We have a PortChannel that has been up and running for several years. Initially it was between a Nexus 7010 and two Catalyst 2960S stacked switches. Now it is between the same Nexus 7010 and a stack consisting of a 2960X and two 2960S switches.
Three times in the calendar year 2015, we have had this portchannel go into an errdisable state. The first time, it seemed like a fluke. After the second time, approximately one week later, we increased the log retention. It happened again five weeks later. This is what we found in the logs:
May 18 04:54:25.164: %ETHCNTR-3-LOOP_BACK_DETECTED: Loop-back detected on
GigabitEthernet2/0/52. (chipedstk01-2)
May 18 04:54:25.164: %PM-4-ERR_DISABLE: loopback error detected on Gi2/0/52,
putting Gi2/0/52 in err-disable state (chipedstk01-2)
May 18 04:54:26.167: %LINEPROTO-5-UPDOWN: Line protocol on Interface
GigabitEthernet2/0/52, changed state to down
May 18 04:54:26.167: %LINEPROTO-5-UPDOWN: Line protocol on Interface Port-channel1,
changed state to down
May 18 04:54:27.177: %LINK-3-UPDOWN: Interface Port-channel1, changed state to down
May 18 04:54:27.205: %LINK-3-UPDOWN: Interface GigabitEthernet2/0/52, changed state
to down
However, preceding this event was this:
May 17 18:17:37.926: %SW_MATM-4-MACFLAP_NOTIF: Host eca8.6bf7.7147 in vlan 1 is
flapping between port Po1 and port Gi2/0/24
The device with MAC eca8.6bf7.7147 is a Windows 7 device which actually sits on a shelf in the rack with the switches and projects an image to a monitor elsewhere in the building. My thinking is that an artificial "Loop" is detected because the switch is detecting the MAC address locally on the switch and also on PortChannel 1. But Why?
Here is the Nexus configuration:
interface port-channel6
description P IDF
switchport
switchport mode trunk
switchport trunk allowed vlan 1,4,140
interface Ethernet4/21
description move P1-P2 (yellow)
switchport
switchport mode trunk
switchport trunk allowed vlan 1,4,140
channel-group 6 mode active
interface Ethernet4/24
description P5-P6 (Blue)
switchport
switchport mode trunk
switchport trunk allowed vlan 1,4,140
channel-group 6 mode active
Here is the switch stack configuration:
interface Port-channel1
description Data Center
switchport trunk allowed vlan 1,4,140
switchport mode trunk
end
interface GigabitEthernet2/0/52
description P IDF 1-2 to Data Center Wall
switchport trunk allowed vlan 1,4,140
switchport mode trunk
channel-group 1 mode active
end
interface GigabitEthernet3/0/52
description P IDF 5-6 to Data Center Wall
switchport trunk allowed vlan 1,4,140
switchport mode trunk
channel-group 1 mode active
end
How do I make this trunk more stable? I have temporarily enabled errdisable recovery for loopback on the 2960 switchstack.
Note: The Nexus ports never go into errdisable state.
