GTID的工作原理：
1、master更新数据时，会在事务前产生GTID，一同记录到binlog日志中。
2、slave端的i/o 线程将变更的binlog，写入到本地的relay log中。
3、sql线程从relay log中获取GTID，然后对比slave端的binlog是否有记录。
4、如果有记录，说明该GTID的事务已经执行，slave会忽略。
5、如果没有记录，slave就会从relay log中执行该GTID的事务，并记录到binlog。
6、在解析过程中会判断是否有主键，如果没有就用二级索引，如果没有就用全部扫描。
GTIDS的环境：
一台master 192.168.200.111
多个slave： 192.168.200.112 192.168.200.113
修改master服务器：
[root@localhost ~]# /etc/init.d/mysql stop
[root@localhost ~]# vim /etc/my.cnf
[mysqld]
datadir=/usr/local/mysql/data
socket=/tmp/mysql.sock
symbolic-links=0
#GTID:
gtid_mode=on
enforce_gtid_consistency=on
server_id=1
#binlog
log-bin=mysql-binlog
log-slave-updates=1
binlog_format=row
#relay log
skip_slave_start=1
[mysqld_safe]
log-error=/usr/local/mysql/data/mysql.log
pid-file=/usr/local/mysql/data/mysql.pid
!includedir /etc/my.cnf.d
[root@localhost ~]# /etc/init.d/mysql start
[root@localhost ~]# mysql -uroot -p111111
mysql> grant replication slave on *.* to 'myslave'@'192.168.200.%' identified by '123456';
可能会出错，原因是自定义密码比较简单，就出现了不符合密码策略的问题。
解决方法请查看https://www.cnblogs.com/lyqlyqlyq/p/11683294.html
mysql> flush privileges;
修改slave1服务器
[root@localhost ~]# /etc/init.d/mysql stop
[root@localhost ~]# vim /etc/my.cnf
[mysqld]
datadir=/usr/local/mysql/data
socket=/tmp/mysql.sock
symbolic-links=0
gtid_mode=on
enforce_gtid_consistency=on
server_id=2
read_only=on
log_bin=on
#保证从节点的安全性
master_info_repository=TABLE
relay_log_info_repository=TABLE
[mysqld_safe]
log-error=/usr/local/mysql/data/mysql.log
pid-file=/usr/local/mysql/data/mysql.pid
!includedir /etc/my.cnf.d
[root@localhost ~]# /etc/init.d/mysql start
[root@localhost ~]# mysql -uroot -p111111
mysql> stop slave;
mysql> change master to master_host='192.168.200.111',
-> master_user='myslave',
-> master_password='123456',
-> master_auto_position=1;
mysql> start slave;
修改slave2服务器
[root@localhost ~]# /etc/init.d/mysql stop
[root@localhost ~]# vim /etc/my.cnf
[mysqld]
datadir=/usr/local/mysql/data
socket=/tmp/mysql.sock
symbolic-links=0
gtid_mode=on
enforce_gtid_consistency=on
server_id=3
read_only=on
log_bin=on
#保证从节点的安全性
master_info_repository=TABLE
relay_log_info_repository=TABLE
[mysqld_safe]
log-error=/usr/local/mysql/data/mysql.log
pid-file=/usr/local/mysql/data/mysql.pid
!includedir /etc/my.cnf.d
[root@localhost ~]# /etc/init.d/mysql start
[root@localhost ~]# mysql -uroot -p111111
mysql> stop slave;
mysql> change master to master_host='192.168.200.111',
-> master_user='myslave',
-> master_password='123456',
-> master_auto_position=1;
mysql> start slave;
测试
在master服务器上创建数据
mysql> create database test;
Query OK, 1 row affected (0.01 sec)
然后在slave服务器上查看
mysql> show databases;