redis 哨兵_Redis手摸手教程（三）：Redis哨兵模式

最新推荐文章于 2023-07-21 23:35:18 发布

weixin_39727706

最新推荐文章于 2023-07-21 23:35:18 发布

阅读量91

点赞数

上一节我们搭建了redis主从模式，实现了redis的主从复制，读写分离。主从模式保证了数据备份，但是发生故障依然需要运维人员施工。为了解决主从模式的弊端，本节我们搭建redis哨兵模式。

本节我们搭建一主二从三哨兵。

redis哨兵模式

sentinel 通常翻译成哨兵，用来监控主从节点的健康情况。客户端连接redis主从的时候，先连接 sentinel，sentinel告诉客户端主redis的地址是多少，然后客户端连接上redis并进行后续的操作。

当主节点挂掉的时候，客户端就得不到连接了因而报错了，客户端重新向sentinel询问主master的地址，然后客户端得到了新选举出来的主redis。当原来的主机恢复后，会将其作为新主的从机。

我们直接在上一节的已经搭建好的主从基础上搭建redis哨兵模式，上一节的主从模式节点信息如下：

各容器network信息如下：

 "Containers": {    "3729ea952faade7c3bf37467f00f8a8c6676b9a98e27a897582272f1edce2087": {        "Name": "redis-6382",        "EndpointID": "95f9a05d2b6d10c13eb3d8df706a2700f33a6704ddda290bc6840676eebc6676",        "MacAddress": "02:42:ac:14:00:04",        "IPv4Address": "172.20.0.4/16",        "IPv6Address": ""    },    "a7f4ccfa8e8aa87efb697f3ee5fef686e5138ebb0ad5faac282e7850d8cb176f": {        "Name": "redis-6380",        "EndpointID": "71af5cda52f5026aba9b4040890629d5ca560c2a7208b2bc7f0682a97a19d9bf",        "MacAddress": "02:42:ac:14:00:02",        "IPv4Address": "172.20.0.2/16",        "IPv6Address": ""    },    "ff3d51f269c40912a21b2a8686078d963f292b32d32bce6dad089d9008377a36": {        "Name": "redis-6381",        "EndpointID": "405a1e61ea2afae2fca8b0ce31f6ac83b7c03440668b3f7ec42072a0ebff18e3",        "MacAddress": "02:42:ac:14:00:03",        "IPv4Address": "172.20.0.3/16",        "IPv6Address": ""    }}

现在我们约定sentinel哨兵节点信息：

创建工作目录

mkdir -p /usr/local/redis/sentinel/sentinel-26379/confmkdir -p /usr/local/redis/sentinel/sentinel-26379/datamkdir -p /usr/local/redis/sentinel/sentinel-26380/confmkdir -p /usr/local/redis/sentinel/sentinel-26380/datamkdir -p /usr/local/redis/sentinel/sentinel-26381/confmkdir -p /usr/local/redis/sentinel/sentinel-26381/data

用来存放3个哨兵的配置文件及日志文件。

准备sentinel.conf

wget http://download.redis.io/redis-stable/sentinel.conf

将sentinel.conf复制3份，分别存放到这3个目录下，并修改对应内容：

/usr/local/redis/sentinel/sentinel-26379/conf

dir /datalogfile "sentinel-26379.log"  # 修改日志文件的路径sentinel monitor mymaster redis-6380 6379 2 # redis-6380表示 master 节点地址，最后一个2表示，两个sentinel判定master被动下线后，就进行failover(故障转移)sentinel auth-pass mymaster 123456 # 设置master的密码

/usr/local/redis/sentinel/sentinel-26380/conf

dir /datalogfile "sentinel-26380.log"  # 修改日志文件的路径sentinel monitor mymaster redis-6380 6379 2 # redis-6380表示 master 节点地址，最后一个2表示，两个sentinel判定master被动下线后，就进行failover(故障转移)sentinel auth-pass mymaster 123456 # 设置master的密码

/usr/local/redis/sentinel/sentinel-26381/conf

dir /datalogfile "sentinel-26381.log"  # 修改日志文件的路径sentinel monitor mymaster redis-6380 6379 2 # redis-6380表示 master 节点地址，最后一个2表示，两个sentinel判定master被动下线后，就进行failover(故障转移)sentinel auth-pass mymaster 123456 # 设置master的密码

