Someone told me of a failure scenario that occurred with VPC where I don’t understand how it ever could have happened.
We have a host that connects to vpc Pair. So we create a port-channel down to the host from switch 1 and include 1 physical port, rinse and repeat on the second switch.
So now we have 2 port channels with the same ID on 2 switches and the of course the same vpc ID. We fat finger the configuration and we forget to add spanning-tree type edge to the port channels.
The servers sends a BPDU to the switch and the vpc peer switch puts that vpc member in blocking state. Does that also affect the port channel on the other switch. They have the same ID
In this case port type edge is not configured and also not BPDU guard.
Second scenario is that both vpc member port channel do have guard enabled and a BPDU hits the port channel. The interface will go err dis and port channel down. At that time the other port is orphan and regular STP counts.
Am I correct?
My colleague is saying that the BPDU could affect the whole vpc coming apart and both member interfaces going down.
No comments:
Post a Comment