Redis以主从的模式搭建集群后,如果主节点Master挂掉,虽然可以实现将备用节点Slave切换成主节点,但是Redis本身并没有自动监控机制,需要借助Sentinel哨兵模式,实现监控并实现自动切换。为了实现Sentinel的高可用,需要sentinel也以集群模式来搭建,这里通过一台机器的不同端口来模拟。相关环境信息如下:
1、Redis集群信息:
角色 | IP地址 | 监听端口 |
Master | 127.0.0.1 | 6379 |
Slave | 127.0.0.1 | 6380 |
Slave | 127.0.0.1 | 6381 |
Sentinel集群信息:
哨兵角色 | IP地址 | 监听端口 |
Node-1 | 127.0.0.1 | 26379 |
Node-2 | 127.0.0.1 | 26380 |
Node-3 | 127.0.0.1 | 26381 |
2、Redis集群搭建过程参考上篇,这里不再描述。首先启动Redis的主节点和两个Slave节点。
进入Master节点,查看信息:
[root@VM_0_14_centos bin]# ./redis-cli -h 127.0.0.1 -p -a funnyboy
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
127.0.0.1:> info replication
# Replication
role:master
connected_slaves:
slave0:ip=127.0.0.1,port=,state=online,offset=,lag=
slave1:ip=127.0.0.1,port=,state=online,offset=,lag=
master_replid:d5802af0905736ae28201050ce4871ee2921c16c
master_replid2:
master_repl_offset:
second_repl_offset:-
repl_backlog_active:
repl_backlog_size:
repl_backlog_first_byte_offset:
repl_backlog_histlen:
127.0.0.1:>
显示当前节点role为master,并且连接的slave个数为2,OK。
分别进入两个slave节点查看信息:
[root@VM_0_14_centos bin]# ./redis-cli -h 127.0.0.1 -p -a funnyboy
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
127.0.0.1:> info replication
# Replication
role:slave
master_host:127.0.0.1
master_port:
master_link_status:up
master_last_io_seconds_ago:
master_sync_in_progress:
slave_repl_offset:
slave_priority:
slave_read_only:
connected_slaves:
master_replid:d5802af0905736ae28201050ce4871ee2921c16c
master_replid2:
master_repl_offset:
second_repl_offset:-
repl_backlog_active:
repl_backlog_size:
repl_backlog_first_byte_offset:
repl_backlog_histlen:
127.0.0.1:>
[root@VM_0_14_centos bin]# ./redis-cli -h 127.0.0.1 -p -a funnyboy
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
127.0.0.1:> info replication
# Replication
role:slave
master_host:127.0.0.1
master_port:
master_link_status:up
master_last_io_seconds_ago:
master_sync_in_progress:
slave_repl_offset:
slave_priority:
slave_read_only:
connected_slaves:
master_replid:d5802af0905736ae28201050ce4871ee2921c16c
master_replid2:
master_repl_offset:
second_repl_offset:-
repl_backlog_active:
repl_backlog_size:
repl_backlog_first_byte_offset:
repl_backlog_histlen:
127.0.0.1:>
角色都为slave,并且master信息正常。确保Redis集群OK后,开始准备搭建sentinel集群。
3、Sentinel集群搭建
step1、将redis-sentinel拷贝到redis对应的执行目录bin(该命令是redis-server的一个连接,用redis-server也OK,后面讲到,启动sentinel会有两种方式),然后拷贝sentinel.config到redis配置文件目录config。
step2、分别编辑sentinel的配置文件(没特别强调的保持默认即可)
sentinel-26379.conf配置如下:
daemonize yes #开启后台守护进程
port #端口配置
pidfile "/usr/local/redis/pid/redis-sentinel-26379.pid" #PID文件
logfile "/usr/local/redis/logs/sentinel-26379.log" #日志文件
sentinel monitor mymaster 127.0.0.1 #哨兵监控配置。注意,如果配置了认证,改配置必须在auth-pass配置之前,否则启动报找不到master的错误
sentinel auth-pass mymaster funnyboy #认证配置
sentinel down-after-milliseconds mymaster #master或者slave多少时间(默认30秒)不能使用标记为down状态。
sentinel failover-timeout mymaster #若哨兵在配置值内未能完成故障转移操作,则任务本次故障转移失败。
sentinel-26380.conf配置如下:
