MySQL数据库的主主同步配置

一、架构方案思路

1.两台mysql都可读写，互为主备。本文的实验中：默认只使用一台（DCGH-DB1）负责数据的写入，另一台（DCGH-DB2）备用。

2.DCGH-DB1是DCGH-DB2的主库，DCGH-DB2又是DCGH-DB1的主库，它们互为主从。

3.不足之处：DCGH-DB2可能会一直处于空闲状态（后期经过改进后，可以用它当从库，负责部分查询）。

二、基础环境

操作系统：CentOS Linux release 7.4.1708 (Core) 64Bit
MySQL版本：MySQL Community Server version: 5.7.21 64Bit
主机名及IP：DCGH-DB1（10.1.1.31），DCGH-DB2(10.1.1.32）
两台主机的操作系统，数据库版本完全一致（克隆的）。实际环境中，需要确保操作系统版本完全一致、MySQL版本完全一致、数据完全一致，可能会涉及到数据库的备份与还原。

三、环境初始化

1.在DCGH-DB2上安装MySQL服务器，进行初始化。

[root@DCGH-DB2 ~]# curl -C - -O https://repo.mysql.com//mysql57-community-release-el7-11.noarch.rpm
[root@DCGH-DB2 ~]# rpm -ivh mysql57-community-release-el7-11.noarch.rpm
[root@DCGH-DB2 ~]# yum -y install mysql-community-server
[root@DCGH-DB2 ~]# systemctl start mysqld
[root@DCGH-DB2 ~]# systemctl enable mysqld
[root@DCGH-DB2 ~]# firewall-cmd --permanent --add-port=3306/tcp
[root@DCGH-DB2 ~]# firewall-cmd --reload
[root@DCGH-DB2 ~]# grep -i password /var/log/mysqld.log 
2018-04-11T10:11:40.732848Z 1 [Note] A temporary password is generated for root@localhost: p>-/fQDca3ag
[root@DCGH-DB2 ~]# mysql -u root -p
Enter password: 
mysql> set password=password("DCGH-test-db2");
mysql> exit

如果遭遇如下错误：ERROR 1819 (HY000): Your password does not satisfy the current policy requirements，密码不满足当前策略要求，那么请执行以下命令：

set global validate_password_policy=0;  
set global validate_password_mixed_case_count=0;  
set global validate_password_number_count=3;  
set global validate_password_special_char_count=0;  
set global validate_password_length=3;  

2.关机克隆DCGH-DB2，修改配置，使之满足DCGH-DB1要求，只需修改主机名及IP即可，登录数据库，修改密码。

[root@DCGH-DB2 ~]# hostnamectl set-hostname DCGH-DB1 --static
[root@DCGH-DB2 ~]# nmtui
[root@DCGH-DB2 ~]# systemctl restart network
[root@DCGH-DB1 ~]# mysql -u root -p
Enter password: 
mysql> set password=password("DCGH-test-db1");
mysql> exit

至此，实验环境准备完毕。

四、DCGH-DB1配置

1.修改/etc/my.cnf,在[mysqld]下加入如下内容：

server-id=31
auto-increment-increment=2
auto-increment-offset=1
log-bin=mysql-bin
binlog_format=mixed
relay-log=relay-bin
log-slave-updates

2.重启服务，创建同步用户copy，并授予相应权限，锁表，查出Position备用。

[root@DCGH-DB1 ~]# mysql -u root -A -p
Enter password: 
mysql> grant replication slave,replication client on *.* to 'copy'@10.1.1.32 identified by 'DCGH-test-db2';
mysql> flush privileges;
mysql> flush tables with read lock;
mysql> show master status;
+------------------+----------+--------------+------------------+-------------------+
| File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000001 |      627 |              |                  |                   |
+------------------+----------+--------------+------------------+-------------------+

五、DCGH-DB2配置

1.修改/etc/my.cnf,在[mysqld]下加入如下内容：

server-id=32
auto-increment-increment=2
auto-increment-offset=1
log-bin=mysql-bin
binlog_format=mixed
relay-log=relay-bin
log-slave-updates

2.重启服务，创建同步用户copy，并授予相应权限并在DCGH-DB2上去同步DCGH-DB1上的数据。

[root@DCGH-DB2 ~]# mysql -u root -A -p
Enter password: 
mysql> grant replication slave,replication client on *.* to 'copy'@10.1.1.31 identified by 'DCGH-test-db1';
mysql> CHANGE MASTER TO MASTER_HOST='10.1.1.31',MASTER_USER='copy',MASTER_PASSWORD='DCGH-test-db2',MASTER_LOG_FILE='mysql-bin.000001',MASTER_LOG_POS=627;
mysql> start slave;
mysql> show slave status \G
*************************** 1. row ***************************
               Slave_IO_State: 
                  Master_Host: 10.1.1.31
                  Master_User: copy
                  Master_Port: 3306
                Connect_Retry: 60
              Master_Log_File: mysql-bin.000001
          Read_Master_Log_Pos: 627
               Relay_Log_File: relay-bin.000001
                Relay_Log_Pos: 4
        Relay_Master_Log_File: mysql-bin.000001
             Slave_IO_Running: No
            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: 627
              Relay_Log_Space: 154
              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: NULL
Master_SSL_Verify_Server_Cert: No
                Last_IO_Errno: 1593
                Last_IO_Error: Fatal error: The slave I/O thread stops because master and slave have equal MySQL server UUIDs; these UUIDs must be different for replication to work.
               Last_SQL_Errno: 0
               Last_SQL_Error: 
  Replicate_Ignore_Server_Ids: 
             Master_Server_Id: 31
                  Master_UUID: 
             Master_Info_File: /var/lib/mysql/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: 180411 19:27:07
     Last_SQL_Error_Timestamp: 
               Master_SSL_Crl: 
           Master_SSL_Crlpath: 
           Retrieved_Gtid_Set: 
            Executed_Gtid_Set: 
                Auto_Position: 0
         Replicate_Rewrite_DB: 
                 Channel_Name: 
           Master_TLS_Version: 

