Redis主从复制，哨兵机制

勿以浮沙筑高台

---

Redis主从机制

环境搭建

这里我们搭建三个rdis来实现一个一主二从的redis集群。

• 修改redis.conf文件为三个,端口号分别是6379，6380，6381

  [root@work1 bin]#  ./redis-server ./redis-6379.conf
  [root@work1 bin]#  ./redis-server ./redis-6380.conf
  [root@work1 bin]#  ./redis-server ./redis-6381.conf
  

• 修改配置文件信息

  92 行 port 6380   #端口号
  247 行 pidfile /var/run/redis_6380.pid  #pidfile文件
  260 行 logfile "/usr/local/redis/logs/6380.log"   #配置文件地址
  342 行 dbfilename dump_6380.rdb   #rdb文件
  

• 启动三个Redis

  [root@work1 bin]#  ./redis-server ./redis-6379.conf
  [root@work1 bin]#  ./redis-server ./redis-6380.conf
  [root@work1 bin]#  ./redis-server ./redis-6381.conf
  

• 这里以6379为主，80和81为从
• 分别切换到80和81上执行命令

  SLAVEOF 192.168.3.109 6379    #执行指定master命令
  info replication              #查看当前redis主从info命令
  # Replication
  role:slave
  master_host:192.168.31.109   #masterIP
  master_port:6379             #masterPort
  master_link_status:up
  master_last_io_seconds_ago:9
  master_sync_in_progress:0
  slave_repl_offset:14
  slave_priority:100
  slave_read_only:1
  connected_slaves:0
  master_replid:705152324014710b88db075af98e225910b45d8e
  master_replid2:0000000000000000000000000000000000000000
  master_repl_offset:14
  second_repl_offset:-1
  repl_backlog_active:1
  repl_backlog_size:1048576
  repl_backlog_first_byte_offset:1
  repl_backlog_histlen:14
  127.0.0.1:6380> exit
  [root@work1 bin]# cp redis.conf redis-6381.conf
  cp：是否覆盖"redis-6381.conf"？ n
  [root@work1 bin]# ./redis-cli -p 6381
  

演示

• 在6379主中写入一个节点，在从数据中获得了节点数据。倘若重新写加入一个从Redis也可以从copy中获得节点

   127.0.0.1:6379> set name zhangshan
  OK
  127.0.0.1:6379> exit
  [root@work1 bin]# ./redis-cli -p 6381
  127.0.0.1:6381> get name
  "zhangshan"
  127.0.0.1:6381>
  [root@work1 bin]# ./redis-cli -p 6380
  127.0.0.1:6380> get name
  "zhangshan"
  127.0.0.1:6380>
  

pkill redis

主从实现原理

• 基础准备对RDB的概念请看文章。
  Redis持久化机制（RDB原理和AOF原理）

• Slave（从） 启动成功连接到 Master（主） 后会发送一个 sync （同步）同步命令。

• Master 接收到命令后，启动后台的存盘进程，将 RDB文件全部copy到Slave当中。

• Slave文件运行RDB文件，创建节点。

• 倘若后面还有数据的增加就不是全量Copy，只将增连Copy。

---

Redis哨兵机制

处理问题，当Master(主）程序发生宕机后，Slave(从)是无法重新进行选举的。因为彼此之间并没有进行联系，这时候Redis采用哨兵进行监听（本质就是心跳检测），哨兵进行监听，发现宕机后，重新从中选举master，当宕机的机子重启后，定义为Slave。

哨兵配置

• 到redis启动目录创建配置文件sentinel.conf

  # sentinel monitor 被监控的名字（随便写） host 1
  sentinel monitor myredis 192.168.31.109 6379 1
  

• 启动./redis-sentinel ./sentinel.conf

  [root@work1 bin]# vi sentinel.conf   #编辑config文件
  [root@work1 bin]# ./redis-sentinel ./sentinel.conf    #启动
  1646:X 21 Jun 2021 22:05:20.502 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
  1646:X 21 Jun 2021 22:05:20.502 # Redis version=6.0.7, bits=64, commit=00000000, modified=0, pid=1646, just started
  1646:X 21 Jun 2021 22:05:20.502 # Configuration loaded
  1646:X 21 Jun 2021 22:05:20.503 * Increased maximum number of open files to 10032 (it was originally set to 1024).
                  _._
             _.-``__ ''-._
        _.-``    `.  `_.  ''-._           Redis 6.0.7 (00000000/0) 64 bit
    .-`` .-```.  ```\/    _.,_ ''-._
   (    '      ,       .-`  | `,    )     Running in sentinel mode
   |`-._`-...-` __...-.``-._|'` _.-'|     Port: 26379
   |    `-._   `._    /     _.-'    |     PID: 1646
    `-._    `-._  `-./  _.-'    _.-'
   |`-._`-._    `-.__.-'    _.-'_.-'|
   |    `-._`-._        _.-'_.-'    |           http://redis.io
    `-._    `-._`-.__.-'_.-'    _.-'
   |`-._`-._    `-.__.-'    _.-'_.-'|
   |    `-._`-._        _.-'_.-'    |
    `-._    `-._`-.__.-'_.-'    _.-'
        `-._    `-.__.-'    _.-'
            `-._        _.-'
                `-.__.-'
  
