[2.1] Rack(relay) Controller is rejected after upgrade to 2.1

Update
======

After upgrading a MAAS with 2 rack (relay) controllers, the Rack (relay) Controller running in the same machine as the Region, is being rejected and it is unable to connect to the region:

==> /var/log/maas/rackd.log <==
2016-09-01 10:33:44 [ClusterClient,client] Rack controller REJECTED by the region (via maas:pid=6263).
2016-09-01 10:33:44 [ClusterClient,client] ClusterClient connection lost (HOST:IPv6Address(TCP, '::ffff:192.168.9.2', 60620) PEER:IPv6Address(TCP, '::ffff:192.168.9.2', 5252))

==> /var/log/maas/regiond.log <==
2016-09-01 10:33:44 [RegionServer,204102,::ffff:192.168.9.2] Rack controller 'None' disconnected.
2016-09-01 10:33:44 [RegionServer,204102,::ffff:192.168.9.2] RegionServer connection lost (HOST:IPv6Address(TCP, '::ffff:192.168.9.2', 5252) PEER:IPv6Address(TCP, '::ffff:192.168.9.2', 60620))

Also, maas.log doesn't show a message of it being rejected, it only shows:

Aug 27 17:06:01 maas maas.rpc.rackcontrollers: [INFO] Registering existing rack controller 'maas'.

That said, the second Rack (relay) Controller that is running standalone is successfully connected to the Region (although, service tracking is not working, which is a known bug for alpha 1).

Original Report
===============
Upon upgrade to 2.1a1 I see repeated errors like this in the logs (every two seconds by the looks of it):

Aug 27 17:06:01 maas maas.rpc.rackcontrollers: [INFO] Registering existing rack controller 'maas'.