Questions re %FWM-2-STM_LEARNING_RE_ENABLE messages

[d-fillmore]

Hi - I get this cycle of message regularely on my N5K pair;

%FWM-2-STM_LOOP_DETECT: Loops detected in the network for mac bf80.5bee.2a60 among ports Eth101/1/19 and Eth101/1/20 vlan 120 - Disabling dynamic learn notifications for 180 seconds

%FWM-2-STM_LEARNING_RE_ENABLE: Re enabling dynamic learning on all interfaces

I'm almost certan that it's to do with a badly configured port channel at the server end and these have been ongoing for about a year.

I don't 100% understand the implications of the messages. The LOOP_DETECT message says that the switch is disabling dynamic learning fo 180 seconds. So the switch is not going to populate it's MAC add table with any new entries for 180 seconds.

Does the switch flush it's existing MAC address table at any stage?

If it does flush it, and dynamic learning is disabled, how can traffic pass through the switch?

Is anyone able to step through exactly what the switch does when this occurs?

Many Thanks in advance

Dom

[richbarb]

Look this:

http://www.cisco.com/en/US/products/ps9670/products_qanda_item09186a0080c11900.shtml

[d-fillmore]

Thanks - That gives a bit more information than the error message guides but doesn't answer my question

[Michael Dougherty]

I have seen this issue with a red hat server that was configured with a bond mode 5. In my instance the Ethernet interfaces are being used are attached to VFCs to support FCoE and can not be members of an ether-channel. So it was up to the server to make the LACP channel decision. I would get the same error referencing the VIP of the bond jumping between the 2 physical interfaces. Note the VIP MAC address will be the same as the physical address of the first NIC of the bond. The ultimate workaround was to set the bond to mode 1 active/passive interfaces.

Don't know if this would help in your case but it maybe worth testing. Hope this helps.

Sent from Cisco Technical Support iPad App