I've noticed that with the latest MaaS (v3.0.0, specifically the snap version), that if a remote rackd loses connection to the main regiond service, the the rackd will shutdown the DHCP server:
2021-08-10 22:20:42 ClusterClient,client: [info] ClusterClient connection lost (HOST:IPv6Address(type='TCP', host='::ffff:172.30.255.200', port=48606, flowInfo=0, scopeID=0) PEER:IPv6Address(type=
'TCP', host='::ffff:172.31.255.200', port=5250, flowInfo=0, scopeID=0))
2021-08-10 22:20:42 provisioningserver.rpc.clusterservice: [info] Failure on ping dropping connection to event-loop: maasv2:pid=1354
2021-08-10 22:20:42 ClusterClient,client: [info] ClusterClient connection lost (HOST:IPv6Address(type='TCP', host='::ffff:172.30.255.200', port=38118, flowInfo=0, scopeID=0) PEER:IPv6Address(type=
'TCP', host='::ffff:172.31.255.200', port=5253, flowInfo=0, scopeID=0))
2021-08-10 22:20:42 provisioningserver.rpc.clusterservice: [info] Failure on ping dropping connection to event-loop: maasv2:pid=1355
2021-08-10 22:20:42 provisioningserver.rpc.clusterservice: [info] Lost all connections to region controllers. Stopping service(s) dhcpd.
2021-08-10 22:20:42 ClusterClient,client: [info] ClusterClient connection lost (HOST:IPv6Address(type='TCP', host='::ffff:172.30.255.200', port=55878, flowInfo=0, scopeID=0) PEER:IPv6Address(type=
'TCP', host='::ffff:172.31.255.200', port=5251, flowInfo=0, scopeID=0))
2021-08-10 22:20:43 provisioningserver.rpc.clusterservice: [info] Failure on ping dropping connection to event-loop: maasv2:pid=1356
I was previously using 2.4.2 and didn't run into this issue; is there anyway to keep the DHCP service running in 3.0.0 if the connection to the main regiond is interrupted?