一、环境信息
| 主机名 | ip | 操作系统 | mysql版本 | VIP（虚拟ip） |
| hadoop01 | 192.168.10.200 | centos7_x86 | 5.7 | 192.168.10.253 |
| hadoop03 | 192.168.10.202 | centos7_x86 | 5.7 |
二、mysql集群搭建
两台节点，如果未部署mysql服务，部署文档请看【Mysql】mysql三种安装方式（二进制、yum、docker）-CSDN博客
三、配置文件修改
hadoop01节点的my.cnf配置，默认情况下MySQL的配置文件是/etc/my.cnf
[mysqld_safe]
pid-file=/data/mysql5.7/logs/mysqld.pid
[mysqld]
basedir=/usr/local/mysql5.7
datadir=/data/mysql5.7/data
socket=/data/mysql5.7/run/mysql.sock
log_error=/data/mysql5.7/logs/alert.log
pid_file=/data/mysql5.7/logs/mysqld.pid##新增内容如下
#设置server-id,两节点必须不一样
server-id=1
#开启binlog,并指定存储位置
log-bin=/data/mysql5.7/logs/mysql-bin
#中继日志
relay-log =/data/mysql5.7/logs/mysql-relay-bin
#忽略同步的库
replicate-ignore-db = mysql
replicate-ignore-db = information_schema
replicate-ignore-db = performance_schema
replicate-ignore-db = sys
hadoop03节点的my.cnf配置
[mysqld_safe]
pid-file=/data/mysql5.7/logs/mysqld.pid
[mysqld]
basedir=/usr/local/mysql5.7
datadir=/data/mysql5.7/data
socket=/data/mysql5.7/run/mysql.sock
log_error=/data/mysql5.7/logs/alert.log
pid_file=/data/mysql5.7/logs/mysqld.pid##新增内容如下
#设置server-id,两节点必须不一样
server-id=2
#开启binlog
log-bin=/data/mysql5.7/logs/mysql-bin
#中继日志
relay-log=/data/mysql5.7/logs/mysql-relay-bin
#忽略同步的库
replicate-ignore-db = mysql
replicate-ignore-db = information_schema
replicate-ignore-db = performance_schema
replicate-ignore-db = sys
四、数据保持一致性
如果Hadoop01上已经有mysql数据，那么在执行主主互备之前，需要将mysql的数据保持同步，首先在Hadoop01上备份 mysql数据，执行sql命令：FLUSH TABLES WITH READ LOCK; 不要退出这个终端，否则这个锁就失效了。
在不退出终端的情况下，使用 mysqldump工具来导出数据，再导入宁外一个数据库中。
重启数据库，使之前配置文件参数生效；
五、主主互备模式配置
1）配置hadoop01主，hadoop02从
#登录hadoop01数据库，创建同步用户
mysql> grant replication slave on *.* to 'cuixiaoqin'@'192.168.10.%' identified by '520_1314';
#刷新权限表
mysql> flush privileges;
#查看hadoop01的master_log_file和master_log_pos
mysql> show master status;
+------------------+----------+--------------+------------------+-------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000001 | 454 | | | |
+------------------+----------+--------------+------------------+-------------------+
1 row in set (0.00 sec)#登录hadoop02数据库，配置主从，master_log_file和master_log_pos对应hadoop01的master信息
mysql> change master to master_host='192.168.10.200',master_user='cuixiaoqin',master_password='520_1314',master_log_file='mysql-bin.000001',master_log_pos=454;
#开启从
mysql> start slave;
#查看主从状态，Slave_IO_Running和Slave_SQL_Running都为yes代表正常
mysql> show slave status\G
2）配置hadoop02主，hadoop01从
#登录hadoop02数据库，创建同步用户
mysql> grant replication slave on *.* to 'cuixiaoqin'@'192.168.10.%' identified by '520_1314';
#刷新权限表
mysql> flush privileges;
#查看hadoop02的master_log_file和master_log_pos
mysql> show master status;
mysql> show master status;
+------------------+----------+--------------+------------------+-------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000001 | 606 | | | |
+------------------+----------+--------------+------------------+-------------------+
1 row in set (0.00 sec)#登录hadoop01数据库，配置主从，master_log_file和master_log_pos对应hadoop01的master信息
