Redis主从复制集群的实现

主从复制的特点

一个master可以有多个slave
一个slave只能有一个master
数据流向是从master到slave单向的

1 主从复制的实现

Redis Slave 也要开启持久化并设置和master同样的连接密码，因为后期slave会有提升为master的可能，Slave端切换master同步后会丢失之前的所有数据，而通过持久化可以恢复数据

一旦某个Slave成为一个master的slave，Redis Slave服务会清空当前redis服务器上的所有数据并将master的数据导入到自己的内存，但是如果只是断开同步关系后，则不会删除当前已经同步过的数据。

当配置Redis复制功能时，强烈建议打开主服务器的持久化功能。否则的话，由于延迟等问题，部署的主节点Redis服务应该要避免自动启动。

1.1 启用主从同步

默认redis 状态为master，需要转换为slave角色并指向master服务器的IP+PORT+Password
在从节点执行 REPLICAOF MASTER_IP PORT 指令可以启用主从同步复制功能，早期版本使用SLAVEOF指令

127.0.0.1:6379> REPLICAOF MASTER_IP PORT
127.0.0.1:6379> CONFIG SET masterauth <masterpass>

环境准备

三台机器：
10.0.0.7 Redis-6.2.4 master
10.0.0.17 Redis-6.2.4 slave 1
10.0.0.27 Redis-6.2.4 slave 2

1.2 master

#在mater上设置key1
[root@centos7 ~]#redis-cli
127.0.0.1:6379> AUTH 123456
OK
127.0.0.1:6379> INFO replication
# Replication
role:master
connected_slaves:0
master_failover_state:no-failover
master_replid:69647aaf4b664798db3b7bde649674fad6e320af
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

127.0.0.1:6379> SET key1 v1-master
OK
127.0.0.1:6379> KEYS *
1) "key1"
127.0.0.1:6379> GET key1
"v1-master"

#在master上可以看到所有slave信息
127.0.0.1:6379> INFO replication
# Replication
role:master
connected_slaves:2
slave0:ip=10.0.0.27,port=6379,state=online,offset=6529,lag=0  #slave信息
slave1:ip=10.0.0.17,port=6379,state=online,offset=6529,lag=1
master_failover_state:no-failover
master_replid:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:6529
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:6529

#查看连接状态
[root@centos7 ~]#ss -nt
State      Recv-Q Send-Q              Local Address:Port                             Peer Address:Port
ESTAB      0      0                        10.0.0.7:6379                                10.0.0.17:45026
ESTAB      0      52                       10.0.0.7:22                                   10.0.0.1:51959
ESTAB      0      0                        10.0.0.7:6379                                10.0.0.27:44892

#查看Master日志
[root@centos7 ~]#tail -f /apps/redis/log/redis-6379.log
1293:M 04 May 2022 11:25:37.272 * Replica 10.0.0.27:6379 asks for synchronization
1293:M 04 May 2022 11:25:37.273 * Partial resynchronization request from 10.0.0.27:6379 accepted. Sending 3738 bytes of backlog starting from offset 1490.
1293:M 04 May 2022 11:27:35.089 * 1 changes in 3600 seconds. Saving...
1293:M 04 May 2022 11:27:35.090 * Background saving started by pid 2488
2488:C 04 May 2022 11:27:35.141 * DB saved on disk
2488:C 04 May 2022 11:27:35.142 * RDB: 0 MB of memory used by copy-on-write
1293:M 04 May 2022 11:27:35.193 * Background saving terminated with success
1293:M 04 May 2022 11:29:16.729 # Connection with replica 10.0.0.17:6379 lost.
1293:M 04 May 2022 11:29:16.841 * Replica 10.0.0.17:6379 asks for synchronization
1293:M 04 May 2022 11:29:16.841 * Partial resynchronization request from 10.0.0.17:6379 accepted. Sending 4032 bytes of backlog starting from offset 1490.

