This morning I had a customer network down due to a network loop, it turned out that the loop was caused by a single switch port were an AP1242 was connected to.
I discovered that the AP went down during the night due to the failure of the power supply, incredibly enough the switch port remained up and caused the loop.
The disconnection of the AP from the switch port immediatly solved the loop.
I tried to simulate the issue on an identical switch and found that connecting a non powered 1242 AP generates the following logs:
Apr 4 12:43:52.208: %LINK-3-UPDOWN: Interface GigabitEthernet1/0/47, changed state to up
Apr 4 12:43:53.211: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet1/0/47, changed state to up
Apr 4 12:43:57.028: %ETHCNTR-3-LOOP_BACK_DETECTED: Loop-back detected on GigabitEthernet1/0/47.
Apr 4 12:43:57.028: %PM-4-ERR_DISABLE: loopback error detected on Gi1/0/47, putting Gi1/0/47 in err-disable state
Apr 4 12:43:58.031: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet1/0/47, changed state to down
Apr 4 12:43:59.045: %LINK-3-UPDOWN: Interface GigabitEthernet1/0/47, changed state to down
At the end switch port is automatically err-disabled, maybe on the customer case loop occured before the switch had the chance to err-disable the port.
Original port is configured as a trunk, but tried many different configurations with same behaviour.
Tried to plug different devices and found that many 10/100 (poe) unpowered devices cause switch port to come up, but only AP1242 loops back.
Tried different switch models, but it seems that only 2960X suffer this issue.
Tried to disable keepalive on the port, this prevents switch port to be err-disabled, but I'm sure it not prevents loops.
Anyone experienced something similar?
Switch model is WS-C2960X-48TS-L