mysql> change master to master_host='192.168.10.202',master_user='cuixiaoqin',master_password='520_1314',master_log_file='mysql-bin.000001',master_log_pos=606;
#开启从
mysql> start slave;
#查看主从状态，Slave_IO_Running和Slave_SQL_Running都为yes代表正常
mysql> show slave status\G
六、配置KeepAlived实现MySQL双主高可用
hadoop01和hadoop03都执行相同操作，keepalived官网Keepalived for Linux
#安装keepalived依赖
[root@hadoop01 ~]# yum install gcc gcc-c++ openssl-devel -y
#离线安装，也可以yum install安装
[root@hadoop01 ~]# wget -c https://keepalived.org/software/keepalived-2.1.5.tar.gz
[root@hadoop01 ~]# tar -xvf keepalived-2.1.5.tar.gz
[root@hadoop03 ~]# mv keepalived-2.1.5 keepalived
[root@hadoop03 ~]# mv keepalived /usr/local/
[root@hadoop01 ~]# cd /usr/local/keepalived
[root@hadoop01 keepalived]# ./configure --prefix=/usr/local/keepalived
[root@hadoop01 keepalived]# make && make install#编写监控mysql状态脚本
[root@hadoop01 ~]# vi /etc/keepalived/check_mysql.sh
#!/bin/bash
#检查3306端口是否存在
ss -lntup|grep 3306 > /dev/null 2>&1
if [[ $? -eq 0 ]]; thenexit 0
elsesystemctl stop keepalived
fi
[root@hadoop01 ~]# chmod +x /etc/keepalived/check_mysql.shhadoop01的keepalived.conf配置文件设置
[root@hadoop01 keepalived]# mkdir /etc/keepalived
[root@hadoop01 keepalived]# vi /etc/keepalived/keepalived.conf
! Configuration File for keepalivedglobal_defs {#在 keepalived 状态变化时接收通知邮件的地址notification_email {csdn@126.com}#邮件的发件人地址notification_email_from ops@wangshibo.cn#发送通知邮件所使用的 SMTP 服务器地址smtp_server 127.0.0.1#连接到 SMTP 服务器的超时时间smtp_connect_timeout 30#为 VRRP 实例指定了一个唯一的标识符router_id MYSQL-01#启用脚本安全性检查enable_script_security
}#定义了一个 VRRP 脚本 chk_mysql，用于监控脚本的执行状态
vrrp_script chk_mysql {#指定了要执行的脚本路径script "/etc/keepalived/check_mysql.sh"#设定了脚本执行的时间间隔，单位是秒interval 2#指定用户user root
}#定义了一个 VRRP 实例，VI_1 是这个实例的名称
vrrp_instance VI_1 {#设定角色为BACKUP，都是设成BACKUP则以优先级为主要参考state BACKUP#指定了用于 VRRP 的网络接口interface ens33#指定了 VRRP 实例的虚拟路由器 ID，这个 ID 在相同的 VRRP 组中必须唯一。virtual_router_id 51#设定节点的优先级priority 101#指定了主节点向备份节点广播状态信息的间隔时间，单位秒advert_int 1#设置认证方式，auth_type PASS 表示使用简单的密码认证，auth_pass 1111 是认证密码authentication {auth_type PASSauth_pass 1111}#虚拟 IP 地址virtual_ipaddress {192.168.10.253}#用于跟踪脚本的状态，chk_mysql 是定义的脚本名称，keepalived 会根据这个脚本的执行结果来决定是否调整状态track_script {chk_mysql}
}hadoop02的keepalived.conf配置文件设置
[root@hadoop03 ~]# mkdir /etc/keepalived
[root@hadoop03 ~]# vi /etc/keepalived/keepalived.conf
! Configuration File for keepalivedglobal_defs {notification_email {csdn@126.com}notification_email_from ops@wangshibo.cnsmtp_server 127.0.0.1smtp_connect_timeout 30#设置 VRRP 实例的一个唯一的标识符router_id MYSQL-02enable_script_security
}vrrp_script chk_mysql {script "/etc/keepalived/check_mysql.sh"interval 2user root
}vrrp_instance VI_1 {state BACKUPinterface ens33virtual_router_id 51#设置优先级为100priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.10.253}track_script {chk_mysql}
}hadoop01和hadoop02重启keepalived服务
systemctl start keepalived
systemctl status keepalived七、验证keepalived虚拟ip地址是否正常
#目前vip在Hadoop01节点
#将hadoop01节点mysql停掉，查看虚拟ip地址是否在hadoop02上
