一．体系架构

在Keepalived + amoeba高可用负载均衡架构中，keepalived负责实现High-availability (HA) 功能控制前端机VIP（虚拟网络地址），当有设备发生故障时，热备服务器可以瞬间将VIP自动切换过来，实际运行中体验只有2秒钟切换时间，
通过amoeba mysql中间件对master和slave进行读写分离，主master负责写，两台slave负责读，备master负责实时同步主master的数据

二. 优点

读写分离
mysql slave支持横向扩展，提高性能
数据库内容支持定期备份，在发生重大操作失误后可以进行回退操作
mysql master热备，能在master出现问题后自动切换到另一台备份master上继续使用

三. 系统环境

两台负载机器安装：centos7.2+docker+amoeba+keepalived，分别命名为：MYSQL_MASTER，MYSQL_BACKUP。
多台mysql slave，这里使用两台

服务器操作系统 IP地址安装软件
MYSQL_MASTER Centos 7.2 64位 10.141.1.31 docker+amoeba+keepalived
MYSQL_BACKUP Centos 7.2 64位 10.141.1.5 docker+amoeba+keepalived
MYSQL_SLAVE_1 Centos 7.2 64位 10.141.1.12 docker+mysql
MYSQL_SLAVE_2 Centos 7.2 64位 10.141.1.25 docker+mysql
虚拟IP 10.141.1.8

四. 搭建环境

1. 主机准备

全部主机执行命令

setenforce 0 #关闭selinux
sed -i 's/SELINUX=enforcing/SELINUX=disabled/g' /etc/selinux/config
systemctl stop firewalld #关闭防火墙
systemctl stop iptables #关闭iptables

2. docker安装(全部主机执行命令)

a. 在线安装

参考: (https://docs.docker.com/install/linux/docker-ce/centos/#uninstall-old-versions)

yum install docker

b. 离线二进制安装：

参考：(https://www.jianshu.com/p/46b9a351f749)

3. 安装mysql（全部安装）

a. 拉取镜像

docker pull mysql:5.7

b. 修改mysql配置文件，vim mysqld.cnf

[mysqld]
pid-file    = /var/run/mysqld/mysqld.pid
socket      = /var/run/mysqld/mysqld.sock
datadir     = /var/lib/mysql
#log-error  = /var/log/mysql/error.log
# By default we only accept connections from localhost
#bind-address   = 127.0.0.1
# Disabling symbolic-links is recommended to prevent assorted security risks
symbolic-links=0
log-bin=mysql-bin
server_id = 11               #备份写12，不能一样
auto-increment-offset = 1    #备份mysql填写0
auto-increment-increment = 2

c. 启动mysql

docker run --name mysql -e MYSQL_ROOT_PASSWORD=1234 -p 3306:3306 -v /${PWD}/mysqld.cnf:/etc/mysql/mysql.conf.d/mysqld.cnf -d mysql:5.7

4. 设置mysql双主多从

a. 赋权

（1）master给backup和两个slave增加复制账号

grant replication slave on *.* to 'repl'@'10.141.1.5' identified by '123456';
grant replication slave on *.* to 'repl'@'10.141.1.12' identified by '123456';
grant replication slave on *.* to 'repl'@'10.141.1.25' identified by '123456';
flush  privileges;

（2）backup给master和两个slave增加复制账号

grant replication slave on *.* to 'repl'@'10.141.1.31' identified by '123456';
grant replication slave on *.* to 'repl'@'10.141.1.12' identified by '123456';
grant replication slave on *.* to 'repl'@'10.141.1.25' identified by '123456';
flush  privileges;

（3）设置backup同步master

show master status; #在master执行，记住File和Position 值
change master to master_host='10.141.1.31',master_port=3306,master_user='repl',master_password='123456',master_log_file='mysql-bin.000003',master_log_pos=1207; #在backup执行，master_log_file为上面的File值，master_log_pos为Position 值
start slave; #启动
show slave status\G; 查看Slave_IO_Running和Slave_SQL_Running两个值为yes即成功

image.png

（4）设置master同步backup

show master status; #在backup执行，记住File和Position 值
change master to master_host='10.141.1.5',master_port=3306,master_user='repl',master_password='123456',master_log_file='mysql-bin.000003',master_log_pos=1050; #在master执行，master_log_file为上面的File值，master_log_pos为Position 值
start slave; #启动
show slave status\G; 查看Slave_IO_Running和Slave_SQL_Running两个值为yes即成功

image.png

（5）设置slave同步master

show master status; #在master执行，记住File和Position 值
change master to master_host='10.141.1.8',master_port=3306,master_user='repl',master_password='123456',master_log_file='mysql-bin.000003',master_log_pos=1207; #在backup执行，master_log_file为上面的File值，master_log_pos为Position 值
start slave; #启动
show slave status\G; 查看Slave_IO_Running和Slave_SQL_Running两个值为yes即成功

