My mysql master-slave replication configuration method

server-id I used the last number of the IP. The maximum log size is 512MB, and records older than 30 days are deleted, which can be adjusted freely according to the actual environment.

first step:

Main mysql my.cnf configuration:

wrote

server-id = 228
log-bin=mysql-bin
expire_logs_days = 30

Connect to the main database:

mysql -uroot -pxxxxxxxx

GRANT REPLICATION SLAVE ON *.* to 'sync_xxx'@'%' identified by 'xxxxxxxx';

flush privileges;

View permissions: select * from user where user = 'sync_xxx';

You can see that the value of the Repl_slave_priv key pair is Y.

Step 2:

Configure my.cnf from database and restart:

wrote

server-id = 220
log-bin=mysql-bin
relay_log = mysql-relay-bin
log_slave_updates = 1
read_only = 1
replicate-wild-ignore-table =mysql.%
replicate-wild-ignore-table =test.%
replicate-wild-ignore-table =log.%
replicate-wild-ignore-table =information_schema.%
replicate-wild-ignore-table =performance_schema.%

The third step is also the most important and most important step. To ensure that the data of the two databases are completely consistent, you must stop the web service of the main database, then dump the data, then start the main database, and check the log status, because this backup data is imported From the library, the log status is configured to the slave library, which can ensure that the slave library data is completely consistent with the master library.

At this point, it must be ensured that there are no read and write operations in the two databases (disconnecting services such as the web that write to the database).

Dump the main database data (if the data is too large to be dumped, you can stop mysql and copy the entire data directory.), and then check the log status:

1. Connect to the database:

Check the log status: show master status; see the following information.

mysql> show master status;

+------------------+----------+--------------+------------------+

+------------------+----------+--------------+------------------+

| mysql-bin.000003 | 1230 | | |

+------------------+----------+--------------+------------------+

1 row in set (0.00 sec)

有了备份文件，有了备份文件以后没有进行任何写操作的主数据库的master状态，那么主数据库所在服务器的web服务就可以恢复了，因为从库导入了相同数据，且可以配置这个日志起点。

主库导出的数据导入到从数据库。

连接到从数据库，执行：注意：mysql-bin.000002来自查看主库状态得到的文件名字段File，MASTER_LOG_POS是Position字段。

 
 CHANGE MASTER TO 

 
 MASTER_HOST= 
 '192.168.1.228' 
 , 

 
 MASTER_USER= 
 'xxxxx' 
 , 

 
 MASTER_PASSWORD= 
 'xxxxxx.' 
 , 

 
 MASTER_PORT= 
 3306 
 , 

 
 MASTER_LOG_FILE= 
 'mysql-bin.000003' 
 , 

 
 MASTER_LOG_POS=1230 
 , 

 
 MASTER_CONNECT_RETRY= 
 10 
 ; 

  执行：
  

  start slave;
  

  查看从数据库状态：show slave status\G;
  

  配置中已经设置从数据库为只读，防止从数据库写入任何数据，检查一下：
  

  mysql> show global variables like 'read_only'; +---------------+-------+ | Variable_name | Value | +---------------+-------+ | read_only | ON | +---------------+-------+ 1 row in set (0.01 sec)
  

 
 这里要注意，虽然设置了只读，但root用户仍然有权限写入数据，所以所有写数据库的服务都不可以用root用户来连接，另外，如果创建用户的时候，是如下方法，也会导致创建了拥有supper权限的root类用户： 

grant all privileges on *.* to [email protected] identified by 'password';
flush privileges;
相当于创建了另一个超级用户！！！
所以，创建用户只能用如下方法：
grant all privileges on db_name.* to [email protected] identified by 'password';
flush privileges;
验证设置只读成功的方法：
连接mysql数据库
更新一条数据看看能不能成功。
mysql> update account set nick='xxx';
ERROR 1290 (HY000): The MySQL server is running with the --read-only option so it cannot execute this statement

主从配置完毕，去主库变更一个数据，然后从库执行sql查询看看有没有随之变化吧，如果变化了，就启动主从服务器上的web服务即可。

主从配置的作用主要是做到双机热备，但对外提供服务的只有主数据库，当然，从数据库可用于数据读取，但绝对不能有写入操作，否则从数据库数据是不会同步到主数据库中的，而且对于采用自动递增id的情况，会导致从数据库因为添加了数据，而无法从主数据库从不过来（id重复的错误）。

如果主从数据库都要对外提供服务，则需要相互配置主从数据库，可以多台主机互为主从，但需要设置id的递增方法，见文章：

http://www.cnblogs.com/top5/archive/2009/04/15/1436399.html

My mysql master-slave replication configuration method

Guess you like