公司新项目需要使用redis集群，综合考虑了一些方案，最后选择了Redis sentinel，
先在虚拟机部署测试环境如下：
sentinel 2台
redis 实例 3个
部分配置如下，采用的默认配置，记得开启认证及最大内存限制，即使是测试环境也要开启认证，
好的习惯是安全生产第一步。
redis-6380：
bind 192.168.162.128
protected-mode yes
port 6380
tcp-backlog 511
timeout 0
tcp-keepalive 300
daemonize yes
supervised no
pidfile "/var/run/redis_6380.pid"
loglevel notice
logfile "/var/log/redis_6380.log"
masterauth "123456"
slave-serve-stale-data yes
slave-read-only yes
repl-diskless-sync no
repl-diskless-sync-delay 5
repl-disable-tcp-nodelay no
slave-priority 100
requirepass "123456"
maxmemory 50M
appendonly no
appendfilename "appendonly.aof"
redis-6381：
bind 192.168.162.128
protected-mode yes
port 6381
tcp-backlog 511
timeout 0
slaveof 192.168.162.128 6380
masterauth 123456
slave-serve-stale-data yes
slave-read-only yes
repl-diskless-sync no
repl-diskless-sync-delay 5
repl-disable-tcp-nodelay no
slave-priority 100
requirepass 123456
maxmemory 50M
sentinel1.conf
daemonize yes
port 26379
dir "/tmp"
sentinel monitor mymaster 192.168.162.128 6380 2
sentinel down-after-milliseconds mymaster 6000
sentinel failover-timeout mymaster 60000
sentinel auth-pass mymaster 123456
sentinel parallel-syncs mymaster 1
logfile "/var/log/sentinel1.log"
protected-mode no
启动后sentinel 会自动修改配置文件，修改后如下
[root@linux-node1 redis]# cat sentinel1.conf
daemonize yes
port 26379
dir "/tmp"
sentinel myid b777f8a0c7f7cf219804a3f140535a58ed067b96
sentinel monitor mymaster 192.168.162.128 6380 2
sentinel down-after-milliseconds mymaster 6000
sentinel failover-timeout mymaster 60000
sentinel auth-pass mymaster 123456
logfile "/var/log/sentinel1.log"
protected-mode no
# Generated by CONFIG REWRITE
sentinel config-epoch mymaster 0
sentinel leader-epoch mymaster 0
sentinel known-slave mymaster 192.168.162.130 6380
sentinel known-slave mymaster 192.168.162.128 6381
sentinel known-sentinel mymaster 192.168.162.130 26379 94680c01a0112bb24d7cc131bd37b388da00b087
sentinel current-epoch 0
================我是分割线================
redis-6380：
bind 192.168.162.130
protected-mode yes
port 6380
slaveof 192.168.162.128 6380
masterauth 123456
slave-serve-stale-data yes
slave-read-only yes
repl-diskless-sync no
repl-diskless-sync-delay 5
repl-disable-tcp-nodelay no
slave-priority 100
requirepass "123456"
maxmemory 50M
sentinel1.conf
daemonize yes
port 26379
dir "/tmp"
sentinel monitor mymaster 192.168.162.128 6380 2
sentinel down-after-milliseconds mymaster 6000
sentinel failover-timeout mymaster 60000
sentinel auth-pass mymaster 123456
sentinel parallel-syncs mymaster 1
logfile "/var/log/sentinel1.log"
protected-mode no
启动后被自动修改为：
[root@linux-node2 redis]# cat sentinel1.conf
daemonize yes
port 26379
dir "/tmp"
sentinel myid 94680c01a0112bb24d7cc131bd37b388da00b087
sentinel monitor mymaster 192.168.162.128 6380 2
sentinel down-after-milliseconds mymaster 6000
sentinel failover-timeout mymaster 60000
sentinel auth-pass mymaster 123456
logfile "/var/log/sentinel1.log"
protected-mode no
# Generated by CONFIG REWRITE
sentinel config-epoch mymaster 0
sentinel leader-epoch mymaster 0
sentinel known-slave mymaster 192.168.162.130 6380
sentinel known-slave mymaster 192.168.162.128 6381
sentinel known-sentinel mymaster 192.168.162.128 26379 b777f8a0c7f7cf219804a3f140535a58ed067b96
sentinel current-epoch 0
================我是分割线================
注意事项：
- 如果多台机器redis实例端口有相同的如我都有6380端口， bind地址不要写成127.0.0.1，指定为本机ip，sentinel进行管理时会修改配置文件，将其他机器的相同端口的bing ip 修改为127.0.0.1（如果前面写了127，此处将是个坑）。
- redis、sentinel需要开启认证。
- 已经挂断的实例如果是master的话，再次启动时记得修改配置文件，将slaveof指向当前的master，并注意同步rdb或者aof文件。
- 启动宕掉的实例时不要用这种方式启动/usr/local/bin/redis-server /etc/redis/6380.conf --slaveof 192.168.162.128 6381 ，如果这种方式启动的话，配置会被修改为127.0.0.1
- 多sentinel 时记得protected-mode no ，否则多哨兵无法通知对方进行决策Sdown、Odown。