5. 读写分离

a. 分配amoeba账号，全部执行

grant all on *.* to 'amoeba'@'%' identified by '123';

b. 启动amoeba服务

docker run --name amoeba -it -d -p 8066:8066 -v /${PWD}/conf:/usr/local/amoeba/conf commanderhu/amoeba

启动后修改conf下的amoeba.xml ，dbServers.xml 文件，然后重启

docker restart amoeba

c. 测试

mysql -uamoeba -p123 -P8066 -h10.141.1.8
show databases;

6. keepalived安装配置

参考（https://www.jianshu.com/p/72dfd4d4cf9bc）

master配置

! Configuration File for keepalived

global_defs {
   notification_email {
     acassen@firewall.loc
     failover@firewall.loc
     sysadmin@firewall.loc
   }
   notification_email_from Alexandre.Cassen@firewall.loc
   smtp_server 192.168.200.1
   smtp_connect_timeout 30
   router_id LVS_DEVEL
   vrrp_skip_check_adv_addr
   #vrrp_strict
   vrrp_garp_interval 0
   vrrp_gna_interval 0
}

vrrp_script chk_mysql_port {     #检测mysql服务是否在运行。有很多方式，比如进程，用脚本检测等等
    script "/root/chk_mysql.sh"   #这里通过脚本监测
    interval 2                   #脚本执行间隔，每2s检测一次
    weight -10                    #脚本结果导致的优先级变更，检测失败（脚本返回非0）则优先级 -5
    fall 2                    #检测连续2次失败才算确定是真失败。会用weight减少优先级（1-255之间）
    rise 1                    #检测1次成功就算成功。但不修改优先级
}

vrrp_instance VI_1 {
    state MASTER
    #nopreempt
    notify_master /root/to_master.sh
    unicast_src_ip  10.141.1.29
    unicast_peer {
                 10.141.1.24
    }
    interface eth0
    virtual_router_id 58
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        10.141.1.8
    }

track_script {               
   chk_mysql_port             
}
}

slave配置

! Configuration File for keepalived

global_defs {
   notification_email {
     acassen@firewall.loc
     failover@firewall.loc
     sysadmin@firewall.loc
   }
   notification_email_from Alexandre.Cassen@firewall.loc
   smtp_server 192.168.200.1
   smtp_connect_timeout 30
   router_id LVS_DEVEL
   vrrp_skip_check_adv_addr
   #vrrp_strict
   vrrp_garp_interval 0
   vrrp_gna_interval 0
}
vrrp_script chk_mysql_port {     #检测mysql服务是否在运行。有很多方式，比如进程，用脚本检测等等
    script "/root/chk_mysql.sh"   #这里通过脚本监测
    interval 2                   #脚本执行间隔，每2s检测一次
    weight -10                    #脚本结果导致的优先级变更，检测失败（脚本返回非0）则优先级 -5
    fall 2                    #检测连续2次失败才算确定是真失败。会用weight减少优先级（1-255之间）
    rise 1                    #检测1次成功就算成功。但不修改优先级
}

vrrp_instance VI_1 {
    state BACKUP
    notify_master /root/to_master.sh
    #nopreempt
    unicast_src_ip  10.141.1.24
    unicast_peer {
                 10.141.1.29
    }
    interface eth0
    virtual_router_id 58
    priority 95
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        10.141.1.8
    }

   track_script {
     chk_mysql_port
   }
}

slave变master执行的脚本

re=`mysql -h10.141.1.29 -uroot -p1234 -e "show master status;" | grep -v File`

echo $re

binlog=`echo $re | awk '{print $1}'`
position=`echo $re | awk '{print $2}'`
echo $binlog ------ $position

mysql -h10.141.1.30 -uroot -p1234 -e "stop slave;change master to master_host='10.141.1.8',master_port=3306,master_user='repl',master_password='123456',master_log_file='$binlog',master_log_pos=$position;start slave;show slave status\G;"
mysql -h10.141.1.4 -uroot -p1234 -e "stop slave;change master to master_host='10.141.1.8',master_port=3306,master_user='repl',master_password='123456',master_log_file='$binlog',master_log_pos=$position;start slave;show slave status\G;"

检查脚本

counter=$(netstat -na|grep "LISTEN"|grep "3306"|wc -l)
if [ "${counter}" -eq 0 ]; then
    #systemctl stop keepalived
    exit 1
fi

mysql高可用-读写分离-双主多从