实现主从容错切换迁移

在3主3从【案例1/4】的基础上，实现主从容错切换迁移，示意图如下：

一、数据读写存储

1、启动6机构成的集群并通过exec进入（任意一台都行）：

docker exec -it redis-node-1 /bin/bash

2、对6381新增两个key

# 登陆任意redis（重点：此为单机登陆命令）
root@localhost:/data$ redis-cli -p 6381

# 查看数据，为空
127.0.0.1:6381> keys *
(empty array)

# 设置k1=v1 失败
127.0.0.1:6381> set k1 v1
(error) MOVED 12706 12.114.161.16:6383

# 设置k2=v2 、k3=v3 成功
127.0.0.1:6381> set k2 v2
OK
127.0.0.1:6381> set k3 v3
OK

# 设置k4=v4 失败
127.0.0.1:6381> set k4 v4
(error) MOVED 8455 12.114.161.16:6382

因为，总共有16384个槽位，平均分给3个主机，每个槽位段分别为16384/3个槽位。
所以报错原因是：登陆的第一台机器存储的槽位区间，无法存储k1和k4的值。【解决办法：见第3步】

3、集群的形式登陆【redis-cli -p 6381 -c】
登陆命令 redis-cli -p 6381 -c ，加入**-c**,代表以集群的形式登陆后，优化路由后，插入信息时，可根据哈希槽位进行自由切换。不再局限于某一台机器。

# 以集群的形式登陆（重点：此为集群登陆命令）
root@localhost:/data# redis-cli -p 6381 -c
127.0.0.1:6381> keys *
1) "k3"
2) "k2"
127.0.0.1:6381> set k1 v1
-> Redirected to slot [12706] located at 12.114.161.16:6383
OK
12.114.161.16:6383> # 已经切换到6383号机器
12.114.161.16:6383> set k4 v4
-> Redirected to slot [8455] located at 12.114.161.16:6382
OK
12.114.161.16:6382> # 已经切换到6382号机器

4、查看集群信息【redis-cli --cluster check ip地址：端口】

# 进入任意台节点机器
docker exec -it redis-node-1 /bin/bash

# 查看集群槽位信息
redis-cli --cluster check 12.114.161.16:6381

root@localhost:/data# redis-cli --cluster check 12.114.161.16:6381

# 该主机   数据值 | 哈希槽位数 | 从主机个数
12.114.161.16:6381 (6e961a47...) -> 2 keys | 5461 slots | 1 slaves.  
12.114.161.16:6382 (f097fec9...) -> 1 keys | 5462 slots | 1 slaves.
12.114.161.16:6383 (78ac7be1...) -> 1 keys | 5461 slots | 1 slaves.
[OK] 4 keys in 3 masters.
0.00 keys per slot on average.
>>> Performing Cluster Check (using node 12.114.161.16:6381)
M: 6e961a4765b555189708bebb69badf7dfad25cd5 12.114.161.16:6381
   slots:[0-5460] (5461 slots) master
   1 additional replica(s)
S: 4ed46e0368698cd9d5af2ee84631c878b8ebc4d0 12.114.161.16:6386
   slots: (0 slots) slave
   replicates 6e961a4765b555189708bebb69badf7dfad25cd5
M: f097fec937f54d147d316c1c62e26cb67c9fd059 12.114.161.16:6382
   slots:[5461-10922] (5462 slots) master
   1 additional replica(s)
S: 0e34147ce2544cd90f5cce78d5493ae9e8625dfe 12.114.161.16:6384
   slots: (0 slots) slave
   replicates f097fec937f54d147d316c1c62e26cb67c9fd059
M: 78ac7be13522bd4ffd6fcf900c6c149c6938ecc2 12.114.161.16:6383
   slots:[10923-16383] (5461 slots) master
   1 additional replica(s)
S: 9bc2417a9cd6545ef2445eb4aa0610d586acd73b 12.114.161.16:6385
   slots: (0 slots) slave
   replicates 78ac7be13522bd4ffd6fcf900c6c149c6938ecc2
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.

显示如下内容：
1、3台主机 、3台从机器的信息（id、端口、存储数据等）
2、每台主机的哈希槽位区间、和存取数值多少

二、容错切换迁移

选择任意一组“主从机器”，① 如果主机宕机后，从机是否后替补上；②如果宕机的主机再次恢复，是否会重新“夺回” 主主机地位，还是顺延作为“从机”？

1、先停止主机redis-node-1，再次查看集群信息。命令：

# 停止主机redis-node-1
[root@localhost ~]# docker stop redis-node-1
redis-node-1

# 查看集群节点信息
12.114.161.16:6386> cluster nodes

12.114.161.16:6386> cluster nodes
f097fec937f54d147d316c1c62e26cb67c9fd059 12.114.161.16:6382@16382 master - 0 1705309024699 2 connected 5461-10922
78ac7be13522bd4ffd6fcf900c6c149c6938ecc2 12.114.161.16:6383@16383 master - 0 1705309023000 3 connected 10923-16383
6e961a4765b555189708bebb69badf7dfad25cd5 12.114.161.16:6381@16381 master,fail - 1705308720558 1705308717000 1 disconnected
0e34147ce2544cd90f5cce78d5493ae9e8625dfe 12.114.161.16:6384@16384 slave f097fec937f54d147d316c1c62e26cb67c9fd059 0 1705309024000 2 connected
4ed46e0368698cd9d5af2ee84631c878b8ebc4d0 12.114.161.16:6386@16386 myself,master - 0 1705309023000 7 connected 0-5460
9bc2417a9cd6545ef2445eb4aa0610d586acd73b 12.114.161.16:6385@16385 slave 78ac7be13522bd4ffd6fcf900c6c149c6938ecc2 0 1705309025703 3 connected

说明：

• 6381@16381 master,fail ： 该主机（redis-node-1）为master，已宕机
• 6386@16386 myself,master ：该主机（redis-node-6）顶替成为master

2、先还原之前的3主3从，再查看集群信息

# 停止主机redis-node-1
[root@localhost ~]# docker start redis-node-1
redis-node-1

# 查看集群节点信息
12.114.161.16:6386> cluster nodes

12.114.161.16:6386> cluster nodes
f097fec937f54d147d316c1c62e26cb67c9fd059 12.114.161.16:6382@16382 master - 0 1705309203416 2 connected 5461-10922
78ac7be13522bd4ffd6fcf900c6c149c6938ecc2 12.114.161.16:6383@16383 master - 0 1705309206000 3 connected 10923-16383
6e961a4765b555189708bebb69badf7dfad25cd5 12.114.161.16:6381@16381 slave 4ed46e0368698cd9d5af2ee84631c878b8ebc4d0 0 1705309205421 7 connected
0e34147ce2544cd90f5cce78d5493ae9e8625dfe 12.114.161.16:6384@16384 slave f097fec937f54d147d316c1c62e26cb67c9fd059 0 1705309207426 2 connected
4ed46e0368698cd9d5af2ee84631c878b8ebc4d0 12.114.161.16:6386@16386 myself,master - 0 1705309204000 7 connected 0-5460
9bc2417a9cd6545ef2445eb4aa0610d586acd73b 12.114.161.16:6385@16385 slave 78ac7be13522bd4ffd6fcf900c6c149c6938ecc2 0 1705309206422 3 connected

说明：

• 6381@16381 slave ： 该主机（redis-node-1）虽然已经恢复，但沦为为slave从机器
• 6386@16386 myself,master：该主机（redis-node-6）没有变化，依然为master

总结：集群中，master主机宕机后，slave主机顶替上来，成为新master主机； 而且宕机的机器恢复后，将会成为slave机器