Centos7搭建Redis主从复制及sentinel监控
Centos7搭建Redis主从复制及sentinel监控
- 主从复制
- sentinel监控，实现故障转移failover
主从复制
本文介绍如何搭建Redis一主二从环境，需要先准备3个Redis单实例节点环境。
单实例节点环境搭建请参考：Centos7安装Redis单节点 https://blog.csdn.net/iceliooo/article/details/102892062
| IP | 角色：端口 |
|---|---|
| 192.168.56.105 | Master：6379 |
| 192.168.56.106 | replica：6379 |
| 192.168.56.107 | replica：6379 |
修改106及107节点redis.conf配置，添加如下配置信息：
286 # replicaof
287 replicaof 192.168.56.105 6379
重启105，106， 107 节点Redis服务，查看主从复制状态：
105Master节点：
[root@k8s-master src]# redis-server ../redis.conf
23141:C 04 Nov 2019 14:37:31.847 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
23141:C 04 Nov 2019 14:37:31.847 # Redis version=5.0.5, bits=64, commit=00000000, modified=0, pid=23141, just started
23141:C 04 Nov 2019 14:37:31.847 # Configuration loaded
[root@k8s-master src]# redis-cli
127.0.0.1:6379> info replication
NOAUTH Authentication required.
127.0.0.1:6379> auth ******
OK
127.0.0.1:6379> info replication
# Replication
role:master
connected_slaves:0
master_replid:c3bda75e828e43caee07f27dc191f98e661b82c2
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:0
second_repl_offset:-1
repl_backlog_active:0
repl_backlog_size:1048576
repl_backlog_first_byte_offset:0
repl_backlog_histlen:0
106 replica节点：
[root@k8s-node1 src]# redis-server ../redis.conf
13197:C 04 Nov 2019 14:38:40.254 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
13197:C 04 Nov 2019 14:38:40.254 # Redis version=5.0.5, bits=64, commit=00000000, modified=0, pid=13197, just started
13197:C 04 Nov 2019 14:38:40.254 # Configuration loaded
[root@k8s-node1 src]# redis-cli
127.0.0.1:6379> info replication
NOAUTH Authentication required.
127.0.0.1:6379> auth ******
OK
127.0.0.1:6379> info repilication
127.0.0.1:6379> info replication
# Replication
role:slave
master_host:192.168.56.105
master_port:6379
master_link_status:down
master_last_io_seconds_ago:-1
master_sync_in_progress:0
slave_repl_offset:1
master_link_down_since_seconds:1572849566
slave_priority:100
slave_read_only:1
connected_slaves:0
master_replid:c1b5f8a15f72d059bc1d523a12c4a4380391c35d
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:0
second_repl_offset:-1
repl_backlog_active:0
repl_backlog_size:1048576
repl_backlog_first_byte_offset:0
repl_backlog_histlen:0
发现主从配置失败，主要是因为Master节点配置了访问密码，所以需要在replica节点配置Master节点的访问秘密：
294 # masterauth 重启Master及replica节点，主从配置成功：
127.0.0.1:6379> info replication
# Replication
role:master
connected_slaves:2
slave0:ip=192.168.56.106,port=6379,state=online,offset=254,lag=1
slave1:ip=192.168.56.107,port=6379,state=online,offset=254,lag=1
master_replid:c4b26a9a991c54e3110d3cf9479d997937532002
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:254
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:254
127.0.0.1:6379> info replication
# Replication
role:slave
master_host:192.168.56.105
master_port:6379
master_link_status:up
master_last_io_seconds_ago:2
master_sync_in_progress:0
slave_repl_offset:56
slave_priority:100
slave_read_only:1
connected_slaves:0
master_replid:c4b26a9a991c54e3110d3cf9479d997937532002
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:56
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:56
Master节点执行命令，replica节点成功接收
127.0.0.1:6379> set iceliooo male
OK
127.0.0.1:6379>
从节点默认为自读
127.0.0.1:6379> get iceliooo
"male"
127.0.0.1:6379> set iceliooo male
(error) READONLY You can't write against a read only replica.
