely.

-CMP