MySQL架构备份之M-S-S级联备份
M--S1--S2 级联复制
master—>slave1—>slave2
master需要开启二进制日志
中间的slave1也需要打开二进制日志,但是它默认不把应用master的操作记录到自己的二进制日志。所以需要打开一个参数让它记录,才可以传给第三级的从;然后在中间从和第三级从之间再做一次AB复制就可以了。
打开log-slave-updates=1,让第一台传过来relay日志记录到自己的二进制日志。
思路:
先搭建好主从——>然后加入slave2
环境准备及要求:
1、关闭防火墙和selinux
2、hosts文件中三台服务器主机名和ip地址一一对应起来
3、系统时间需要同步
4、master和slave的数据库版本保持一致(系统版本保持一致)
思路:
先搭建好主从——>然后加入slave2
1、master和中间slave1必须开启二进制日志
2、slave2必须开启中继日志
3、master和slave1和slave2的server-id必须不一样 范围(2^23-1)
4、master和slave的初始数据一致
M--S1--S2级联架构操作步骤
环境:
master:192.168.1.166
slave1:192.168.1.114
slave2:192.168.1.180
1、环境准备
# master: [root@master ~]# vim /etc/hosts 192.168.1.114 slave1 # slave1: [root@slave1 ~]# vim /etc/hosts 192.168.1.166 master 192.168.1.180 slave2 # slave2: [root@slave2 ~]# vim /etc/hosts 192.168.1.114 slave1
2、修改配置文件
# master: root@master ~]# vim /etc/my.cnf log-bin=mysql-bin server-id = 1 # slave1: [root@slave1 ~]# vim /etc/my.cnf log-bin=mysql-bin log-slave-updates=1 server-id = 2 relay-log=/data/DB/relay.log # slave2: [root@slave2 ~]# vim /etc/my.cnf server-id = 3 relay-log=/data/DB/relay.log
3、数据初始化一致
情况1:master和slave都没有业务数据,直接初始化master和slave的数据库
情况2:master上面有业务数据
masteer: 1>停止数据库 [root@master ~]# /etc/init.d/mysqld stop 2>物理级别备份 [root@master ~]# cd /data/DB/ && ll [root@master DB]# tar -cvzf /tmp/mysql.tar.gz * [root@master DB]# scp /tmp/mysql.tar.gz 192.168.1.114:/tmp [root@master DB]# scp /tmp/mysql.tar.gz 192.168.1.180:/tmp 3>slave1和slave2进行数据恢复 [root@slave1 ~]# cd /data/DB/ && ll [root@slave1 DB]# tar -xf /tmp/mysql.tar.gz -C . [root@slave1 DB]# /etc/init.d/mysqld start [root@slave2 ~]# cd /data/DB/ && ll [root@slave2 DB]# tar -xf /tmp/mysql.tar.gz -C . [root@slave2 DB]# /etc/init.d/mysqld start 4>查看slave和master上面的数据是否一样 [root@master DB]# echo "show databases;" | mysql -p123 [root@slave1 DB]# echo "show databases;" | mysql -p123 [root@slave2 DB]# echo "show databases;" | mysql -p123
4、在master上面创建对slave1授权用户及查看binlog信息
1>创建授权用户 mysql> grant replication slave on *.* to 'slave'@'192.168.1.114' identified by '123'; 2>查看binglog信息,只有打开二进制日志,这句命令才有结果,表示当前数据库的二进制日志写到什么位置 mysql> show master status \G *************************** 1. row *************************** File: mysql-bin.000013 Position: 331 Binlog_Do_DB: Binlog_Ignore_DB: Executed_Gtid_Set: 1 row in set (0.03 sec)
5、在slave1上配置同步信息
slave1: mysql> change master to master_host='192.168.1.166', master_user='slave', master_password='123', master_port=3306, master_log_file='mysql-bin.000013', master_log_pos=331; 参数说明: master_host: master ip master_user: 同步用户 master_password: 密码 master_port: 端口 master_log_file: master上面查看到的二进制日志名 master_log_pos: master上面查看到的POS值 mysql> start slave; mysql> show slave status \G; Slave_IO_Running: Yes Slave_SQL_Running: Yes 如果查看状态这两个值都为Yes, 说明M--S就已经做好
6、在slave1上面对slave2授权
先停止slave slave1: mysql> stop slave; Query OK, 0 rows affected (0.03 sec) mysql> grant replication slave on *.* to 'slave'@'192.168.1.180' identified by '123'; mysql> show master status \G *************************** 1. row *************************** File: mysql-bin.000017 Position: 425 Binlog_Do_DB: Binlog_Ignore_DB: Executed_Gtid_Set: 1 row in set (0.00 sec)
7、在slave2上配置同步信息
slave2: mysql> change master to master_host='192.168.1.114', master_user='slave', master_password='123', master_port=3306, master_log_file='mysql-bin.000017', master_log_pos=425; mysql> start slave; mysql> show slave status \G; Slave_IO_Running: Yes Slave_SQL_Running: Yes
8、在回到slave1上面启动slave
slave1: mysql> start slave;
9、测试验证
先查看原有的三个服务器的db01库中的table03表, master: mysql> select * from db01.table03; +------+------+ | id | name | +------+------+ | 1 | haha | | 2 | wowo | | 4 | yoyo | | 1 | haha | | 2 | wowo | | 3 | lili | | 4 | yoyo | +------+------+ 7 rows in set (0.10 sec) slave1: mysql> select * from db01.table03; +------+------+ | id | name | +------+------+ | 1 | haha | | 2 | wowo | | 4 | yoyo | | 1 | haha | | 2 | wowo | | 3 | lili | | 4 | yoyo | +------+------+ 7 rows in set (0.17 sec) slave2: mysql> select * from db01.table03; +------+------+ | id | name | +------+------+ | 1 | haha | | 2 | wowo | | 4 | yoyo | | 1 | haha | | 2 | wowo | | 3 | lili | | 4 | yoyo | +------+------+ 7 rows in set (0.44 sec) 在master上面删除一条数据 mysql> delete from db01.table03 where name="lili"; Query OK, 1 row affected (0.38 sec) 再次查看 master: mysql> select * from db01.table03; +------+------+ | id | name | +------+------+ | 1 | haha | | 2 | wowo | | 4 | yoyo | | 1 | haha | | 2 | wowo | | 4 | yoyo | +------+------+ 6 rows in set (0.00 sec) slave1: mysql> select * from db01.table03; +------+------+ | id | name | +------+------+ | 1 | haha | | 2 | wowo | | 4 | yoyo | | 1 | haha | | 2 | wowo | | 4 | yoyo | +------+------+ 6 rows in set (0.00 sec) slave2: mysql> select * from db01.table03; +------+------+ | id | name | +------+------+ | 1 | haha | | 2 | wowo | | 4 | yoyo | | 1 | haha | | 2 | wowo | | 4 | yoyo | +------+------+ 6 rows in set (0.00 sec)
总结:
测试验证在master上面操作,会同步到slave1上面,相对slave2来说,也等于在slave1上面进行了相同的操作,所有slave2上面也会发生相应的改变