1.3 slave 1

[root@centos7 ~]#redis-cli
127.0.0.1:6379> AUTH 123456
OK
127.0.0.1:6379> INFO replication  #查看当前角色默认为master
# Replication
role:master
connected_slaves:0
master_failover_state:no-failover
master_replid:2c539210cad9f2660ca76c7ca35a4c41a2aef84e
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

#在slave上设置master的IP和端口
127.0.0.1:6379> REPLICAOF 10.0.0.7 6379
OK
#在slave上设置master的密码，才可以同步
127.0.0.1:6379> CONFIG SET masterauth 123456
OK
127.0.0.1:6379> INFO replication
# Replication
role:slave  #角色变为slave
master_host:10.0.0.7  #指向master
master_port:6379
master_link_status:up
master_last_io_seconds_ago:5
master_sync_in_progress:0
slave_repl_offset:98
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:98
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:98

#在slave1上查看已经同步成功
127.0.0.1:6379> KEYS *
1) "key1"
127.0.0.1:6379> GET key1
"v1-master"

#修改slave节点配置文件
[root@centos7 ~]#vim /apps/redis/etc/redis.conf
# replicaof <masterip> <masterport>
replicaof 10.0.0.7 6379  #指定master的IP和端口号

# masterauth <master-password>
masterauth 123456  #指定master的密码

#重启redis服务
[root@centos7 ~]#systemctl restart redis.service

#再次登录查看连接状态
[root@centos7 ~]#redis-cli
127.0.0.1:6379> AUTH 123456
OK
127.0.0.1:6379> INFO replication
# Replication
role:slave
master_host:10.0.0.7
master_port:6379
master_link_status:up
master_last_io_seconds_ago:11
master_sync_in_progress:0
slave_repl_offset:5745
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:5745
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1490
repl_backlog_histlen:4256

#查看slave日志
[root@centos7 ~]#tail -f /apps/redis/log/redis-6379.log
2517:S 04 May 2022 11:29:16.828 * DB loaded from disk: 0.001 seconds
2517:S 04 May 2022 11:29:16.829 * Before turning into a replica, using my own master parameters to synthesize a cached master: I may be able to synchronize with the new master with just a partial transfer.
2517:S 04 May 2022 11:29:16.829 * Ready to accept connections
2517:S 04 May 2022 11:29:16.829 * Connecting to MASTER 10.0.0.7:6379
2517:S 04 May 2022 11:29:16.833 * MASTER <-> REPLICA sync started
2517:S 04 May 2022 11:29:16.834 * Non blocking connect for SYNC fired the event.
2517:S 04 May 2022 11:29:16.836 * Master replied to PING, replication can continue...
2517:S 04 May 2022 11:29:16.838 * Trying a partial resynchronization (request 3ecc9b36d6175a927c0c46bc5681ca340b4f49d0:1490).
2517:S 04 May 2022 11:29:16.840 * Successful partial resynchronization with master.
2517:S 04 May 2022 11:29:16.840 * MASTER <-> REPLICA sync: Master accepted a Partial Resynchronization.

#slave状态只读无法写入数据
127.0.0.1:6379> set key1 v1-slave
(error) READONLY You can't write against a read only replica.

#取消主从复制，在slave上执行 REPLIATOF NO ONE 指令可以取消主从复制
#取消主从复制后，会断开和master的连接不再主从复制, 但不会清除slave上已有的数据
127.0.0.1:6379> REPLICAOF no one
OK
127.0.0.1:6379> INFO replication
# Replication
role:master
connected_slaves:0
master_failover_state:no-failover
master_replid:4c9236e7046236ebe873d01190864029e9cf508a
master_replid2:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_repl_offset:9105
second_repl_offset:9106
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1490
repl_backlog_histlen:7616

127.0.0.1:6379> KEYS *
1) "key1"
127.0.0.1:6379> GET key1
"v1-master"

1.4 slave 2

[root@centos7 ~]#redis-cli
127.0.0.1:6379> AUTH 123456
OK
127.0.0.1:6379> INFO replication
# Replication
role:master
connected_slaves:0
master_failover_state:no-failover
master_replid:5c563b18d77232c8b1100322003c015d3c44131c
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

