1、主从的my.cnf的server_id=x 不一致
主库操作,创建一个用户，然后授权该用户可以再从上进行复制所有库
>create user 'rep1'@'192.168.6.128' identified by '123456'; >grant replication slave on *.* to 'rep1'@'192.168.6.128' identified by '123456'; >flush PRIVILEGES;
2、把主库的数据文件做全量备份
[root@lvs-master mysql]# bin/mysqldump -uroot --master-data=2 --single-transaction --set-gtid-purged=OFF -p123456 --databases $(/home/work/mysql/bin/mysql -uroot -p123456 -N -e "show databases"|egrep -v "performance_schema|information_schema|azkaban"|xargs echo)|gzip > all_dbs_fulldump.sql.gz
3、把主库的备份文件传到从库，解压
[work@lvs-slave mysql]$ gunzip all_dbs_fulldump.sql.gz [work@lvs-slave mysql]$ ./bin/mysql -u root -p -h localhost -S /home/work/mysql/tmp/mysql.sock < all_dbs_fulldump.sql
4、从库操作
停掉slave,配置主库的信息，启动slave，查看状态
>stop slave; >change master to master_host='192.168.6.130',master_user='rep1',master_password='123456',master_port=3306,MASTER_AUTO_POSITION=1; >start slave; >show slave statusG
如果这里出现：
Last_SQL_Error: Error 'Operation CREATE USER failed for 'rep1'@'192.168.6.128'' on query. Default database: ''. Query: 'CREATE USER 'rep1'@'192.168.6.128' IDENTIFIED WITH 'mysql_native_password' AS '*6BB4837EB74329105EE4568DDA7DC67ED2CA2AD9'
则需要删除本地的'rep1'@'192.168.6.128' 账户，然后再进行配置master的Gtid同步
>drop user 'rep1'@'192.168.6.128'; >FLUSH PRIVILEGES; >stop slave; >change master to master_host='192.168.6.130',master_user='rep1',master_password='123456',master_port=3306,MASTER_AUTO_POSITION=1; >start slave; >show slave statusG
检测
Slave_IO_Running和Slave_SQL_Running 是否都为yes
Retrieved_Gtid_Set和Executed_Gtid_Set都有值