gtid原理:
1.当一个事务在主库端执行并提交时,产生GTID,一同记录到binlog日志中;
2.binlog传输到slave,并存储到slave的relaylog后,读取这个GTID的这个值设置gtid_next变量,即告诉Slave,下一个要执行的GTID值;
3.sql线程从relay log中获取GTID,然后对比slave端的binlog是否有该GTID;
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就会执行该GTID事务,并记录该GTID到自身的binlog,在读取执行事务前会先检查其他session持有该GTID,确保不被重复执行;
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
相比传统主从优点:
更简单的搭建主从复制;
比传统的复制更加安全;
GTID是连续的没有空洞的,保证数据的一致性,零丢失。
数据库 Gtid 主从复制如下, 如需实现多从, 将以下实验从案例再配置一次即可
操作系统 | MySQL版本 | IP | 角色 |
---|---|---|---|
Centos7.4 | MysQL5.7 | 192.168.230.16 | master |
Centos7.4 | MysQL5.7 | 192.168.230.15 | slave |
主库配置
修改配置文件
[root@master ~]# vim /etc/my.cnf
log-bin=mysql-bin
server-id=1
gtid-mode = ON
enforce_gtid_consistency = 1
重启mysql
[root@master ~]# service mysqld restart
准备数据库
mysql> create database test;
mysql> use test;
mysql> create table jaja(id int,name varchar(20));
mysql> insert into jaja values (1,'aa'),(2,'bb');
mysql> select * from jaja;
+------+------+
| id | name |
+------+------+
| 1 | aa |
| 2 | bb |
+------+------+
授权允许能够远程连接的主机
mysql> grant replication slave on *.* to 'rep'@'192.168.230.15' identified by 'rep123.com';
mysql> flush privileges;
导出当前数据库
[root@master ~]# mysqldump -uroot -pcs123.com --all-databases --single-transaction --master-data=1 --flush-logs > $(date +%F-%H-%M)-all.sql
[root@master ~]# ls
2019-02-28-19-21-all.sql
将备份文件发送给slave
[root@master ~]# scp 2019-02-28-19-21-all.sql root@192.168.230.15:/root/
从库配置
注意,如果之前在从库上面部署过传统主从,请先初始化数据库环境
检查是否能使用远程账户登录
[root@slave ~]# mysql -urep -p'rep123.com' -h 192.168.230.16
修改配置文件
[root@slave ~]# vim /etc/my.cnf
server-id = 2
gtid-mode = ON
enforce_gtid_consistency = 1
重启服务
[root@slave ~]# service mysqld restart
导入数据
[root@slave ~]# mysql -uroot -pcs123.com -e "source 2019-02-28-19-21-all.sql"
mysql> use test;
mysql> select * from jaja;
+------+------+
| id | name |
+------+------+
| 1 | aa |
| 2 | bb |
+------+------+
指向master,gtid自动协商同步
mysql> change master to master_host='192.168.230.16',master_user='rep',master_password='rep123.com',master_auto_position=1;
启动slave角色
mysql> start slave;
查看角色是否同步
mysql> show slave status\G
*************************** 1. row ***************************
Slave_IO_State: Waiting for master to send event
Master_Host: 192.168.230.16
Master_User: rep
Master_Port: 3306
Connect_Retry: 60
Master_Log_File: mysql-bin.000008
Read_Master_Log_Pos: 194
Relay_Log_File: slave-relay-bin.000002
Relay_Log_Pos: 367
Relay_Master_Log_File: mysql-bin.000008
Slave_IO_Running: Yes
Slave_SQL_Running: Yes
Replicate_Do_DB:
Replicate_Ignore_DB:
Replicate_Do_Table:
Replicate_Ignore_Table:
Replicate_Wild_Do_Table:
Replicate_Wild_Ignore_Table:
Last_Errno: 0
Last_Error:
Skip_Counter: 0
Exec_Master_Log_Pos: 194
Relay_Log_Space: 574
Until_Condition: None
Until_Log_File:
Until_Log_Pos: 0
Master_SSL_Allowed: No
Master_SSL_CA_File:
Master_SSL_CA_Path:
Master_SSL_Cert:
Master_SSL_Cipher:
Master_SSL_Key:
Seconds_Behind_Master: 0
Master_SSL_Verify_Server_Cert: No
Last_IO_Errno: 0
Last_IO_Error:
Last_SQL_Errno: 0
Last_SQL_Error:
Replicate_Ignore_Server_Ids:
Master_Server_Id: 1
Master_UUID: b27045f6-36bb-11e9-b404-000c29c87115
Master_Info_File: /opt/data/master.info
SQL_Delay: 0
SQL_Remaining_Delay: NULL
Slave_SQL_Running_State: Slave has read all relay log; waiting for more updates
Master_Retry_Count: 86400
Master_Bind:
Last_IO_Error_Timestamp:
Last_SQL_Error_Timestamp:
Master_SSL_Crl:
Master_SSL_Crlpath:
Retrieved_Gtid_Set:
Executed_Gtid_Set: b27045f6-36bb-11e9-b404-000c29c87115:1-7539
Auto_Position: 1
Replicate_Rewrite_DB:
Channel_Name:
Master_TLS_Version:
主要看这个位置,两个yes就对了
Slave_IO_Running: Yes
Slave_SQL_Running: Yes
验证
主库添加数据
mysql> use test;
mysql> insert into jaja values (3,'aaa'),(4,'bbb');
mysql> select * from jaja;
+------+------+
| id | name |
+------+------+
| 1 | aa |
| 2 | bb |
| 3 | aaa |
| 4 | bbb |
+------+------+
从库查看
mysql> select * from jaja;
+------+------+
| id | name |
+------+------+
| 1 | aa |
| 2 | bb |
| 3 | aaa |
| 4 | bbb |
+------+------+
4 rows in set (0.01 sec)