Redis+Sentinel集群安装与配置
1.环境介绍
集群机器 redis服务 哨兵(sentinel) 节点描述
虚拟机：192.168.20.105 安装redis服务,端口6379 安装sentinel服务,端口26379 主节点
虚拟机：192.168.20.29 安装redis服务,端口6379 安装sentinel服务,端口26379 从节点
虚拟机：192.168.20.108 安装redis服务,端口6379 安装sentinel服务,端口26379 从节点
即：3个redis服务，三个sentinel服务，192.168.20.105 为主节点，实现按照上篇文章介绍配置好主从关系。
2.相关启动命令
cd到/usr/redis/bin目录可以执行以下命令：
redis-server /usr/redis/etc/redis.conf#启动redis服务
./redis-sentinel /usr/redis/etc/sentinel.conf#启动sentinel哨兵服务
./redis-cli -p 6379 -a 123456#打开本机redis客户端连接，连接上之后可执行set、get、keys *等命令
ps -ef|grep redis#查看redis状态
3.配置文件
虚拟机：192.168.20.105
redis.conf配置文件：
daemonize yes#后台启动
pidfile "/var/run/redis.pid"
port 6379
timeout 0
tcp-keepalive 0
loglevel notice
logfile ""
databases 16
save 900 1
save 300 10
save 60 10000
stop-writes-on-bgsave-error yes
rdbcompression yes
rdbchecksum yes
dbfilename "dump.rdb"
dir "/usr/redis/data"
masterauth "123456"#如果做故障切换，不论主从节点都要填写密码且要保持一致
slave-serve-stale-data yes
slave-read-only yes
repl-disable-tcp-nodelay no
slave-priority 98
requirepass "123456"#当前redis密码
appendonly yes
# appendfsync always
appendfsync everysec
# appendfsync no
no-appendfsync-on-rewrite no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
lua-time-limit 5000
slowlog-log-slower-than 10000
slowlog-max-len 128
notify-keyspace-events ""
hash-max-ziplist-entries 512
hash-max-ziplist-value 64
list-max-ziplist-entries 512
list-max-ziplist-value 64
set-max-intset-entries 512
zset-max-ziplist-entries 128
zset-max-ziplist-value 64
activerehashing yes
client-output-buffer-limit normal 0 0 0
client-output-buffer-limit slave 256mb 64mb 60
client-output-buffer-limit pubsub 32mb 8mb 60
hz 10
aof-rewrite-incremental-fsync yes
# Generated by CONFIG REWRITE
port 26379
sentinel monitor mymaster 192.168.20.105 6379 2#2表示在sentinel集群中只要有两个节点检测到redis主节点出故障就进行切换，单sentinel节点无效(自己测试发现的)
sentinel down-after-milliseconds mymaster 3000#如果3s内mymaster无响应，则认为mymaster宕机了
sentinel failover-timeout mymaster 10000#如果10秒后,mysater仍没活过来，则启动failover
sentinel auth-pass mymaster 123456#redis主节点密码
sentinel config-epoch mymaster 89
daemonize yes
dir "/usr/redis/bin"#指定工作目录
sentinel known-slave mymaster 192.168.20.29 6379#sentinel自动生成的
sentinel known-slave mymaster 192.168.20.108 6379#sentinel自动生成的
sentinel known-sentinel mymaster 192.168.20.108 26379 695a7bf69bf7c0bdacb3a01ea262319cfca85de9#sentinel自动生成的
# Generated by CONFIG REWRITE
sentinel known-sentinel mymaster 192.168.20.29 26379 b1fba42ff8b363b2aea5c52e5dfa2c1ae52869e9#sentinel自动生成的虚拟机：192.168.20.29
redis.conf配置文件：
daemonize yes
pidfile "/var/run/redis.pid"
port 6379
timeout 0
tcp-keepalive 0
loglevel notice
logfile ""
databases 16
save 900 1
save 300 10
save 60 10000
stop-writes-on-bgsave-error yes
rdbcompression yes
rdbchecksum yes
dbfilename "dump.rdb"
dir "/usr/redis/data"
masterauth "123456"#主节点密码
slave-serve-stale-data yes
slave-read-only yes
repl-disable-tcp-nodelay no
slave-priority 98
requirepass "123456"
appendonly yes
# appendfsync always
appendfsync everysec
# appendfsync no
no-appendfsync-on-rewrite no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
lua-time-limit 5000
slowlog-log-slower-than 10000
slowlog-max-len 128
notify-keyspace-events ""
hash-max-ziplist-entries 512
