What about STP with equipments not managing BPDU?

[aurorasanz]

Hello everybody,

I have been wondering about a issue with this situation. On a customer network, there are some servers connected to a switch Cisco, with portfast and BPDUguard at interfaces where servers are connected. The thing is that these servers have teaming configured between 2 interfaces connected to this switch. And there is also a Netapp cabinet which is connected and which has the same MAC on both interfaces (belonging to the timing).

The doubt I have is what can happen regarding loops since the switch sees the same MAC on both interfaces. This cabinet does not support BPDU so it is not going to be recognized by the switch as a dual connection. And the customer is having some problems and I don't know if this could be due to the same MAC. Here to prevent this, what could be done? I have configured LACP on both interfaces in order to convert it to the same virtual interface.

But is there anything else could be done?

For this scenario, or other as a hub/switch without BPDU, BPDU guard or BPDU filter does not give any solution, does it?

Thank you very much, I am really worried about this as I have been reading many discussions in this form and I have not found an answer to this particular situacion.

[Giuseppe Larosa]

Hello Aurora,

>> I have configured LACP on both interfaces in order to convert it to the same virtual interface.

You should be fine if the other device speaks LACP. With LACP you can build controlled bundles with LACP frames exchanged on each member link of the bundle.

For experience people managing the server have to configure their NICs to send and receive LACP messages.

You can use

show etherchannel ch# port

to see the state of the ports.

This issue is not strictly related to STP and STP BPDUs as STP does not attempt to learn MAC addresses on ports, but it just builds loop free L2 topologies by putting in STP blocking state few ports

I agree that without a proper bundling configuration the switch may see a MAC address flapping between the two ports and this could lead the network administrator to think of a possible bridging loop that does not exist.

Hope to help

Giuseppe