redis主从复制

科技2023-10-26 76

//从库配置 slaveof 主库IP 主库端口 //取消从库配置，使当前数据库停止与其他数据库的同步，转化为主数据库 slaveof no one //主库 127.0.0.1:6379> info replication # Replication role:master connected_slaves:0 master_replid:7f39ff900c939bf7795372bf8160ff31aa0fd26e master_replid2:f97aced19f0231382f5c223b557c22d6a12d5e70 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 //从库一 [root@centos bin]# ./redis-server ./6380.conf [root@centos bin]# ./redis-cli -p 6380 //查看从库的信息 127.0.0.1:6380> info replication # Replication role:master connected_slaves:0 master_replid:6bd8662e4fec3c530d4f5c2ecc9d4110b9da90a9 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 //从库配置一 127.0.0.1:6380> slaveof 127.0.0.1 6379 OK //再次查看从库信息 127.0.0.1:6380> info replication # Replication role:slave master_host:127.0.0.1 master_port:6379 master_link_status:up master_last_io_seconds_ago:6 master_sync_in_progress:0 slave_repl_offset:1608 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:0a026d10eda541c2cb2e00b21099ad5a830881d2 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:1608 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:1608 //从库二 [root@centos bin]# ./redis-server ./6381.conf [root@centos bin]# ./redis-cli //配置从库二 127.0.0.1:6379> slaveof 127.0.0.1 6379 OK Already connected to specified master //再次查看从库信息 127.0.0.1:6381> info replication # Replication role:slave master_host:127.0.0.1 master_port:6379 master_link_status:up master_last_io_seconds_ago:6 master_sync_in_progress:0 slave_repl_offset:1790 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:0a026d10eda541c2cb2e00b21099ad5a830881d2 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:1790 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:29 repl_backlog_histlen:1762 //再次查看主库信息 127.0.0.1:6379> info replication # Replication role:master connected_slaves:2 slave0:ip=127.0.0.1,port=6380,state=online,offset=417,lag=1 //从库一 slave1:ip=127.0.0.1,port=6381,state=online,offset=417,lag=1 //从库二 master_replid:0a026d10eda541c2cb2e00b21099ad5a830881d2 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:417 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:417

主库可读也可写(主要写数据)，但是从库只可以读数据

127.0.0.1:6381> set k1 v1 (error) READONLY You can't write against a read only replica.

从库会备份主库所有的数据（并不是只备份配置从库之后时间段的数据）

当主库掉线时（使用shutdown命令，在使用exit命令直接使用exit命令不行），并不会消除和从库的关系，当主库重新上线时依然是之前从库的主库

当从库掉线时，会消除和主库的关系以及之前所有备份的数据（注意备份文件dump.rdb和aof的影响），即当从库重新上线时为master，需要手动重新配置从库，而且此时会复制主库之前所有的数据

slaveof 127.0.0.1 6379

薪火相传

从库既可以作为上一个主库的从库，也可以作为下一个从库的主库

//主库 127.0.0.1:6379> //6379的从库 6381的主库 127.0.0.1:6380> slaveof 127.0.0.1 6379 //6380的从库 127.0.0.1:6381> slaveof 127.0.0.1 6380

反客为主

//当主库下线时 127.0.0.1:6379> shutdown //成为主库 127.0.0.1:6380> slaveof no one //6380的从库 127.0.0.1:6381> slaveof 127.0.0.1 6380

总结：当slaveof启动成功连接到master后会发送一个sync命令，master在接收到命令后会启动后台的存盘进程，同时收集所有接收到用于修改数据集的命令，在后台进程执行完毕之后，master将传送整个数据文件到slave，已完成一次完全同步

全量复制：在slave服务接收到数据库文件之后，将其存盘并加载到内存中

增量复制：master继续将新的所有收集到的命令一次传给slave完成同步

只要重新连接master，一次完全同步（增量复制）将被自动执行

哨兵模式(sentinel)

