个人学习笔记，谢绝转载！！！

原文：https://www.cnblogs.com/wshenjin/p/15296600.html

配置三个redis实例：

# cluster
cluster-enabled yes
cluster-config-file nodes.conf
cluster-node-timeout 15000

启动三个Redis实例：

docker run -tid --network=net_192.168.1 --ip=192.168.1.11 -p 6011:6001 -p 16011:16001 --name redis_cluster_node01 redis-4.0.14

docker run -tid --network=net_192.168.1 --ip=192.168.1.12 -p 6012:6001 -p 16012:16001 --name redis_cluster_node02 redis-4.0.14

docker run -tid --network=net_192.168.1 --ip=192.168.1.13 -p 6013:6001 -p 16013:16001 --name redis_cluster_node03 redis-4.0.14

创建集群:

[root@ ]# redis-trib.rb create 192.168.1.11:6001 192.168.1.12:6001 192.168.1.13:6001
>>> Creating cluster
>>> Performing hash slots allocation on 3 nodes...
Using 3 masters:
192.168.1.11:6001
192.168.1.12:6001
192.168.1.13:6001
M: 468b51d1918e82ad6ca81b3772e469be1000e23f 192.168.1.11:6001
   slots:0-5460 (5461 slots) master
M: e91829ab91203f38f37de7114e82ba5883dcff3c 192.168.1.12:6001
   slots:5461-10922 (5462 slots) master
M: 6dce80e6cc8a8e6863165087581496409f7e73ad 192.168.1.13:6001
   slots:10923-16383 (5461 slots) master
Can I set the above configuration? (type 'yes' to accept): yes
>>> Nodes configuration updated
>>> Assign a different config epoch to each node
>>> Sending CLUSTER MEET messages to join the cluster
Waiting for the cluster to join.
>>> Performing Cluster Check (using node 192.168.1.11:6001)
M: 468b51d1918e82ad6ca81b3772e469be1000e23f 192.168.1.11:6001
   slots:0-5460 (5461 slots) master
   0 additional replica(s)
M: 6dce80e6cc8a8e6863165087581496409f7e73ad 192.168.1.13:6001
   slots:10923-16383 (5461 slots) master
   0 additional replica(s)
M: e91829ab91203f38f37de7114e82ba5883dcff3c 192.168.1.12:6001
   slots:5461-10922 (5462 slots) master
   0 additional replica(s)
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.

此时的Redis Cluster集群正常的，却无法对外部网络提供业务。
这是因为，Cluster节点需要告诉用户或者是其他节点连接自己的IP和端口，默认情况下，Redis会自动检测自己的IP和从配置中获取绑定的PORT，告诉客户端或者是其他节点。而在Docker环境中，如果使用的不是host网络模式，在容器内部的IP和PORT都是隔离的，那么客户端和其他节点无法通过节点公布的IP和PORT建立连接。

效果如下：

1.1.1.1:6012> CLUSTER NODES
6dce80e6cc8a8e6863165087581496409f7e73ad 192.168.1.13:6001@16001 master - 0 1631788081031 3 connected 10923-16383
468b51d1918e82ad6ca81b3772e469be1000e23f 192.168.1.11:6001@16001 master - 0 1631788082034 1 connected 0-5460
e91829ab91203f38f37de7114e82ba5883dcff3c 192.168.1.12:6001@16001 myself,master - 0 1631788079000 2 connected 5461-10922
1.1.1.1:6012> get mkey02897
-> Redirected to slot [15915] located at 192.168.1.13:6001
Could not connect to Redis at 192.168.1.13:6001: Connection timed out

从Redis4.0将兼容 NAT 和 Docker， 如下配置：

cluster-announce-ip：要宣布的IP地址。
cluster-announce-port：要宣布的数据端口。
cluster-announce-bus-port：要宣布的集群总线端口

##每个节点都要配置
192.168.1.11:6001> config set cluster-announce-bus-port 16011
OK
192.168.1.11:6001> config set cluster-announce-port 6011
OK
192.168.1.11:6001> config set cluster-announce-ip 1.1.1.1
OK

配置了以后，Redis节点会将配置中的这些IP和PORT告知客户端或其他节点，而这些IP和PORT是通过Docker转发到容器内的临时IP和PORT的。

配置后的效果：

1.1.1.1:6012> CLUSTER NODES
6dce80e6cc8a8e6863165087581496409f7e73ad 1.1.1.1:6013@16013 master - 0 1631788348900 3 connected 10923-16383
468b51d1918e82ad6ca81b3772e469be1000e23f 1.1.1.1:6011@16011 master - 0 1631788349904 1 connected 0-5460
e91829ab91203f38f37de7114e82ba5883dcff3c 1.1.1.1:6001@16001 myself,master - 0 1631788346000 2 connected 5461-10922
1.1.1.1:6012> get mkey02897
-> Redirected to slot [15915] located at 1.1.1.1:6013
"__ac268adff36ba00140264436b7436b749a"

学习参考：https://www.jianshu.com/p/57aac41e26e8