127.0.0.1:6379>
sentinel监控，实现故障转移failover
主从复制解决了数据备份问题，提升了性能，但Master节点不可用，将出现单点故障，需要手动切换Master节点，费时费力，还会造成服务不可用问题。
下面介绍如何通过sentinel实现Master自动切换。
sentinel监控主从集群中所有Redis服务节点，sentinel本身也会存在单点故障问题，sentinel通过相互监控来解决。sentinel本身没有主从之分。
Redis主从及Sentinel部署分布
| IP | 角色：端口 |
|---|---|
| 192.168.56.105 | Master：6379/sentinel：26379 |
| 192.168.56.106 | replica：6379 /sentinel：26379 |
| 192.168.56.107 | replica：6379/sentinel：26379 |
修改sentinel.conf配置文件
[root@k8s-master redis-5.0.5]# cp sentinel.conf sentinel.conf.bak
[root@k8s-master redis-5.0.5]# vi sentinel.conf
主要修改如下配置，其它使用默认值：
daemonize yes
port 16379
protected-mode no
dir "/usr/local/software/redis-5.0.5/sentinel-temp"
sentinel monitor redis-master 192.168.56.105 6379 2
启动3台服务器的redis服务和sentinel服务
[root@k8s-master redis-5.0.5]# cd src/
[root@k8s-master src]# redis-sentinel ../sentinel.conf
2489:X 04 Nov 2019 17:12:28.225 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
2489:X 04 Nov 2019 17:12:28.225 # Redis version=5.0.5, bits=64, commit=00000000, modified=0, pid=2489, just started
2489:X 04 Nov 2019 17:12:28.225 # Configuration loaded
[root@k8s-master src]# ps -ef|grep redis
root 2490 1 0 17:12 ? 00:00:00 redis-sentinel *:16379 [sentinel]
root 3064 11611 0 17:13 pts/1 00:00:00 grep --color=auto redis
root 29488 1 0 14:50 ? 00:00:09 redis-server *:6379
[root@k8s-master src]#
模拟Master故障：
106上查看集群状态：
127.0.0.1:6379> auth centos
OK
127.0.0.1:6379> info replication
# Replication
role:slave
master_host:192.168.56.105
master_port:6379
master_link_status:up
master_last_io_seconds_ago:7
master_sync_in_progress:0
slave_repl_offset:12465
slave_priority:100
slave_read_only:1
connected_slaves:0
master_replid:c4b26a9a991c54e3110d3cf9479d997937532002
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:12465
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:12465
127.0.0.1:6379>
关闭Master节点：
127.0.0.1:6379> SHUTDOWN
not connected>
再次查看106上集群状态，30秒之后发现master节点切换失败：
127.0.0.1:6379> info replication
# Replication
role:slave
master_host:192.168.56.105
master_port:6379
master_link_status:down
master_last_io_seconds_ago:-1
master_sync_in_progress:0
slave_repl_offset:12605
master_link_down_since_seconds:35
slave_priority:100
slave_read_only:1
connected_slaves:0
master_replid:c4b26a9a991c54e3110d3cf9479d997937532002
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:12605
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:12605
127.0.0.1:6379>
失败原因主要是没有在sentinel.conf配置文件中配置Redis集群节点的访问密码（这里要求集群中每个节点的访问密码必须一致）
121 sentinel monitor mymaster 192.168.56.107 6379 2
122 sentinel auth-pass mymaster ******
这里还需要注意，“sentinel auth-pass mymaster ******” 必须配置在“sentinel monitor mymaster 192.168.56.107 6379 2” 之后，否则会出现如下异常：
[root@k8s-master src]# redis-sentinel ../sentinel.conf
*** FATAL CONFIG FILE ERROR ***
Reading the configuration file, at line 104
>>> 'sentinel auth-pass mymaster ******'
No such master with specified name.
重新启动各个节点的redis-sentinel服务及redis服务，106上查看集群状态：
127.0.0.1:6379> info replication
# Replication
role:slave
master_host:192.168.56.105
master_port:6379
master_link_status:up
master_last_io_seconds_ago:1
master_sync_in_progress:0
slave_repl_offset:1596
slave_priority:100
slave_read_only:1
connected_slaves:0
master_replid:09787ba9f25f74e8caac6c07808997c7d3d4a421
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:1596
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:1596
关闭Master节点，模拟Master节点故障：
127.0.0.1:6379> info replication
# Replication
role:master
connected_slaves:2
slave0:ip=192.168.56.106,port=6379,state=online,offset=6786,lag=0
slave1:ip=192.168.56.107,port=6379,state=online,offset=6786,lag=1
master_replid:09787ba9f25f74e8caac6c07808997c7d3d4a421
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:6786
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:6786
127.0.0.1:6379> SHUTDOWN
not connected>
30秒后，106上查看集群状态，切换成功，107成为新的Master节点：
127.0.0.1:6379> info replication
# Replication
role:slave
master_host:192.168.56.107
master_port:6379
master_link_status:up
master_last_io_seconds_ago:0
master_sync_in_progress:0
slave_repl_offset:11842
slave_priority:100
slave_read_only:1
connected_slaves:0
master_replid:569347a188791772e9f7ff3fa11048d0e6a86b01
master_replid2:09787ba9f25f74e8caac6c07808997c7d3d4a421
master_repl_offset:11842
second_repl_offset:8517
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:11842
启动105redis节点，模拟原Master节点恢复，新Master中恢复到两个replica节点：
127.0.0.1:6379> info replication
# Replication
role:master
connected_slaves:2
slave0:ip=192.168.56.106,port=6379,state=online,offset=478823,lag=1
slave1:ip=192.168.56.105,port=6379,state=online,offset=478966,lag=0
master_replid:569347a188791772e9f7ff3fa11048d0e6a86b01
master_replid2:09787ba9f25f74e8caac6c07808997c7d3d4a421
master_repl_offset:478966
second_repl_offset:8517
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:478966