#在slave上设置master的IP和端口
127.0.0.1:6379> REPLICAOF 10.0.0.7 6379
OK
#在slave上设置master的密码，才可以同步
127.0.0.1:6379> CONFIG SET masterauth 123456
OK
127.0.0.1:6379> INFO replication
# Replication
role:slave
master_host:10.0.0.7
master_port:6379
master_link_status:up
master_last_io_seconds_ago:4
master_sync_in_progress:0
slave_repl_offset:644
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:644
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:603
repl_backlog_histlen:42

#在slave2上查看已经同步成功
127.0.0.1:6379> KEYS *
1) "key1"
127.0.0.1:6379> GET key1
"v1-master"

#修改slave节点配置文件
[root@centos7 ~]#vim /apps/redis/etc/redis.conf
# replicaof <masterip> <masterport>
replicaof 10.0.0.7 6379

# masterauth <master-password>
masterauth 123456

#重启redis服务
[root@centos7 ~]#systemctl restart redis.service

#再次登录查看连接状态
[root@centos7 ~]#redis-cli
127.0.0.1:6379> AUTH 123456
OK
127.0.0.1:6379> INFO replication
# Replication
role:slave
master_host:10.0.0.7
master_port:6379
master_link_status:up
master_last_io_seconds_ago:5
master_sync_in_progress:0
slave_repl_offset:5353
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:5353
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1490
repl_backlog_histlen:3864

#slave状态只读无法写入数据
127.0.0.1:6379> set key1 v1-slave
(error) READONLY You can't write against a read only replica.

1.5 主从复制故障恢复实现

master故障后，只能手动提升一个slave为新master，不支持自动切换。

之后将其它的slave节点重新指定新的master为master节点

Master的切换会导致master_replid发生变化，slave之前的master_replid就和当前master不一致，从而会引发所有 slave的全量同步。

假设当前主节点10.0.0.7故障，提升10.0.0.17为新的master

1.51 master

[root@centos7 ~]#systemctl stop redis.service

1.52 slave 1

127.0.0.1:6379> INFO replication
# Replication
role:slave
master_host:10.0.0.7
master_port:6379
master_link_status:down  #显示down，表示无法连接master
master_last_io_seconds_ago:-1
master_sync_in_progress:0
slave_repl_offset:11233
master_link_down_since_seconds:241
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:11233
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1490
repl_backlog_histlen:9744

#停止slave同步并提升为新的master
127.0.0.1:6379> REPLICAOF NO ONE
OK
127.0.0.1:6379> INFO replication
# Replication
role:master
connected_slaves:0
master_failover_state:no-failover
master_replid:df2f9e8391795cc0a3b7023b9a90a0b9db553ad3
master_replid2:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_repl_offset:11233
second_repl_offset:11234
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1490
repl_backlog_histlen:9744

#测试能否写入数据
127.0.0.1:6379> set date 20220504
OK
127.0.0.1:6379> get date
"20220504"
127.0.0.1:6379> dbsize
(integer) 2

1.53 slave 2

127.0.0.1:6379> INFO replication
# Replication
role:slave
master_host:10.0.0.7
master_port:6379
master_link_status:down
master_last_io_seconds_ago:-1
master_sync_in_progress:0
slave_repl_offset:11233
master_link_down_since_seconds:332
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:11233
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1490
repl_backlog_histlen:9744

#修改10.0.0.27节点指向新的master节点10.0.0.17
127.0.0.1:6379> REPLICAOF 10.0.0.17 6379
OK
127.0.0.1:6379> INFO replication
# Replication
role:slave
master_host:10.0.0.17
master_port:6379
master_link_status:up
master_last_io_seconds_ago:4
master_sync_in_progress:0
slave_repl_offset:11391
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:df2f9e8391795cc0a3b7023b9a90a0b9db553ad3
master_replid2:3ecc9b36d6175a927c0c46bc5681ca340b4f49d0
master_repl_offset:11391
second_repl_offset:11234
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1490
repl_backlog_histlen:9902

#查看数据同步
127.0.0.1:6379> get date
"20220504"
127.0.0.1:6379> dbsize
(integer) 2