Updates:
I found the problem is actually because I deploy maas controller in the cloud, it does have a public ip, but not showed in the interface, and rack controller is trying to connect it’s internal ip
It was good on 2.5 beta version,
after upgradation, seems there are plenty bugs in provisioningserver.rpc.clusterservice,
First of all it doesn’t support https. which is already been report to Bug #1777712 “TLS doesn't work for inter-controller communicatio...” : Bugs : MAAS
Even I switch to http, I still got:
2018-12-19 11:49:30 provisioningserver.rpc.clusterservice: [info] Event-loop hk0-1-2-3-4-0:pid=17130 (::ffff:192.168.1.198:5251): User timeout caused connection failure.
2018-12-19 11:49:30 provisioningserver.rpc.clusterservice: [info] Event-loop hk0-1-2-3-4-0:pid=17132 (::ffff:192.168.1.198:5250): User timeout caused connection failure.
2018-12-19 11:49:30 provisioningserver.rpc.clusterservice: [info] Event-loop hk0-1-2-3-4-0:pid=17133 (::ffff:192.168.1.198:5252): User timeout caused connection failure.
2018-12-19 11:49:30 provisioningserver.rpc.clusterservice: [info] Event-loop hk0-1-2-3-4-0:pid=17134 (::ffff:192.168.1.198:5253): User timeout caused connection failure.
2018-12-19 11:49:31 provisioningserver.rpc.clusterservice: [info] Making connections to event-loops: hk0-1-2-3-4-0:pid=17130, hk0-1-2-3-4-0:pid=17132, hk0-1-2-3-4-0:pid=17133, hk0-1-2-3-4-0:pid=17134
2018-12-19 11:50:01 provisioningserver.rpc.clusterservice: [info] Event-loop hk0-1-2-3-4-0:pid=17130 (::ffff:192.168.1.198:5251): User timeout caused connection failure.
2018-12-19 11:50:01 provisioningserver.rpc.clusterservice: [info] Event-loop hk0-1-2-3-4-0:pid=17132 (::ffff:192.168.1.198:5250): User timeout caused connection failure.
2018-12-19 11:50:01 provisioningserver.rpc.clusterservice: [info] Event-loop hk0-1-2-3-4-0:pid=17133 (::ffff:192.168.1.198:5252): User timeout caused connection failure.
2018-12-19 11:50:01 provisioningserver.rpc.clusterservice: [info] Event-loop hk0-1-2-3-4-0:pid=17134 (::ffff:192.168.1.198:5253): User timeout caused connection failure.
I have no idea why the network address been resolved to ::ffff:192.168.1.198, while they should use ipv4.