MAC flapping on Mikrotik with Ubiquiti APs

Scratching my head here. I have two Ubiquiti sectors sharing a bridge on a Mikrotik 493G. AP1 is in port 1 of the Mikrotik. AP2 is in port 2 of Mikrotik. BR0 consists of ports 1 and 2. MAC of AP1 is mapped to port 1 correctly. Intermittently though, in the bridge table of the Mikrotik, MAC of AP1 gets mapped to Port 2.

This happens for multiple (if not all) MACs downstream from AP1 (bridge mode).

TCPDUMP on AP2 shows frame with source MAC of AP1 just before the host entry changes and pings stop dropping. To begin with, these were CDP packets. I turned off CDP and Discovery and now the problem comes from Multicast Listener Report packets.

Downstream devices from AP1 consist of all kinds of SOHO devices so not only Ubiquiti.

When I was seeing the CDP packets, I saw them sometimes as single (which did not lead to a mac flap) and sometimes two at a time (which lead to mac flap). I assume this is a packet being looped which is causing the update to the bridging table on the Mikrotik.

What am I missing? Why do I see frames sourced from AP1 bouncing around or looping behind AP2?

The offending packets are these. When this sequence of packets occur, the port flaps.

21:16:51.270063 11:11:11:11:11:11 > 33:33:00:00:00:01, ethertype IPv6 (0x86dd), length 86: (hlim 1, next-header Options (0) payload length: 32) fe80::218:e7ff:fe8c:538a > ff02::1: HBH (rtalert: 0x0000) (padn)[icmp6 sum ok] ICMP6, multicast listener query, length 24max resp delay: 10 addr: ::

21:16:51.275740 22:22:22:22:22:22 > 33:33:ff:22:22:22, ethertype IPv6 (0x86dd), length 86: (hlim 1, next-header Options (0) payload length: 32) fe80::227:22ff:fe22:2222 > ff02::1:ff22:2222: HBH (rtalert: 0x0000) (padn)[icmp6 sum ok] ICMP6, multicast listener report, length 24max resp delay: 0 addr: ff02::1:ff28:89f9

This was taken on AP2. 22:22:22:22:22:22 is the MAC of AP1. 11:11:11:11:11:11 is the MAP of a client router on AP3.