How Did That Happen?

I had a customer call and say the PoE board on a switch was down and that 7 APs were impacted. So I make the 2 hour drive to the customer’s site. These are good friends of mine so I don’t mind at all.

Upon arrival I ask the usual questions…

Storms?
Surges?
Reboot the switch?
Area impacted by the outage?

No weather issues. Yes, they rebooted the switch. Half the APs in the area are flashing?

Half are flashing? First thought is the PoE is fine. Why are half the APs acting up?

So we go to the IDF and look at the switch. At first glance the activity lights look great. All AP cables look normal. Time to grab the serial adapter and check the switch.

Time check PoE, just in case. Sh inlinepower showed that it was delivering PoE just fine. I execute sh conf vlan. At first glance it looks ok. I was mostly looking to make sure the switch had a config. I then execute sh conf port ge5.12-33. Whoa! Half the ports have a different PVID. They run the APs on the MGMT VLAN. I get that fixed and then look closer at the VLAN config. The ports for two downstream switches weren’t tagged for the MGMT VLAN. I fix those as well.

Everything comes back up and teachers say they are now working. I make sure to save config, twice. I’m kind of untrusting that way. I capture a copy of the config.

No idea why the config did that. It’s like it reverted to a previous version but that can’t be since the set port vlan command does them all at once, saying that you define all the ports.