I'm seeing this error in rackd.log. There's no specific way to reproduce this error, but for context, I was simply running the CI at which point this error just appeared after all machines were deployed.
2018-07-27 00:41:46 provisioningserver.utils.services: [info] Starting beaconing for interfaces: {'virbr0'}
2018-07-27 00:41:46 provisioningserver.utils.services: [info] Beaconing process for virbr0 started.
2018-07-27 00:41:46 twisted.internet.defer: [critical] Unhandled error in Deferred:
2018-07-27 00:41:46 twisted.internet.defer: [critical]
Traceback (most recent call last):
Failure: twisted.internet.error.MulticastJoinError: (b'\xe0\x00\x00v', b'\n\xf5\x88\x06', 98, 'Address already in use')
2018-07-27 00:41:46 provisioningserver.utils.services: [info] observe-beacons[virbr0]: tcpdump: listening on virbr0, link-type EN10MB (Ethernet), capture size 16384 bytes
2018-07-27 00:41:52 provisioningserver.utils.services: [info] New interface monitoring state: {'ens3': {'neighbour': False, 'mdns': True}, 'ens4': {'neighbour': True, 'mdns': True}, 'virbr0': {'neighbour': True, 'mdns': True}}
2018-07-27 00:41:52 provisioningserver.utils.services: [info] Starting neighbour discovery for interfaces: {'virbr0'}
2018-07-27 00:41:52 provisioningserver.utils.services: [info] Neighbour observation process for virbr0 started.
2018-07-27 00:41:53 provisioningserver.utils.services: [info] observe-arp[virbr0]: tcpdump: listening on virbr0, link-type EN10MB (Ethernet), capture size 64 bytes
I believe this bug was caused because MAAS was picking up interfaces attached to bridges for the purpose of attaching VMs or containers, in addition to the bridges themselves.
That is, this should have been resolved with the following commit:
commit 842646142d42d95 453f4b64c209f93 b09f8aea7c
Author: Mike Pontillo <email address hidden>
Date: Fri Aug 24 21:08:12 2018 +0000
LP: #1788780 - Ignore tunnel interfaces (such as those created by hypervisors) when gathering interfaces for MAAS management.