mysql5.7主从配置

一、主库(10.159.186.149)配置

修改配置文件/etc/my.cnf：开启binlog日志，添加server_id
重启mysql:service mysql restart
登录mysql:mysql -u root -p
访问mysql库:use mysql;
新建存库可访问账号slave并赋予权限：
select user,authentication_string,host from user;
create user 'slaver'@'10.159.176.139' identified by 'slaver';
grant replication slave on *.* to 'slaver'@'10.159.176.139';
flush privileges;
检查账号:
select user,authentication_string,host from user;
show grants for 'slaver'@'10.159.176.139';
检查主库的状态：show master status;

二、从库(10.159.76.139)配置：

修改配置文件/etc/my.cnf:添加server_id,修改为只读(对root用户不生效)
重启mysql:service mysql restart
登录mysql:mysql -u root -p
检查slave从数据库额各参数
show variables like '%log_bin%';
show variables like '%server_id%';
配置复制参数，slave从库连接master主库的配置
mySQL>change master to master_host='10.159.176.54',master_user='slaver',master_password='slaver',master_log_file='mysql-bin.000011',master_log_pos=432;
启动从库的同步开关
start slave;
查看复制状态
show slave status\G
设置只读账户：
mysql> create user 'read_user'@'%' identified by '222222';
mysql> grant select on *.* to 'read_user'@'%';
mysql> flush privileges;

三、mysql多线程复制配置(可以看作是基于table的)

修改从库配置文件/etc/my.cnf，添加以下内容：
    slave-parallel-type=LOGICAL_CLOCK
    slave-parallel-workers=16       # ###打开16个线程，原先的进程变成调度器
    ###下面是优化参数
    master_info_repository=TABLE
    relay_log_info_repository=TABLE
    relay_log_recovery=ON
    symbolic-links=0

查看：
select * from slave_relay_log_info\G;
show processlist;

完整的配置文件

完整的从库配置文件(/etc/my.cnf)
        [mysqld]
        datadir=/usr/local/data/mysql-data
        basedir=/usr/local/src/mysql
        socket=/usr/local/data/mysql-data/mysql.sock
        user=mysql
        port=3306
        character-set-server=utf8
        # 取消密码验证
        #skip-grant-tables
        # Disabling symbolic-links is recommended to prevent assorted security risks
        symbolic-links=0
        slave-parallel-type=LOGICAL_CLOCK
        slave-parallel-workers=16       # ###打开16个线程，原先的进程变成调度器
        ###下面是优化参数
        master_info_repository=TABLE
        relay_log_info_repository=TABLE
        relay_log_recovery=ON
        # skip-grant-tables
        server-id=2
        read-only=true
        [mysqld_safe]
        log-error=/var/log/mysqld.log
        pid-file=/var/run/mysqld/mysqld.pid
完整的主库配置文件(/etc/my.cnf)
        [mysqld]
        datadir=/usr/local/data/mysql-data
        basedir=/usr/local/src/mysql
        socket=/usr/local/data/mysql-data/mysql.sock
        user=mysql
        port=3306
        character-set-server=utf8
        # 取消密码验证
        #skip-grant-tables
        # Disabling symbolic-links is recommended to prevent assorted security risks
        symbolic-links=0
        # skip-grant-tables
        server-id=1
        log-bin=/usr/local/data/mysql-data/binlog/mysql-bin
        binlog-format=Row
        [mysqld_safe]
        log-error=/var/log/mysqld.log
        pid-file=/var/run/mysqld/mysqld.pid

四、异常信息

 报错信息：The slave I/O thread stops because master and slave have equal MySQL server UUIDs;
 解决：将mysql data目录下边的auto.cnf移除
 mv  /data/mysql-data/auto.cnf /data/mysql-data/auto.cnf.bak