多台slave的mysql Replication的配置
5.1.24版本的配置请看：http://www.masalife.com/archives/173
一、先修改服务器的配置文件
1、Master服务器配置简单，修改my.cnf为:
server-id = 1
log-bin
set-variable=binlog-ignore-db=mysql
2、slave1的配置加入
server-id = 2
master-host = 172.16.20.135
master-user = rep
master-password = cnrep
master-port = 3306
log-bin
set-variable=replicate-ignore-db=mysql
set-variable=replicate-do-db=AliSMS
set-variable=replicate-do-db=lcd
set-variable=replicate-do-db=loginmanager
set-variable=replicate-do-db=samis
set-variable=replicate-do-db=sareport
set-variable=replicate-do-db=syslog
set-variable=replicate-do-db=web_speed
log-slave-updates
3、slave2服务器的配置
server-id = 3
master-host = 172.16.20.3
master-user = rep1
master-password = cnrep
master-port = 3306
set-variable=replicate-ignore-db=mysql
set-variable=replicate-do-db=AliSMS
set-variable=replicate-do-db=lcd
set-variable=replicate-do-db=loginmanager
set-variable=replicate-do-db=samis
set-variable=replicate-do-db=sareport
set-variable=replicate-do-db=syslog
set-variable=replicate-do-db=web_speed
二、重启master数据库
三、然后锁定master数据库的表：
mysql>FLUSH TABLES WITH READ LOCK;
四、在master数据库中添加用于slave1同步的用户，并赋予相关权限：
mysql>GRANT REPLICATION SLAVE ON *.* TO rep@sa_cfengine1 IDENTIFIED BY ‘cnrep’;
mysql>GRANT FILE,SELECT,REPLICATION SLAVE ON *.* TO rep@sa_cfengine1 IDENTIFIED BY ‘cnrep’;
五、在slave1数据库中添加用于slave2同步的用户，并赋予相关权限：
mysql>GRANT REPLICATION SLAVE ON *.* TO rep1@sa_cfengine2 IDENTIFIED BY ‘cnrep’;
mysql>GRANT FILE,SELECT,REPLICATION SLAVE ON *.* TO rep1@sa_cfengine2 IDENTIFIED BY ‘cnrep’;
六、同步数据库：
方法很多，可以打包之后scp，再解压，由于sa_cfengine1到mysql master服务器通道打通了，切sa_cfengine2到sa_cfengine1通道也打了，故直接scp整个数据库目录即可。
注意：此时要注意删除同步过来的日志文件，最好把与数据库无关的文件全删除（可以将非目录的文件全删了）。
七、重启salve1的mysql，起来之后锁定表
八、重启slave2的mysql，然后先后给slave1和master服务器的mysql表解锁
mysql> UNLOCK TABLES；
九、分别登录slave1和slave2的mysql，查看同步状态：
mysql>SHOW SLAVE STATUS\G
*************************** 1. row ***************************
Slave_IO_State: Waiting for master to send event
Master_Host: 172.16.20.135
Master_User: rep
Master_Port: 3306
Connect_Retry: 60
Master_Log_File: mysql-bin.000051
Read_Master_Log_Pos: 13856842
Relay_Log_File: sa_cfengine1-relay-bin.000013
Relay_Log_Pos: 624419
Relay_Master_Log_File: mysql-bin.000051
** Slave_IO_Running: Yes**
Slave_SQL_Running: Yes
Replicate_Do_DB: AliSMS,lcd,loginmanager,samis,sareport,syslog,web_speed
Replicate_Ignore_DB: mysql,mysql
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: 13856842
Relay_Log_Space: 624419
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
1 row in set (0.01 sec)
注意标注为红色的地方，两个都是yes说明一切正常，否则要检查原因，可以看error log查找原因后做相应的处理。
十、测试：
在master数据库中update在同步列表中的一个表的一个字段，如果slave服务器的做相应改变，则测试用过。
原创文章如转载，请注明：转载自五四陈科学院[http://www.54chen.com]
Posted by 54chen linux
« yahoo、taobao云计算利器之“云”端的小飞象—Hadoop 如何解决时间同步引起oracle 不稳定的问题 »