HA Causes MAC address Flapping on Cisco Switch


#1

For high Availability, the Master LAN and Salve LAN are connected to two different Cisco Switches (Which in turn are interconnected via trunks)

LB1 (LAN) – (port-1)-SW1 (Trunk1) == (Trunk1)SW2-(port-5)–(LAN)LB2

This causes Flapping on cisco switch, as the same MAC address is seen on SW1 coming from port-1 and Trunk1
Same is happening on SW2, same MAC is seen on interfaces: port-5 and Trunk1
This makes setup unstable, and the two Peplink devices are fighting who will be the master.
Any solution is welcome!


#2

Master and slave are using same virtual MAC address on LAN port after HA was enabled and before they communicate with each other.

After master and slave are managed to communicate, MAC address for slave will be changed back to its physical MAC address.

I suspect master and slave lost communication for your case.

  1. Sw1-port1 was configured as Access port? If so, what is the Vlan ID?

  2. Sw2-port5 was configured as Access port? If so, what is the Vlan ID?

  3. Which Vlan ID you allowed in the trunk port on both Sw1 and 2?


#3

I agree with you. There must be somewhere traffic is getting blocked in the path between the two load balancers.
I checked with the customer, and they have a Core Switch (SW3) in between the two Switches (SW1 & SW2). Maybe something is blocked at that level.
Both SW1-(port-1) and SW2-(port-5) are Access Ports in VLAN-ID=10. Trunks are allowing this VLAN, and this can be witnessed by the fact that each switch (SW1 and SW2) is seeing the same MAC address from two interfaces (which proves the path is open between the two switches). What is the communication traffic between Load-balancers that need to establish understanding about Master/Salve?
Can it be that Multicast filtering is in between?


#4

We are using VRRP. You may find here for the details.