daemonize yes #开启后台守护进程
port #端口配置
pidfile "/usr/local/redis/pid/redis-sentinel-26380.pid" #PID文件
logfile "/usr/local/redis/logs/sentinel-26380.log" #日志文件
sentinel monitor mymaster 127.0.0.1 #哨兵监控配置。注意,如果配置了认证,改配置必须在auth-pass配置之前,否则启动报找不到master的错误
sentinel auth-pass mymaster funnyboy #认证配置
sentinel down-after-milliseconds mymaster #master或者slave多少时间(默认30秒)不能使用标记为down状态。
sentinel failover-timeout mymaster #若哨兵在配置值内未能完成故障转移操作,则任务本次故障转移失败。
sentinel-26381.conf配置如下:
daemonize yes #开启后台守护进程
port #端口配置
pidfile "/usr/local/redis/pid/redis-sentinel-26381.pid" #PID文件
logfile "/usr/local/redis/logs/sentinel-26381.log" #日志文件
sentinel monitor mymaster 127.0.0.1 #哨兵监控配置。注意,如果配置了认证,改配置必须在auth-pass配置之前,否则启动报找不到master的错误
sentinel auth-pass mymaster funnyboy #认证配置
sentinel down-after-milliseconds mymaster #master或者slave多少时间(默认30秒)不能使用标记为down状态。
sentinel failover-timeout mymaster #若哨兵在配置值内未能完成故障转移操作,则任务本次故障转移失败。
step3、启动哨兵监控程序:
[root@VM_0_14_centos redis]# ./bin/redis-sentinel ./config/sentinel-.conf
[root@VM_0_14_centos redis]# ./bin/redis-sentinel ./config/sentinel-.conf
[root@VM_0_14_centos redis]# ./bin/redis-sentinel ./config/sentinel-.conf
启动有两种方式:
一是执行:redis-sentinel sentinel.conf
二是执行:redis-server sentinel --sentinel
step4、通过哨兵连接,并检查信息:
[root@VM_0_14_centos bin]# ./redis-cli -h 127.0.0.1 -p -a funnyboy
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
127.0.0.1:> info sentinel
# Sentinel
sentinel_masters:
sentinel_tilt:
sentinel_running_scripts:
sentinel_scripts_queue_length:
sentinel_simulate_failure_flags:
master0:name=mymaster,status=ok,address=127.0.0.1:,slaves=,sentinels=
127.0.0.1:>
可以看到监控的redis服务,一个Master、两个Slave、sentinels = 3 说明配置OK。
4、模拟场景:Redis Master节点挂掉,查看Redis集群状态。
step1、关掉Master节点:
[root@VM_0_14_centos bin]# ./redis-cli -p -a funnyboy
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
127.0.0.1:> shutdown
not connected>
[root@VM_0_14_centos bin]#
step2、通过哨兵查看集群状态:
127.0.0.1:> info sentinel
# Sentinel
sentinel_masters:
sentinel_tilt:
sentinel_running_scripts:
sentinel_scripts_queue_length:
sentinel_simulate_failure_flags:
master0:name=mymaster,status=ok,address=127.0.0.1:,slaves=,sentinels=
127.0.0.1:>
通过sentinel信息可以看到,Master节点已经自动切换到6380端口了,说明主节点挂掉后,6380 Slave节点自动升级成为了Master节点。
step3、启动6379 redis服务,然后查看节点角色,此时6379变成了Slave,6380为Master节点,OK。
[root@VM_0_14_centos redis]# ./bin/redis-server ./config/redis-.conf
[root@VM_0_14_centos redis]# ./bin/redis-cli -p -a funnyboy
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
127.0.0.1:>
127.0.0.1:> info replication
# Replication
role:slave
master_host:127.0.0.1
master_port:
master_link_status:up
master_last_io_seconds_ago:
master_sync_in_progress:
slave_repl_offset:
slave_priority:
slave_read_only:
connected_slaves:
master_replid:c77763408266bcebf233bdc9e59e3bcf14dc7a08
master_replid2:
master_repl_offset:
second_repl_offset:-
repl_backlog_active:
repl_backlog_size:
repl_backlog_first_byte_offset:
repl_backlog_histlen: