redis sentinel监控高可用集群实现的配置步骤

Redis Sentinel是Redis分布式系统的监控工具，它能够监控Redis集群中每个节点的运行状态，并在节点故障时进行自动故障转移，从而保证Redis集群的高可用性。下面是采用Redis Sentinel实现高可用集群监控的完整配置步骤：

1. 安装Redis Sentinel

首先需要安装Redis Sentinel。可以通过以下命令进行安装：

sudo apt-get install redis-server redis-tools redis-sentinel

1. 配置Redis主从及Sentinel节点

主从节点分别运行在不同的机器上，Sentinel节点也可以运行在不同的机器上。配置文件的示例如下：

主节点：

# /etc/redis/redis.conf

port 6379 # main port
bind 127.0.0.1 # bind to local IP address
dir /var/lib/redis # data directory
logfile "/var/log/redis/redis-server.log"

requirepass password # set password for redis server
masterauth password # set password for slave server to connect
masterinfoformat "ip=%s,port=%s,state=%s,info=%s" # set the format of master node info that is sent to the slave node

从节点：

# /etc/redis/redis.conf

port 6380 # slave port
slaveof 127.0.0.1 6379 # set the address and port of the master node
bind 127.0.0.1 # bind to local IP address
dir /var/lib/redis # data directory
logfile "/var/log/redis/redis-server.log"

requirepass password # set password for redis server
masterauth password # set password for slave server to connect

Sentinel节点：

# /etc/redis/sentinel.conf

port 26379 # sentinel port
bind 127.0.0.1 # bind to local IP address
dir /var/lib/redis # data directory
logfile "/var/log/redis/redis-sentinel.log"

sentinel monitor mymaster 127.0.0.1 6379 2 # set sentinel to monitor the master node, and set the quorum to 2
sentinel down-after-milliseconds mymaster 5000 # set the time after which a node is considered down
sentinel failover-timeout mymaster 60000 # set the time after which a failover is considered successful
sentinel auth-pass mymaster password # set the password for monitoring

1. 启动Redis主从及Sentinel节点

在启动前，需要确保主从节点的Redis服务已启动。然后可以依次启动主从节点和Sentinel节点：

启动主节点：

redis-server /etc/redis/redis.conf

启动从节点：

redis-server /etc/redis/redis.conf --port 6380 --slaveof 127.0.0.1 6379

启动Sentinel节点：

redis-sentinel /etc/redis/sentinel.conf

1. 检查Redis Sentinel是否生效

在已启动Redis Sentinel节点的机器上，运行以下命令：

redis-cli -p 26379 sentinel master mymaster

则会显示当前master节点的信息，如下所示：

1) "name"
2) "mymaster"
3) "ip"
4) "127.0.0.1"
5) "port"
6) "6379"
7) "runid"
8) "long-string-of-characters"
9) "flags"
10) "master"
11) "pending-commands"
12) "0"
13) "last-ok-ping-reply"
14) "429"
15) "last-ping-reply"
16) "429"
17) "info-refresh"
18) "5847"
19) "num-other-sentinels"
20) "0"
21) "quorum"
22) "2"
23) "failover-timeout"
24) "60000"
25) "parallel-syncs"
26) "1"

1. 故障转移测试

在主节点所在的机器上，在Redis配置文件中的bind后加上0.0.0.0，并重启Redis服务。然后可以观察到Sentinel节点会自动检测到主节点故障，进行故障转移，将从节点提升为主节点。运行以下命令，检查主节点信息：

redis-cli -p 26379 sentinel masters

切换完成后，从节点会输出如下信息：

slave: ip=127.0.0.1,port=6380,state=online,offset=XXX,lag=0,slave_read_only=1

示例1：添加Sentinel节点

可以通过添加Sentinel节点的方式，提高Redis集群的可用性。示例如下：

1. 在另外一台机器上配置Redis Sentinel：

# /etc/redis/sentinel.conf

port 26380 # sentinel port
bind 0.0.0.0 # bind to all IP addresses
dir /var/lib/redis # data directory
logfile "/var/log/redis/redis-sentinel.log"

sentinel monitor mymaster 127.0.0.1 6379 2 # set sentinel to monitor the master node, and set the quorum to 2
sentinel down-after-milliseconds mymaster 5000 # set the time after which a node is considered down
sentinel failover-timeout mymaster 60000 # set the time after which a failover is considered successful
sentinel auth-pass mymaster password # set the password for monitoring

1. 启动新的Sentinel节点：

redis-sentinel /etc/redis/sentinel.conf

1. 在已启动的Sentinel节点中添加新的Sentinel节点：

redis-cli -p 26379 sentinel add mymaster 127.0.0.1 26380

示例2：查看Redis Sentinel状态

可以通过以下命令查看每个Sentinel节点的状态：

redis-cli -p 26379 sentinel sentinels mymaster

输出示例如下：

1) 1) "name"
   2) "127.0.0.1:26379"
   3) "ip"
   4) "127.0.0.1"
   5) "port"
   6) "26379"
   7) "runid"
   8) "f35aa380bd0595e2f5d85e22ee5fec5eaaa18476"
   9) "flags"
  10) "sentinel"
  11) "pending-commands"
  12) "0"
  13) "last-ok-ping-reply"
  14) "0"
  15) "last-ping-reply"
  16) "0"
  17) "info-refresh"
  18) "1644"
  19) "num-other-sentinels"
  20) "3"
  21) "quorum"
  22) "2"
  23) "failover-timeout"
  24) "60000"
  25) "parallel-syncs"
  26) "1"
2) 1) "name"
   2) "127.0.0.1:26381"
   3) "ip"
   4) "127.0.0.1"
   5) "port"
   6) "26381"
   7) "runid"
   8) "c63ec77f2b11654a3e7a23acacc046ff5d43caf5"
   9) "flags"
  10) "sentinel"
  11) "pending-commands"
  12) "0"
  13) "last-ok-ping-reply"
  14) "0"
  15) "last-ping-reply"
  16) "0"
  17) "info-refresh"
  18) "1644"
  19) "num-other-sentinels"
  20) "3"
  21) "quorum"
  22) "2"
  23) "failover-timeout"
  24) "60000"
  25) "parallel-syncs"
  26) "1"

可见当前存在两个Sentinel节点。