Now my test is such 192.168.200.1 swarm cluster manager
192.168.200.2 and 192.168.200.3 is work
Inside the 192.168.200.1 I create a service such as 200.1:8080

I sent you a request at this time: 200.1 8080 then the three nodes will do load balancing but if I come again a 192.168.200.4 joined the swarm of cluster, and then expands, make 200.4 with service, this time, I put the node 200.4 promoted to manager, who is so VIP IP ah: 200.1 or 200.4 8080:8080? How to schedule?