基于docker的keepalived+MySQL主从实现MySQL高可用

因生产需要对MySQL做高可用，同时，资源有限，因此采用双节点主从+keepalived方式实现高勇。另外因需要大批量部署MySQL集群，需要采用模板化部署，本方案采用将MySQL容器化，实现MySQL模板化配置部署。

部署环境及软件：centos7.6、两个节点、keepalived-2.2.7、MySQL5.7.30

MySQL容器化部署配置参数

mysql采用mysql官方基础镜像mysql5.7
主节点ip：192.168.112.121
从节点ip：192.168.112.122
在主从节点上分别部署mysql容器，启动脚本

MySQL主从节点关键配置如下：

主节点（192.168.112.121）my.cnf配置

[client]
port=3306
default-character-set=utf8

[mysqld]
port=3306
server-id=1
log-bin=mysql-bin
binlog_do_db=xxx_db
binlog_format=mixed
character-set-server=utf8

从节点（192.168.112.122）my.cnf配置

[client]
port=3306
default-character-set=utf8

[mysqld]
port=3306
server-id=2
log-bin=mysql-bin
binlog_do_db=xxx_db
binlog_format=mixed
character-set-server=utf8

MySQL启动脚本

主数据库192.168.112.121

#!/bin/bash
docker rm -f msyql-server-1 || true
BASE_DIR=/home/services/mysql/mysql-1
docker run -d -p 3306:3306 \
   --name mysql-server-1 \
   --privileged=true --user=mysql \
   -e MYSQL_ROOT_PASSWORD=xxxx \
   -v $BASE_DIR/conf/my.cnf:/etc/mysql/my.cnf \
   -v $BASE_DIR/data:/var/lib/mysql \
   mysql:5.7

从数据库192.168.112.122

#!/bin/bash
docker rm -f msyql-server-2 || true
BASE_DIR=/home/services/mysql/mysql-1
docker run -d -p 3306:3306 \
   --name mysql-server-2 \
   --privileged=true --user=mysql \
   -e MYSQL_ROOT_PASSWORD=xxxx \
   -v $BASE_DIR/conf/my.cnf:/etc/mysql/my.cnf \
   -v $BASE_DIR/data:/var/lib/mysql \
   mysql:5.7

启动数据库之后做数据库主从配置

MySQL主从配置

登录主数据库创建主从同步账户slave_user，并查看master状态，记录下file、psition值，配置mysql从库时使用
操作步骤如下：

#授权slave_user复制权限
grant replication slave on *.* to slave_user@'%' identified BY '15dGt73@Adm';
flush PRIVILEGES;
# 需要登录验证账户及密码是否正确，如果能够正常登录，则继续后续步骤，否则需要排查原因并解决
mysql -h192.168.112.121 -uslave_user -p15dGt73@Adm
#查看master状态，找到master_log_file和MASTER_LOG_POS的值
show master status;
mysql> show master status;
+------------------+----------+--------------+------------------+-------------------+
| File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000003 |     2117| probe_db      |                  |                   |
+------------------+----------+--------------+------------------+-------------------+
1 row in set (0.01 sec)
# 选择从数据库执行
change master TO
MASTER_HOST='192.168.112.121',
master_port=3306,
master_user='slave_user',
master_password='15dGt73@Adm',
master_log_file='mysql-bin.000003',
MASTER_LOG_POS=2117;

验证从节点Slave_IO_Running、Slave_SQL_Running都为yes则主从配置成功

验证数据库主从同步是否生效，
第一组测试：在主库创建probe_db数据库，从库自动创建probe_db数据库，经验证从库自动创建；

第二组测试：数据库添加表，从库是否会自动创建数据表
在主数据库中执行创建数据库，如果从数据库也自动创建，则表示主从同步生效

CREATE TABLE `probe_db`.`ping_test_result` (
  `id` INT NOT NULL,
  `ip` VARCHAR(32) NOT NULL,
  `package_loss_ratio` FLOAT NOT NULL)
ENGINE = InnoDB
DEFAULT CHARACTER SET = utf8
COLLATE = utf8_esperanto_ci;

从数据库自动创建表
第三组测试：从主数据插入一组数据，验证是否会将数据同步到从库；

INSERT INTO ping_test_result(id, ip, package_loss_ratio) VALUES(1, '8.8.8.8', 0.1);
INSERT INTO ping_test_result(id, ip, package_loss_ratio) VALUES(2, '8.8.8.8', 0.1);
INSERT INTO ping_test_result(id, ip, package_loss_ratio) VALUES(3, '8.8.8.8', 0.1);

检测从库也是自动生产3条数据，数据同步生效
第四组测试：往从库里插入一组数据，验证从库是否自动将数据同步到主库；

INSERT INTO ping_test_result(id, ip, package_loss_ratio) VALUES(4, '8.8.8.8', 0.1);
INSERT INTO ping_test_result(id, ip, package_loss_ratio) VALUES(5, '8.8.8.8', 0.1);
INSERT INTO ping_test_result(id, ip, package_loss_ratio) VALUES(6, '8.8.8.8', 0.1);

