docker搭建redis主从哨兵集群的实现步骤

1：获取redis镜像

docker pull redis:6.2.7

2：创建redis主从+哨兵docker-compose文件

cd /opt/docker/redis
vi docker-compose.yml

docker-compose.yml的内容如下

version: \’3\’
services:
master:
image: redis:6.2.7 ## 镜像
container_name: redis-master
command: redis-server /etc/redis/redis.conf –requirepass 123456 –masterauth 123456
volumes:
– /opt/docker/redis/data/redis_data1:/data
– /opt/docker/redis/conf/redis1.conf:/etc/redis/redis.conf
network_mode: \”host\”
slave1:
image: redis:6.2.7 ## 镜像
container_name: redis-slave-1
volumes:
– /opt/docker/redis/data/redis_data2:/data
– /opt/docker/redis/conf/redis2.conf:/etc/redis/redis.conf
command: redis-server /etc/redis/redis.conf –slaveof redis-master 6379 –requirepass 123456 –masterauth 123456
depends_on:
– master
network_mode: \”host\”
slave2:
image: redis:6.2.7 ## 镜像
container_name: redis-slave-2
volumes:
– /opt/docker/redis/data/redis_data3:/data
– /opt/docker/redis/conf/redis3.conf:/etc/redis/redis.conf
command: redis-server /etc/redis/redis.conf –slaveof redis-master 6379 –requirepass 123456 –masterauth 123456
depends_on:
– master
network_mode: \”host\”
sentinel1:
image: redis:6.2.7 ## 镜像
container_name: redis-sentinel-1
command: redis-sentinel /usr/local/etc/redis/sentinel.conf
volumes:
– /opt/docker/redis/conf/sentinel1.conf:/usr/local/etc/redis/sentinel.conf
network_mode: \”host\”
depends_on:
– master
– slave1
– slave2
sentinel2:
image: redis:6.2.7 ## 镜像
container_name: redis-sentinel-2
command: redis-sentinel /usr/local/etc/redis/sentinel.conf
volumes:
– /opt/docker/redis/conf/sentinel2.conf:/usr/local/etc/redis/sentinel.conf
network_mode: \”host\”
depends_on:
– master
– slave1
– slave2
sentinel3:
image: redis:6.2.7 ## 镜像
container_name: redis-sentinel-3
command: redis-sentinel /usr/local/etc/redis/sentinel.conf
volumes:
– /opt/docker/redis/conf/sentinel3.conf:/usr/local/etc/redis/sentinel.conf
network_mode: \”host\”
depends_on:
– master
– slave1
– slave2

3：redis配置和哨兵配置说明

查看配置文件的目录树

cd /opt/docker/redis
tree ./

结构如下

redis1.conf，redis2.conf，redis3.conf配置如下

#redis1.conf
bind 0.0.0.0
port 6379
protected-mode no
slave-read-only no

#redis2.conf
bind 0.0.0.0
port 6380
protected-mode no
slave-read-only no

#redis3.conf
bind 0.0.0.0
port 6381
protected-mode no
slave-read-only no

sentinel1.conf，sentinel1.conf，sentinel1.conf配置：

#sentinel1.conf
port 26379
dir /tmp
sentinel monitor mymaster 192.168.18.131 6379 2
sentinel auth-pass mymaster 123456
sentinel down-after-milliseconds mymaster 30000
sentinel parallel-syncs mymaster 1
sentinel failover-timeout mymaster 10000
sentinel deny-scripts-reconfig yes

#sentinel2.conf
port 26380
dir /tmp
sentinel monitor mymaster 192.168.18.131 6379 2
sentinel auth-pass mymaster 123456
sentinel down-after-milliseconds mymaster 30000
sentinel parallel-syncs mymaster 1
sentinel failover-timeout mymaster 10000
sentinel deny-scripts-reconfig yes

#sentinel3.conf
port 26381
dir /tmp
sentinel monitor mymaster 192.168.18.131 6379 2
sentinel auth-pass mymaster 123456
sentinel down-after-milliseconds mymaster 30000
sentinel parallel-syncs mymaster 1
sentinel failover-timeout mymaster 10000
sentinel deny-scripts-reconfig yes

说明：

【sentinel monitor mymaster 192.168.18.131 6379 2】192.168.18.131为服务器的IP地址，6379为redis master的端口号【sentinel auth-pass mymaster 123456】设置主节点的密码【sentinel down-after-milliseconds mymaster 30000】表示在一段时间范围内sentinel向master发送的心跳PING没有回复则认为master不可用了。【sentinel parallel-syncs mymaster 1】的parallel-syncs表示设置在故障转移之后，同时可以重新配置使用新master的slave的数量。数字越低，更多的时间将会用故障转移完成，但是如果slaves配置为服务旧数据，你可能不希望所有的slave同时重新同步master。因为主从复制对于slave是非阻塞的，当停止从master加载批量数据时有一个片刻延迟。通过设置选项为1，确信每次只有一个slave是不可到达的。【sentinel failover-timeout mymaster 10000 】表示10秒内mymaster还没活过来，则认为master宕机了。

redis_data1,redis_data2,redis_data3为空文件夹，用于存放redis数据文件