hash-max-ziplist-value 64
list-max-ziplist-entries 512
list-max-ziplist-value 64
set-max-intset-entries 512
zset-max-ziplist-entries 128
zset-max-ziplist-value 64
activerehashing yes
client-output-buffer-limit normal 0 0 0
client-output-buffer-limit slave 256mb 64mb 60
client-output-buffer-limit pubsub 32mb 8mb 60
hz 10
aof-rewrite-incremental-fsync yes
# Generated by CONFIG REWRITE
slaveof 192.168.20.105 6379#配置主节点信息
port 26379
sentinel monitor mymaster 192.168.20.105 6379 2
sentinel down-after-milliseconds mymaster 3000
sentinel failover-timeout mymaster 10000
sentinel auth-pass mymaster 123456
sentinel config-epoch mymaster 89
daemonize yes
dir "/usr/redis/bin"
sentinel known-slave mymaster 192.168.20.108 6379
sentinel known-slave mymaster 192.168.20.29 6379
sentinel known-sentinel mymaster 192.168.20.108 26379 695a7bf69bf7c0bdacb3a01ea262319cfca85de9
# Generated by CONFIG REWRITE
sentinel known-sentinel mymaster 192.168.20.105 26379 a8d94fbe0f1120278136c84a612ea84d57d2501c虚拟机：192.168.20.108
redis.conf配置文件：
daemonize yes
pidfile "/var/run/redis.pid"
port 6379
timeout 0
tcp-keepalive 0
loglevel notice
logfile ""
databases 16
save 900 1
save 300 10
save 60 10000
stop-writes-on-bgsave-error yes
rdbcompression yes
rdbchecksum yes
dbfilename "dump.rdb"
dir "/usr/redis/data"
masterauth "123456"
slave-serve-stale-data yes
slave-read-only yes
repl-disable-tcp-nodelay no
slave-priority 100
requirepass "123456"
appendonly yes
# appendfsync always
appendfsync everysec
# appendfsync no
no-appendfsync-on-rewrite no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
lua-time-limit 5000
slowlog-log-slower-than 10000
slowlog-max-len 128
notify-keyspace-events ""
hash-max-ziplist-entries 512
hash-max-ziplist-value 64
list-max-ziplist-entries 512
list-max-ziplist-value 64
set-max-intset-entries 512
zset-max-ziplist-entries 128
zset-max-ziplist-value 64
activerehashing yes
client-output-buffer-limit normal 0 0 0
client-output-buffer-limit slave 256mb 64mb 60
client-output-buffer-limit pubsub 32mb 8mb 60
hz 10
aof-rewrite-incremental-fsync yes
# Generated by CONFIG REWRITE
slaveof 192.168.20.105 6379
port 26379
sentinel monitor mymaster 192.168.20.105 6379 2
sentinel down-after-milliseconds mymaster 3000
sentinel failover-timeout mymaster 10000
sentinel auth-pass mymaster 123456
sentinel config-epoch mymaster 89
daemonize yes
dir "/usr/redis/bin"
sentinel known-slave mymaster 192.168.20.108 6379
sentinel known-slave mymaster 192.168.20.29 6379
sentinel known-sentinel mymaster 192.168.20.105 26379 a8d94fbe0f1120278136c84a612ea84d57d2501c
# Generated by CONFIG REWRITE
sentinel known-sentinel mymaster 192.168.20.29 26379 b1fba42ff8b363b2aea5c52e5dfa2c1ae52869e9
4.sentinel支持集群(废话)
即使有一些sentinel进程宕掉了，依然可以进行redis集群的主备切换；
如果只有一个sentinel进程，如果这个进程运行出错，或者是网络堵塞，那么将无法实现redis集群的主备切换（单点问题）;
如果有多个sentinel，redis的客户端可以随意地连接任意一个sentinel来获得关于redis集群中的信息。
以上配置文件除加了注释的地方外，其他地方基本都是redis自动生成的，只要注意一下加注释的点即可。
可能在配置过程中会出现故障不能切换的问题，极大可能是因为在原主节点配置文件中没有配置主节点密码，因为sentinel不会自动给主节点配置文件加上新的主节点密码。
Redis目前的复制是异步的，只保证最终一致性，而不是强一致性（主从数据库的更新还是分先后，先主后从）。要是一致性要求高的应用，目前还是读写都在主库上。
如果想要一句话讲清楚sentinel实现redis切换的原理那就是：sentinel心跳检测到主节点出现异常后，通过修改redis.conf配置文件的形式实现redis主从节点之间的故障切换。