新建sentinel.conf文件 sentinel monitor 被监控数据库的名字监控主机的ip 票数(最终票数大于票数的数据库成为主库) //启动哨兵模式 ./redis-sentinel ./sentinel.conf 3600:X 17 Jul 2020 13:57:16.215 # Sentinel ID is 0d755d0aeb17d57bdcfac3e5990575f4986e222e 3600:X 17 Jul 2020 13:57:16.215 # +monitor master post6379 127.0.0.1 6379 quorum 1 3600:X 17 Jul 2020 13:57:16.217 * +slave slave 127.0.0.1:6381 127.0.0.1 6381 @ post6379 127.0.0.1 6379 3600:X 17 Jul 2020 13:57:16.219 * +slave slave 127.0.0.1:6380 127.0.0.1 6380 @ post6379 127.0.0.1 6379 //主库下线 127.0.0.1:6379> SHUTDOWN not connected> exit 3600:X 17 Jul 2020 14:00:14.585 # +sdown master post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:14.585 # +odown master post6379 127.0.0.1 6379 #quorum 1/1 3600:X 17 Jul 2020 14:00:14.585 # +new-epoch 1 3600:X 17 Jul 2020 14:00:14.585 # +try-failover master post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:14.586 # +vote-for-leader 0d755d0aeb17d57bdcfac3e5990575f4986e222e 1 3600:X 17 Jul 2020 14:00:14.586 # +elected-leader master post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:14.586 # +failover-state-select-slave master post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:14.662 # +selected-slave slave 127.0.0.1:6381 127.0.0.1 6381 @ post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:14.662 * +failover-state-send-slaveof-noone slave 127.0.0.1:6381 127.0.0.1 6381 @ post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:14.729 * +failover-state-wait-promotion slave 127.0.0.1:6381 127.0.0.1 6381 @ post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:14.915 # +promoted-slave slave 127.0.0.1:6381 127.0.0.1 6381 @ post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:14.915 # +failover-state-reconf-slaves master post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:14.965 * +slave-reconf-sent slave 127.0.0.1:6380 127.0.0.1 6380 @ post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:15.936 * +slave-reconf-inprog slave 127.0.0.1:6380 127.0.0.1 6380 @ post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:15.936 * +slave-reconf-done slave 127.0.0.1:6380 127.0.0.1 6380 @ post6379 127.0.0.1 6379 3600:X 17 Jul 2020 14:00:15.990 # +failover-end master post6379 127.0.0.1 6379 //选举6381为主库 3600:X 17 Jul 2020 14:00:15.990 # +switch-master post6379 127.0.0.1 6379 127.0.0.1 6381 3600:X 17 Jul 2020 14:00:15.991 * +slave slave 127.0.0.1:6380 127.0.0.1 6380 @ post6379 127.0.0.1 6381 3600:X 17 Jul 2020 14:00:15.991 * +slave slave 127.0.0.1:6379 127.0.0.1 6379 @ post6379 127.0.0.1 6381 3600:X 17 Jul 2020 14:00:46.001 # +sdown slave 127.0.0.1:6379 127.0.0.1 6379 @ post6379 127.0.0.1 6381 //主库变成从库 127.0.0.1:6379> info replication # Replication role:slave master_host:127.0.0.1 master_port:6380 master_link_status:up master_last_io_seconds_ago:0 master_sync_in_progress:0 slave_repl_offset:28439 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:57f61dc1ed803e1b709eed6eaf5f10935786c059 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:28439 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:26779 repl_backlog_histlen:1661 //从库变成主库 127.0.0.1:6380> info replication # Replication role:master connected_slaves:2 slave0:ip=127.0.0.1,port=6381,state=online,offset=65784,lag=1 slave1:ip=127.0.0.1,port=6379,state=online,offset=65784,lag=1 master_replid:57f61dc1ed803e1b709eed6eaf5f10935786c059 master_replid2:e64c672ae3237d6d05788b4781de8007d6f65b5d master_repl_offset:65784 second_repl_offset:20617 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:65784

同一组sentinel可以同时监控多个master

复制的缺点：由于所有的写操作都是在master上，然后同步更新到slave上，所以从master到slave会有一定的延迟，当slave很多时，会使延迟加重

Processed: 0.008, SQL: 8

redis主从复制

目录

文章目录

Replication

一主二仆

薪火相传

反客为主

哨兵模式(sentinel)