Symptom

When RSTP topology change event is detected on a vPC domain member switch and vPC domain members (vPC peers) are configured with the "peer-switch", the information about the topology change is not propagated between the vPC peers. As a result, the MAC address tables of the vPC peer switches are not properly flushed, and the topology change information is not correctly propagated via RSTP through the vPC members to other switches. Consequently, the stale entries in MAC address tables can lead to traffic blackholing and loss of connectivity. In particular, on the vPC domain member switch that has detected the RSTP topology change, the MAC address table entries pointing out the peer-link will not be flushed. Additionally, its vPC peer will not be informed about the topology change via RSTP, and will not propagate this topology change information via RSTP to other switches.

Conditions

1.) The vPC peers are configured with "peer-switch" 2.) The vPC peers are running Rapid-PVST 3.) A vPC domain member detects an RSTP topology change

Workaround

Remove the "peer-switch" from both vPC peers.

Further Problem Description