  1646:X 21 Jun 2021 22:05:20.504 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
  1646:X 21 Jun 2021 22:05:20.611 # Sentinel ID is a184b83f8cb622f0cd4b86d8f718e535b01d924e
  1646:X 21 Jun 2021 22:05:20.611 # +monitor master myredis 127.0.0.1 6379 quorum 1    #监听Master
  1646:X 21 Jun 2021 22:05:20.611 * +slave slave 192.168.31.109:6380 192.168.31.109 6380 @ myredis 127.0.0.1 6379  #监听Slave
  1646:X 21 Jun 2021 22:05:20.616 * +slave slave 192.168.31.109:6381 192.168.31.109 6381 @ myredis 127.0.0.1 6379  #监听Slave
  

演示

• 切换到Master 6379上，shutdown掉主

  [root@work1 bin]# ./redis-cli
  127.0.0.1:6379> exit
  [root@work1 bin]# ./redis-cli shutdown    #shutdown掉主
  [root@work1 bin]#
  

• 查看哨兵监听显示，多出来以来内容，，发现纪元+1+new-epoch 2，新的Master是6381

  
   1646:X 21 Jun 2021 22:18:59.281 # +new-epoch 2    #新纪元
   1646:X 21 Jun 2021 22:18:59.281 # +try-failover master myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:18:59.292 # +vote-for-leader a184b83f8cb622f0cd4b86d8f718e535b01d924e 2
   1646:X 21 Jun 2021 22:18:59.292 # +elected-leader master myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:18:59.292 # +failover-state-select-slave master myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:18:59.377 # +selected-slave slave 127.0.0.1:6381 127.0.0.1 6381 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:18:59.377 * +failover-state-send-slaveof-noone slave 127.0.0.1:6381 127.0.0.1 6381 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:18:59.461 * +failover-state-wait-promotion slave 127.0.0.1:6381 127.0.0.1 6381 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:18:59.586 # -sdown master myredis 192.168.31.109 6381      #选举出的新Master
   1646:X 21 Jun 2021 22:18:59.587 # -odown master myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:00.233 # +promoted-slave slave 127.0.0.1:6381 127.0.0.1 6381 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:00.233 # +failover-state-reconf-slaves master myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:00.290 * +slave-reconf-sent slave 192.168.31.109:6380 192.168.31.109 6380 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:01.283 * +slave-reconf-inprog slave 192.168.31.109:6380 192.168.31.109 6380 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:01.283 * +slave-reconf-done slave 192.168.31.109:6380 192.168.31.109 6380 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:01.338 * +slave-reconf-sent slave 127.0.0.1:6380 127.0.0.1 6380 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:02.292 * +slave-reconf-inprog slave 127.0.0.1:6380 127.0.0.1 6380 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:02.292 * +slave-reconf-done slave 127.0.0.1:6380 127.0.0.1 6380 @ myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:02.383 # +failover-end master myredis 192.168.31.109 6381
   1646:X 21 Jun 2021 22:19:02.383 # +switch-master myredis 192.168.31.109 6381 127.0.0.1 6381
   1646:X 21 Jun 2021 22:19:02.384 * +slave slave 127.0.0.1:6379 127.0.0.1 6379 @ myredis 127.0.0.1 6381
   1646:X 21 Jun 2021 22:19:02.384 * +slave slave 192.168.31.109:6380 192.168.31.109 6380 @ myredis 127.0.0.1 6381
   1646:X 21 Jun 2021 22:19:02.384 * +slave slave 127.0.0.1:6380 127.0.0.1 6380 @ myredis 127.0.0.1 6381
   1646:X 21 Jun 2021 22:19:02.384 * +slave slave 192.168.31.109:6381 192.168.31.109 6381 @ myredis 127.0.0.1 6381
   1646:X 21 Jun 2021 22:19:12.550 * +convert-to-slave slave 192.168.31.109:6381 192.168.31.109 6381 @ myredis 127.0.0.1 6381
   1646:X 21 Jun 2021 22:19:32.398 # +sdown slave 127.0.0.1:6379 127.0.0.1 6379 @ myredis 127.0.0.1 6381
   1646:X 21 Jun 2021 22:20:09.608 # +sdown master myredis 127.0.0.1 6381
   1646:X 21 Jun 2021 22:20:09.608 # +odown master myredis 127.0.0.1 6381 #quorum 1/1
  

小结：

哨兵的作用，就是在中心节点宕机时，重新选举出新的Master。但是有个可能就算当哨兵宕机时，选举就不存在了，因此需要多个哨兵搭建集群。