本文共 5317 字，大约阅读时间需要 17 分钟。

redis复制可以实现主机（master）和从机（salve）之间数据同步，这样做可以实现两个目标：读写分离和数据备份。

“复制”是一种架构思想，redis实现了这种思想，除此之外，其他数据库例如MySQL也支持复制。

redis复制的实现

主从配置和状态

• 配从不配主

  • 配置从库即可，不用配置主库
  • 从库配置使用 slaveof <host> <port>
  • 取消slave配置 slaveof no one

• 默认从库只读

• master宕机后，slave会保持原有角色，等待master恢复，不过slave会显示master的连接状态为down

  • master恢复后，slave会自动连接

127.0.0.1:6380> info replication# Replicationrole:slavemaster_host:127.0.0.1master_port:6379master_link_status:down...

• 在没有配置文件配置配置的情况下，slave宕机之后重启会变为新的master，需要重新连接原来的master才可以继续工作

slave传递（薪火相传）

• slave同样接受其他服务的slaveof请求，使slave可以传递，从而降低master的压力
• 有slave的slave角色仍然时slave
• slave传递时，数据的同步会有延时

# 有slave的slave复制信息127.0.0.1:6380> info replication# Replicationrole:slavemaster_host:127.0.0.1master_port:6379master_link_status:upmaster_last_io_seconds_ago:5master_sync_in_progress:0slave_repl_offset:2670slave_priority:100slave_read_only:1connected_slaves:1slave0:ip=127.0.0.1,port=6381,state=online,offset=2670,lag=0master_failover_state:no-failovermaster_replid:0e3ec53a7c32143e7ce760250c80a5f806892b0fmaster_replid2:0000000000000000000000000000000000000000master_repl_offset:2670second_repl_offset:-1repl_backlog_active:1repl_backlog_size:1048576repl_backlog_first_byte_offset:1repl_backlog_histlen:2670

slave变master(反客为主)

• master宕机后，需要手动选择一个slave（记作slave1）设置为master

• 其他slave默认会等待master恢复，需要手动转移到新的master

• 使用哨兵模式可以实现master宕机时自动将slave变为master

哨兵模式详见作者的另一篇博文

示例

一主两从

环境准备

• 在同一台设备运行三个redis实例，端口分别为6379、6380、6381

• 拷贝三份redis.conf配置文件，分别修改以下属性

• 启动三个redis服务

$ ./redis-server xxx/redis_6379.conf$ ./redis-server xxx/redis_6380.conf$ ./redis-server xxx/redis_6381.conf

• 至此，三个redis服务同时启动，完成了一台主机运行多个redis服务
• 客户端连接服务

$ redis-cli -p 6379$ redis-cli -p 6380$ redis-cli -p 6381

客户端操作

• 查看复制信息 info replication

127.0.0.1:6379> info replication# Replicationrole:masterconnected_slaves:0master_failover_state:no-failovermaster_replid:e4a19fe31c0f617b3da3a3255e0b63aa03722c4dmaster_replid2:0000000000000000000000000000000000000000master_repl_offset:0second_repl_offset:-1repl_backlog_active:0repl_backlog_size:1048576repl_backlog_first_byte_offset:0repl_backlog_histlen:0127.0.0.1:6380> info replication# Replicationrole:masterconnected_slaves:0master_failover_state:no-failovermaster_replid:67e5beb30c793be37999063b9f9eb9470eae8842master_replid2:0000000000000000000000000000000000000000master_repl_offset:0second_repl_offset:-1repl_backlog_active:0repl_backlog_size:1048576repl_backlog_first_byte_offset:0repl_backlog_histlen:0127.0.0.1:6381> info replication# Replicationrole:masterconnected_slaves:0master_failover_state:no-failovermaster_replid:126f4a1cfc0ede33ad893a9a459a717b078eef0fmaster_replid2:0000000000000000000000000000000000000000master_repl_offset:0second_repl_offset:-1repl_backlog_active:0repl_backlog_size:1048576repl_backlog_first_byte_offset:0repl_backlog_histlen:0

三个redis实例角色均为 master

• 将6380和6381设置为6379的slave

step 3～5: 证明6380和6381已经成功成为6379的slave，并且数据已经同步（添加为slave之前的数据也会同步）

step 6: 默认情况下slave为只读，无法写入数据

此时查看复制信息 info replication

127.0.0.1:6379> info replication# Replicationrole:master  # 6379为masterconnected_slaves:2slave0:ip=127.0.0.1,port=6380,state=online,offset=1092,lag=1slave1:ip=127.0.0.1,port=6381,state=online,offset=1092,lag=0master_failover_state:no-failovermaster_replid:36f2fe0ee9028ba075cdba2bb820b9803221d3b4master_replid2:0000000000000000000000000000000000000000master_repl_offset:1092second_repl_offset:-1repl_backlog_active:1repl_backlog_size:1048576repl_backlog_first_byte_offset:1repl_backlog_histlen:1092127.0.0.1:6379> 127.0.0.1:6380> info replication# Replicationrole:slave # 6380为slavemaster_host:127.0.0.1master_port:6379master_link_status:upmaster_last_io_seconds_ago:6master_sync_in_progress:0slave_repl_offset:1064slave_priority:100slave_read_only:1connected_slaves:0master_failover_state:no-failovermaster_replid:36f2fe0ee9028ba075cdba2bb820b9803221d3b4master_replid2:0000000000000000000000000000000000000000master_repl_offset:1064second_repl_offset:-1repl_backlog_active:1repl_backlog_size:1048576repl_backlog_first_byte_offset:1repl_backlog_histlen:1064127.0.0.1:6381> info replication# Replicationrole:slave # 6380为slavemaster_host:127.0.0.1master_port:6379master_link_status:upmaster_last_io_seconds_ago:7master_sync_in_progress:0slave_repl_offset:1022slave_priority:100slave_read_only:1connected_slaves:0master_failover_state:no-failovermaster_replid:36f2fe0ee9028ba075cdba2bb820b9803221d3b4master_replid2:0000000000000000000000000000000000000000master_repl_offset:1022second_repl_offset:-1repl_backlog_active:1repl_backlog_size:1048576repl_backlog_first_byte_offset:155repl_backlog_histlen:868

转载地址：https://console.blog.csdn.net/article/details/115413401 如侵犯您的版权，请留言回复原文章的地址，我们会给您删除此文章，给您带来不便请您谅解！