致命报错来袭：Fatal error: The slave I/O thread stops because master and slave have equal MySQL server UUIDs; these UUIDs must be different。原来是UUID重复了,这是因为咱们的MySQL服务器是克隆出来的，好办，改server-uuid(随便改个数字或字母跟之前uuid不一致即可)！

[root@DCGH-DB2 ~]# vi /var/lib/mysql/auto.cnf
server-uuid=ba5f1c18-3d70-11e8-891f-000c2986a1f0
[root@DCGH-DB2 ~]# systemctl restart mysqld

重复以上步骤：

mysql> CHANGE MASTER TO MASTER_HOST='10.1.1.31',MASTER_USER='copy',MASTER_PASSWORD='DCGH-test-db2',MASTER_LOG_FILE='mysql-bin.000001',MASTER_LOG_POS=627;
mysql> start slave;
mysql> show slave status \G
*************************** 1. row ***************************
               Slave_IO_State: Waiting for master to send event
                  Master_Host: 10.1.1.31
                  Master_User: copy
                  Master_Port: 3306
                Connect_Retry: 60
              Master_Log_File: mysql-bin.000001
          Read_Master_Log_Pos: 627
               Relay_Log_File: relay-bin.000002
                Relay_Log_Pos: 320
        Relay_Master_Log_File: mysql-bin.000001
             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: 627
              Relay_Log_Space: 521
              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: 31
                  Master_UUID: ba5f1c18-3d70-11e8-891f-000c2986a1f9
             Master_Info_File: /var/lib/mysql/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: 
                Auto_Position: 0
         Replicate_Rewrite_DB: 
                 Channel_Name: 
           Master_TLS_Version: 

Slave_IO_Running: Yes，Slave_SQL_Running: Yes，Seconds_Behind_Master: 0，表明目前已经同步了。

六、验证

1.回到DCGH-DB1，解锁表创建数据库DCGHDB。

mysql> unlock tables;
mysql> create database DCGHDB;
mysql> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| DCGHDB             |
| mysql              |
| performance_schema |
| sys                |
+--------------------+

我在DCGHDB2使用show databases命令，结果一致。
2.在DCGH-DB1创建表并插入数据，看是否同步。

mysql> use DCGHDB;
Database changed
mysql> create table test(username varchar(30),password varchar(30));
mysql> insert into test(username,password) values('IVAN DU','DCGH-test');
mysql> select * from test;
+----------+-----------+
| username | password  |
+----------+-----------+
| IVAN DU  | DCGH-test |
+----------+-----------+

在DCGH-DB2查表test的数据是否一致。

mysql> use DCGHDB;
Database changed
mysql> select * from test;
+----------+-----------+
| username | password  |
+----------+-----------+
| IVAN DU  | DCGH-test |
+----------+-----------+

3.在DCGH-DB1更新表test。

mysql> update test set username='DCGH' where username='IVAN DU';
mysql> select * from test;
+----------+-----------+
| username | password  |
+----------+-----------+
| DCGH     | DCGH-test |
+----------+-----------+

在DCGH-DB2进行查看,结果一致。

mysql> select * from test;
+----------+-----------+
| username | password  |
+----------+-----------+
| DCGH     | DCGH-test |
+----------+-----------+

4.经验证，删除表数据、表结构、库都能够顺利同步。不展示，相关验证命令如下：

mysql> delete from test;
mysql> drop table test;
mysql> drop database DCGHDB;
mysql> show tables;

七、拓展知识

配置文件部分选项解读及拓展。

# 服务器的ID,必须唯一，一般设置自己的IP
server-id=31
# 复制过滤：不需要备份的数据库（MySQL库一般不同步）
binlog-ignore-db=mysql
# 开启二进制日志功能，名字可以随便取，最好有含义（比如项目名）
log-bin=DCGH
# 为每个 session 分配的内存,在事务过程中用来存储二进制日志的缓存
binlog_cache_size=1M
# 主从复制的格式(mixed,statement,row,默认格式是 statement)
binlog_format=mixed
# 二进制日志自动删除/过期的天数。默认值为 0,表示不自动删除。
expire_logs_days=7
## 跳过主从复制中遇到的所有错误或指定类型的错误,避免 slave 端复制中断。 
## 如:1062 错误是指一些主键重复,1032 错误是因为主从数据库数据不一致
slave_skip_errors=1062
# 作为从服务器时的中继日志
relay_log=edu-mysql-relay-bin
# log_slave_updates 表示 slave 将复制事件写进自己的二进制日志
log_slave_updates=1
# 主键自增规则，避免主从同步ID重复的问题
auto_increment_increment=2  # 自增因子（每次加2）
auto_increment_offset=1     # 自增偏移（从1开始），单数

八、参考资料

https://dev.mysql.com/doc/refman/5.6/en/replication-howto-newservers.html
https://dev.mysql.com/doc/refman/5.7/en/stored-programs-logging.html
https://dev.mysql.com/doc/refman/5.7/en/replication.html