目录结构

[root@localhost sentinel]# tree.├── sentinel-26379│   ├── conf│   │   └── sentinel.conf│   └── data├── sentinel-26380│   ├── conf│   │   └── sentinel.conf│   └── data└── sentinel-26381    ├── conf    │   └── sentinel.conf    └── data9 directories, 3 files

创建sentinel容器

官网建议设置奇数个哨兵，这样才适合选举机制。
我们逐个执行下面的启动命令，并查看sentinel.conf文件的变化。

docker run -d -p 26379:26379 --name sentinel-26379 --net=redis-master-slaver -v /usr/local/redis/sentinel/sentinel-26379/conf/sentinel.conf:/etc/redis/sentinel.conf -v /usr/local/redis/sentinel/sentinel-26379/data:/data redis:5.0.7 edis-sentinel /etc/redis/sentinel.confdocker run -d -p 26380:26379 --name sentinel-26380 --net=redis-master-slaver -v /usr/local/redis/sentinel/sentinel-26380/conf/sentinel.conf:/etc/redis/sentinel.conf -v /usr/local/redis/sentinel/sentinel-26380/data:/data redis:5.0.7 edis-sentinel /etc/redis/sentinel.confdocker run -d -p 26381:26379 --name sentinel-26381 --net=redis-master-slaver -v /usr/local/redis/sentinel/sentinel-26381/conf/sentinel.conf:/etc/redis/sentinel.conf -v /usr/local/redis/sentinel/sentinel-26381/data:/data redis:5.0.7 edis-sentinel /etc/redis/sentinel.conf

我们启动sentinel-26379后，查看/usr/local/redis/sentinel/sentinel-26379/conf/sentinel.conf文件，有了这些变化：

sentinel myid 327716021d2b88790338b362173a34163a855e56sentinel deny-scripts-reconfig yessentinel monitor mymaster 172.20.0.2 6379 2  #自动替换成了redis-6380的ipsentinel config-epoch mymaster 0sentinel leader-epoch mymaster 0# Generated by CONFIG REWRITEsentinel known-replica mymaster 172.20.0.4 6379sentinel known-replica mymaster 172.20.0.3 6379sentinel current-epoch 0

日志文件(/usr/local/redis/sentinel/sentinel-26379/data/sentinel-26379.log)：

[root@localhost data]# vim sentinel-26379.log 1:X 12 Feb 2020 03:03:09.681 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo1:X 12 Feb 2020 03:03:09.681 # Redis version=5.0.7, bits=64, commit=00000000, modified=0, pid=1, just started1:X 12 Feb 2020 03:03:09.681 # Configuration loaded1:X 12 Feb 2020 03:03:09.683 * Running mode=sentinel, port=26379.1:X 12 Feb 2020 03:03:09.683 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.1:X 12 Feb 2020 03:03:09.687 # Sentinel ID is 327716021d2b88790338b362173a34163a855e561:X 12 Feb 2020 03:03:09.687 # +monitor master mymaster 172.20.0.2 6379 quorum 21:X 12 Feb 2020 03:03:09.689 * +slave slave 172.20.0.3:6379 172.20.0.3 6379 @ mymaster 172.20.0.2 63791:X 12 Feb 2020 03:03:09.690 * +slave slave 172.20.0.4:6379 172.20.0.4 6379 @ mymaster 172.20.0.2 63791:X 12 Feb 2020 03:06:10.373 * +fix-slave-config slave 172.20.0.4:6379 172.20.0.4 6379 @ mymaster 172.20.0.2 63791:X 12 Feb 2020 03:06:10.374 * +fix-slave-config slave 172.20.0.3:6379 172.20.0.3 6379 @ mymaster 172.20.0.2 6379

可以看到它已经检测到了master和2个slave.

我们接下来启动sentinel-26380和sentinel-26381，查看/usr/local/redis/sentinel/sentinel-26379/conf/sentinel.conf文件,新增了2行记录，表示监测到了新的哨兵

sentinel known-sentinel mymaster 172.20.0.7 26379 90137df24fb27cb5821c3a3372c2101379b75018sentinel known-sentinel mymaster 172.20.0.6 26379 d88c61fe5d93be0aaaaff6629196f0034683bb58