Redis高可用
为了保证redis最大程度上能够使用，redis提供了主从同步+Sentinel哨兵机制。
Sentinel 哨兵
下面是官网地址：https://redis.io/topics/sentinel
redis提供的哨兵是用来看护redis实例进程的，可以自动进行故障转移，其功能如下：
- Monitoring. Sentinel constantly checks if your master and replica instances are working as expected.
- Notification. Sentinel can notify the system administrator, or other computer programs, via an API, that something is wrong with one of the monitored Redis instances.
- Automatic failover. If a master is not working as expected, Sentinel can start a failover process where a replica is promoted to master, the other additional replicas are reconfigured to use the new master, and the applications using the Redis server are informed about the new address to use when connecting.
- Configuration provider. Sentinel acts as a source of authority for clients service discovery: clients connect to Sentinels in order to ask for the address of the current Redis master responsible for a given service. If a failover occurs, Sentinels will report the new address.
即：
- 监测。Sentinel不断检查主实例和从实例是否按预期工作。
- 通知。Sentinel可以通过API通知系统管理员和其他计算机程序，其中一个被监视的Redis实例有问题。
- 自动故障切换。如果主设备不能按预期工作，Sentinel可以启动一个故障转移过程，其中一个从设备升级为主设备，其他附加的从设备被重新配置为使用新的主设备，并且使用Redis服务器的应用程序在连接时被告知要使用的新地址。
- 配置提供程序。Sentinel充当客户端服务发现的授权源：客户端连接到Sentinel，以请求负责给定服务的当前Redis主机的地址。如果发生故障转移，哨兵将报告新地址
在redis安装后，会自带sentinel哨兵程序，
修改sentinel.conf配置文件
bind 127.0.0.1 port 26380 daemonize yes logfile /var/log/redis-sentinel.log sentinel monitor mymaster 127.0.0.1 6380 2 sentinel down-after-milliseconds mymaster 30000 sentinel parallel-syncs mymaster 1 sentinel failover-timeout mymaster 180000
- sentinel monitor mymaster 127.0.0.1 6380 2 说明
- mymaster 为sentinel监护的redis主从集群起名
- 127.0.0.1 6300 为主从中任一台机器地址
- 2 表示有两台以的sentinel认为某一台redis宕机后，才会进行自动故障转移。
启动方式：
redis-sentinel sentinel.conf
高可用方案注意事项
- 至少三个sentinel以上
- sentinel要分散运行在不同的机器上
测试
1：查看所有redis进程
[root@peiqi bin]# ps aux | grep redis root 1714 0.1 0.2 147460 10104 ? Sl 08:53 0:24 redis-server 192.168.50.129:6379 root 1753 0.1 0.2 147460 10092 ? Sl 08:54 0:24 redis-server 192.168.50.129:6378 root 1792 0.1 0.3 149508 11884 ? Ssl 08:54 0:24 redis-server 192.168.50.129:7001 [cluster] root 1797 0.1 0.3 149508 11908 ? Ssl 08:55 0:24 redis-server 192.168.50.129:7002 [cluster] root 1802 0.1 0.3 149508 11900 ? Ssl 08:55 0:24 redis-server 192.168.50.129:7003 [cluster] root 44789 0.1 0.3 149508 11916 ? Ssl 10:00 0:21 redis-server 192.168.50.129:7004 [cluster] root 44794 0.1 0.3 149508 11916 ? Ssl 10:01 0:20 redis-server 192.168.50.129:7005 [cluster] root 44812 0.1 0.3 149508 11948 ? Ssl 10:01 0:20 redis-server 192.168.50.129:7006 [cluster] root 45446 0.2 0.2 145412 7752 ? Ssl 15:31 0:00 redis-sentinel 192.168.50.129:26379 [sentinel] root 45451 0.3 0.2 145412 7748 ? Ssl 15:31 0:00 redis-sentinel 192.168.50.129:26380 [sentinel] root 45456 1.0 0.2 145412 7704 ? Ssl 15:31 0:00 redis-sentinel 192.168.50.129:26381 [sentinel] root 45461 0.0 0.0 112812 976 pts/2 S+ 15:31 0:00 grep --color=auto redis [root@peiqi bin]
2：进入主节点，查看主/从复制信息（主节点）
[root@peiqi bin]# redis-cli -h 192.168.50.129 -p 6379 192.168.50.129:6379> info Replication # Replication role:master connected_slaves:1 slave0:ip=192.168.50.129,port=6378,state=online,offset=62446,lag=1 master_replid:551ecb2b80a1cf9de0b97f73ee69fb2cd5128724 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:62732 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:62732 192.168.50.129:6379>
2：进入从节点，查看主/从复制信息（从节点）
[root@peiqi cluster]# redis-cli -h 192.168.50.129 -p 6378 192.168.50.129:6378> info Replication # Replication role:slave master_host:192.168.50.129 master_port:6379 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_repl_offset:71511 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:551ecb2b80a1cf9de0b97f73ee69fb2cd5128724 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:71511 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:71511 192.168.50.129:6378>
3：关掉主节点，查看从节点是自动切换为主节点（这儿要稍微等会，因为配有时间参数）
192.168.50.129:6379> shutdown
not connected>
等会参看从从节点（已将从节点切换为主节点），当开启主节点后，从节点又开始切换回从节点（这是网上说的，但我实践出来的理论是会把原来的主节点变成从节点）
192.168.50.129:6378> info Replication # Replication role:master connected_slaves:0 master_replid:e8b08994005adae4e0520482ac7a08af66a91352 master_replid2:551ecb2b80a1cf9de0b97f73ee69fb2cd5128724 master_repl_offset:146098 second_repl_offset:144360 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:146098
注意：这儿从节点切换为主节的时候会有个个问题！
因为从节点只能都信息，为不能写信息，这是在python中的处理方式如下：
Python客户端使用
# redis 哨兵 REDIS_SENTINELS = [ (‘127.0.0.1‘, ‘26380‘), (‘127.0.0.1‘, ‘26381‘), (‘127.0.0.1‘, ‘26382‘), ] REDIS_SENTINEL_SERVICE_NAME = ‘mymaster‘ from redis.sentinel import Sentinel _sentinel = Sentinel(REDIS_SENTINELS) redis_master = _sentinel.master_for(REDIS_SENTINEL_SERVICE_NAME) redis_slave = _sentinel.slave_for(REDIS_SENTINEL_SERVICE_NAME) 使用示例 # 读数据，master读不到去slave读 try: real_code = redis_master.get(key) except ConnectionError as e: real_code = redis_slave.get(key) # 写数据，只能在master里写 try: current_app.redis_master.delete(key) except ConnectionError as e: logger.error(e)