我确信这个答案就在某个地方，但经过几次尝试后我无法找到或修复它。这是用例：
1.> 我有两个 ec2 实例属于同一个 VPC 但具有不同的安全组
2.> 两个安全组都有 22,80（公共）和来自所有为 CIDR 块 10.20.0.0/16 开放的端口的所有流量
3.> EC2实例的内部IP为10.20.0.51（server-1）和10.20.0.202（server-2）
4.> 我正在使用以下命令在它们上运行两个 dockerized consul 服务器
server-1 : docker run -it -p 8400:8400 -p 8500:8500 -p 8600:53/udp -p 8301:8301 -p 8300:8300 -h node1 progrium/consul -server -advertise 10.20.0.51 -bootstrap-expect 2
server-2 : docker run -it -p 8400:8400 -p 8500:8500 -p 8600:53/udp -p 8301:8301 -p 8300:8300 --name node2 -h node2 progrium/consul -server -advertise 10.20.0.202 -join 10.20.0.51
5.> 他们都开始了，一秒钟后他们互相认识，选举发生，第一个节点被选举，但不久之后 server-2 开始说“成员列表：怀疑节点 1 失败，没有收到确认”和 server-1还说“成员列表：怀疑节点 2 失败，未收到任何确认”
这是 server-1 的日志的样子
2016/01/04 19:18:35 [INFO] serf: EventMemberJoin: node2 10.20.0.202
2016/01/04 19:18:35 [INFO] consul: adding server node2 (Addr: 10.20.0.202:8300) (DC: dc1)
2016/01/04 19:18:35 [INFO] consul: Attempting bootstrap with nodes: [10.20.0.51:8300 10.20.0.202:8300]
2016/01/04 19:18:35 [WARN] raft: Heartbeat timeout reached, starting election
2016/01/04 19:18:35 [INFO] raft: Node at 10.20.0.51:8300 [Candidate] entering Candidate state
2016/01/04 19:18:35 [WARN] raft: Remote peer 10.20.0.202:8300 does not have local node 10.20.0.51:8300 as a peer
2016/01/04 19:18:35 [INFO] raft: Election won. Tally: 2
2016/01/04 19:18:35 [INFO] raft: Node at 10.20.0.51:8300 [Leader] entering Leader state
2016/01/04 19:18:35 [INFO] consul: cluster leadership acquired
2016/01/04 19:18:35 [INFO] consul: New leader elected: node1
2016/01/04 19:18:35 [INFO] raft: pipelining replication to peer 10.20.0.202:8300
2016/01/04 19:18:35 [INFO] consul: member 'node1' joined, marking health alive
2016/01/04 19:18:35 [INFO] consul: member 'node2' joined, marking health alive
2016/01/04 19:18:37 [INFO] memberlist: Suspect node2 has failed, no acks received
2016/01/04 19:18:37 [INFO] agent: Synced service 'consul'
2016/01/04 19:18:39 [INFO] memberlist: Suspect node2 has failed, no acks received
2016/01/04 19:18:41 [INFO] memberlist: Suspect node2 has failed, no acks received
2016/01/04 19:18:42 [INFO] memberlist: Marking node2 as failed, suspect timeout reached
2016/01/04 19:18:42 [INFO] serf: EventMemberFailed: node2 10.20.0.202
2016/01/04 19:18:42 [INFO] consul: removing server node2 (Addr: 10.20.0.202:8300) (DC: dc1)
对于服务器 -2
2016/01/04 19:18:10 [INFO] serf: EventMemberJoin: node2 10.20.0.202
2016/01/04 19:18:10 [INFO] serf: EventMemberJoin: node2.dc1 10.20.0.202
2016/01/04 19:18:10 [INFO] raft: Node at 10.20.0.202:8300 [Follower] entering Follower state
2016/01/04 19:18:10 [INFO] agent: (LAN) joining: [10.20.0.51]
2016/01/04 19:18:10 [INFO] consul: adding server node2 (Addr: 10.20.0.202:8300) (DC: dc1)
2016/01/04 19:18:10 [INFO] consul: adding server node2.dc1 (Addr: 10.20.0.202:8300) (DC: dc1)
2016/01/04 19:18:10 [INFO] serf: EventMemberJoin: node1 10.20.0.51
2016/01/04 19:18:10 [INFO] agent: (LAN) joined: 1 Err: <nil>
2016/01/04 19:18:10 [ERR] agent: failed to sync remote state: No cluster leader
2016/01/04 19:18:10 [INFO] consul: adding server node1 (Addr: 10.20.0.51:8300) (DC: dc1)
2016/01/04 19:18:12 [INFO] memberlist: Suspect node1 has failed, no acks received
2016/01/04 19:18:14 [INFO] memberlist: Suspect node1 has failed, no acks received
2016/01/04 19:18:16 [INFO] memberlist: Suspect node1 has failed, no acks received
2016/01/04 19:18:17 [INFO] memberlist: Marking node1 as failed, suspect timeout reached
2016/01/04 19:18:17 [INFO] serf: EventMemberFailed: node1 10.20.0.51
2016/01/04 19:18:17 [INFO] memberlist: Suspect node1 has failed, no acks received
2016/01/04 19:18:17 [INFO] consul: removing server node1 (Addr: 10.20.0.51:8300) (DC: dc1)
2016/01/04 19:18:19 [INFO] serf: EventMemberJoin: node1 10.20.0.51
2016/01/04 19:18:19 [INFO] consul: adding server node1 (Addr: 10.20.0.51:8300) (DC: dc1)
2016/01/04 19:18:19 [INFO] consul: New leader elected: node1
2016/01/04 19:18:21 [INFO] memberlist: Suspect node1 has failed, no acks received
2016/01/04 19:18:22 [INFO] agent: Synced service 'consul'
2016/01/04 19:18:23 [INFO] memberlist: Suspect node1 has failed, no acks received
2016/01/04 19:18:25 [INFO] memberlist: Suspect node1 has failed, no acks received
2016/01/04 19:18:26 [INFO] memberlist: Marking node1 as failed, suspect timeout reached
2016/01/04 19:18:26 [INFO] serf: EventMemberFailed: node1 10.20.0.51
2016/01/04 19:18:26 [INFO] consul: removing server node1 (Addr: 10.20.0.51:8300) (DC: dc1)
2016/01/04 19:18:26 [INFO] memberlist: Suspect node1 has failed, no acks received
2016/01/04 19:18:40 [INFO] serf: attempting reconnect to node1 10.20.0.51:8301
2016/01/04 19:18:40 [INFO] serf: EventMemberJoin: node1 10.20.0.51
我到底做错了什么。我只想在两个 EC2 实例中运行两个 consul docker 并在它们之间进行通信，而无需显式打开安全组中的端口（当我显式打开它们时，它当然可以工作！）
请有人帮忙。
谢谢