检测主库也是自动生产3条数据，数据同步生效

第五组测试：关闭主库，往从库里插入一组数据，验证从库在主库正常启动后是否自动将数据同步到主库；

INSERT INTO ping_test_result(id, ip, package_loss_ratio) VALUES(7, '8.8.8.8', 0.1);
INSERT INTO ping_test_result(id, ip, package_loss_ratio) VALUES(8, '8.8.8.8', 0.1);
INSERT INTO ping_test_result(id, ip, package_loss_ratio) VALUES(9, '8.8.8.8', 0.1);

到此主从同步配置完成，下面开始配置keepalived服务。

keepalived主从配置

keepalived主要是需要将各种配置项配置正确
keepalived主节点：192.168.112.121
keepalived从节点：192.168.112.122
接下来是安装keepalived服务

安装必要的依赖包
yum -y install gcc openssl-devel popt-devel psmisc
cd /home/services/
wget https://www.keepalived.org/software/keepalived-2.2.7.tar.gz
tar -zxvf keepalived-2.2.7.tar.gz
cd keepalived-2.2.7
./configure --prefix=/opt/keepalived
make && make install

#存放配置文件并将keepalived启动脚本存在合适的位置
mkdir -p /etc/keepalived
cp keepalived/etc/keepalived/keepalived.conf.sample /etc/keepalived/keepalived.conf
cp keepalived/etc/init.d/keepalived /etc/init.d/
cp keepalived/etc/sysconfig/keepalived /etc/sysconfig/
cp bin/keepalived /usr/sbin/

在为主从节点做配置前，
1、需要查看网卡，主从节点的网络均是ens33
2、配置主从节点
3、创建/etc/keepalived/restart.sh

#!/bin/bash
#当mysql服务异常时，将keepalived应用停止，可以及时将vip切换到另一个节点上
killall keepalived

4、其他配置请参考外部资料
https://blog.csdn.net/D1179869625/article/details/126198495

主节点192.168.112.121配置文件如下：

!Confiquration File for keepalived
global defs {
	notification email {
		[email protected]
		[email protected]
		[email protected]
	}
	notification_email_from [email protected]
	smtp_server 10.xx.xx.x
	smtp_connect_timeout 30
	router_id LVS_MASTER
	vrrp_skip_check_adv_addr
	vrrp_strict
	vrrp_garp_interval 0
	vrrp_gna_interval 0
}
	vrrp_instance VI_1{
	    state BACKUP
		interface ens33
		virtual_router_id 51
		priority 90
		advert_int 1
		authentication{
		  auth_type PASS
		  auth_pass 1539
		}	
		virtual_ip_address {
		  192.168.112.120
		}
	}
	virtual_server 192.168.112.121 3306 {
	delay_loop 6
	1b_algo rr
	1b_kind NAT
	persistence_timeout 50
	protocol TCP
	real_server 192.168.112.121 3306{
	   notify_down /etc/keepalived/restart.sh
		TCP_CHECK {
			connect_ip 192.168.112.121
			connect_port 3306
			connect_timeout 3
			nb_get_retry 33
			delay_before_retry 3
		}
	}
}

经验证，将virtual_server服务配置到对应3306端口，一旦数据库异常，将会做故障迁移到另一个数据库实例上。

从节点192.168.112.122配置

!Confiquration File for keepalived
global defs {
	notification email {
		[email protected]
		[email protected]
		[email protected]
	}
	notification_email_from [email protected]
	smtp_server 10.xx.xx.x
	smtp_connect_timeout 30
	router_id LVS_SLAVE
	vrrp_skip_check_adv_addr
	vrrp_strict
	vrrp_garp_interval 0
	vrrp_gna_interval 0
}
	vrrp_instance VI_1{
	    state BACKUP
		interface ens33
		virtual_router_id 51
		priority 90
		advert_int 1
		authentication{
		  auth_type PASS
		  auth_pass 1539
		}	
		virtual_ip_address {
		  192.168.112.120
		}
	}
	virtual_server 192.168.112.122 3306 {
	delay_loop 6
	1b_algo rr
	1b_kind NAT
	persistence_timeout 50
	protocol TCP
	real_server 192.168.112.122 3306{
	   notify_down /etc/keepalived/restart.sh
		TCP_CHECK {
			connect_ip 192.168.112.122
			connect_port 3306
			connect_timeout 3
			nb_get_retry 33
			delay_before_retry 3
		}
	}
}

在两节点上配置keepalived服务为开机自启动并启动服务

systemctl enable keepalived && systemctl start keepalived

检测朱从节点服务状态为下图则表示服务正常运行，

测试一：通过vip（192.168.112.120）地址连接MySQL数据库，如果能正常连接则表示VIP生效；
测试二：关闭其中一个节点的keepalived服务，如果能正常连接则表示keepalived服务生效，否则异常，需要检查keepalived服务配置；
测试三：关闭其中一个数据库服务，如果能正常连接，则表示运行正常，否则异常，需要检查keepalived服务配置是否有误；
测试四：如果关闭主节点keepalived服务，并关闭从节点MySQL服务，如果无法访问，则表示正常。
至此整个高可用方案验证完成。