keepalived结合MHA实现mysql高可用

本例结合本博另一篇文章 MHA-大杀器实现MYSQL单主宕机时,VIP漂移实现高可用，环境请参考上一篇文章

本例使用keepalived-1.1.20.tar.gz版本

[root@node1 keepalived-1.1.20]# ./configure --sysconf=/etc/ #此项表示设置keepalived的根目录

如果编译报错“configure: error: Popt libraries is required” 则yum -y install popt-devel

[root@node1 keepalived-1.1.20]# make && make install

[root@node1 keepalived]# cp /usr/local/sbin/keepalived /usr/sbin/   #拷贝命令文件

#####################主机的配置文件设置####################

[root@node1 keepalived]# cat /etc/keepalived/keepalived.conf 

! Configuration File for keepalived

global_defs { #全局配置项

   router_id node1 #本机标识号，设置全局唯一

}

vrrp_script check_run { #设置检测服务的脚本

script "/root/check_mysql.sh"

interval 1 #每隔一秒检测一次

}

vrrp_instance mha {    #定义keepalived组

    state MASTER #定义本机的keepalived状态

    interface eth0 #设置监听网卡

    virtual_router_id 100 #设置虚拟router-id，所有keepalived的机器都得是一样

    priority 100   #设置权值，master必须高于backup

    advert_int 5 #定义master与backup之间同步检查间隔，单位是秒

    track_interface { #设置监听网卡组

eth0

eth1

}

    authentication { #设置keepalived组的验证类型和密码，所有主机须一样

        auth_type PASS

        auth_pass 1111

    }

track_script { #设置服务监测脚本运行

check_run

}

    virtual_ipaddress { #设置VIP

        10.2.16.250/24 

    }

}

#####################备机的配置文件设置####################

[root@node2 ~]# cat /etc/keepalived/keepalived.conf 

! Configuration File for keepalived

global_defs {

   router_id node2

}

vrrp_script check_run {

        script "/root/check_mysql.sh"

        interval 1

}

vrrp_instance mha {

    state BACKUP

    interface eth0

    virtual_router_id 100

    priority 99

    advert_int 5

track_interface {

eth0

eth1

}

 track_interface {

        eth0

        eth1

        }

  track_script {

        check_run

        }

    authentication {

        auth_type PASS

        auth_pass 1111

    }

    virtual_ipaddress {

10.2.16.250/24

    }

}

mysql服务检测脚本内容，按需修改：

[root@node2 ~]# cat check_mysql.sh 

#!/bin/bash

MYSQL=/usr/bin/mysql

MYSQL_HOST=127.0.0.1

MYSQL_USER=root

MYSQL_PASSWORD=123456

CHECK_TIME=3

#mysql  is working MYSQL_OK is 1 , mysql down MYSQL_OK is 0

MYSQL_OK=1

function check_mysql_helth (){

$MYSQL -h $MYSQL_HOST -u $MYSQL_USER -e "show status;" --password=123456 >/dev/null 2>&1

if [ $? = 0 ] ;then

     MYSQL_OK=1

else

     MYSQL_OK=0

fi

     return $MYSQL_OK

}

while [ $CHECK_TIME -ne 0 ]

do

     let "CHECK_TIME -= 1"

     check_mysql_helth

if [ $MYSQL_OK = 1 ] ; then

     CHECK_TIME=0

     exit 0

fi

if [ $MYSQL_OK -eq 0 ] &&  [ $CHECK_TIME -eq 0 ]

then

     pkill keepalived

exit 1

fi

sleep 1

done

启动Keepalived：

[root@node1 ~]#service keepalived start

启动日志：

[root@node1 ~]# tail -f /var/log/messages

Jun 27 11:48:47 node2 Keepalived: Starting VRRP child process, pid=21545

Jun 27 11:48:47 node2 Keepalived_vrrp: Registering Kernel netlink reflector

Jun 27 11:48:47 node2 Keepalived_vrrp: Registering Kernel netlink command channel

Jun 27 11:48:47 node2 Keepalived_vrrp: Registering gratutious ARP shared channel

Jun 27 11:48:47 node2 Keepalived_vrrp: Opening file ‘/etc/keepalived/keepalived.conf‘.

Jun 27 11:48:47 node2 Keepalived_vrrp: Configuration is using : 65535 Bytes

Jun 27 11:48:47 node2 Keepalived_vrrp: Using LinkWatch kernel netlink reflector...

Jun 27 11:48:47 node2 Keepalived_vrrp: VRRP_Instance(mha) Entering BACKUP STATE

Jun 27 11:48:47 node2 Keepalived_vrrp: VRRP sockpool: [ifindex(2), proto(112), fd(9,10)]

Jun 27 11:48:47 node2 Keepalived_vrrp: VRRP_Script(check_run) succeeded

Jun 27 11:49:03 node2 Keepalived_vrrp: VRRP_Instance(mha) Transition to MASTER STATE

Jun 27 11:49:08 node2 Keepalived_vrrp: VRRP_Instance(mha) Entering MASTER STATE

Jun 27 11:49:08 node2 Keepalived_vrrp: VRRP_Instance(mha) setting protocol VIPs.

Jun 27 11:49:08 node2 Keepalived_vrrp: VRRP_Instance(mha) Sending gratuitous ARPs on eth0 for 10.2.16.250

Jun 27 11:49:13 node2 Keepalived_vrrp: VRRP_Instance(mha) Sending gratuitous ARPs on eth0 for 10.2.16.250

查看 VIP:

[root@node1 ~]# ip addr

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN 

    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00

    inet 127.0.0.1/8 scope host lo

    inet6 ::1/128 scope host 

       valid_lft forever preferred_lft forever

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000

    link/ether 00:0c:29:c5:22:55 brd ff:ff:ff:ff:ff:ff

    inet 10.2.16.253/24 brd 10.2.16.255 scope global eth0

inet 10.2.16.250/24 brd 10.2.16.255 scope global eth0

    inet6 fe80::20c:29ff:fec5:2255/64 scope link 

       valid_lft forever preferred_lft forever

3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000

    link/ether 00:0c:29:c5:22:5f brd ff:ff:ff:ff:ff:ff

    inet 10.0.0.1/24 brd 10.0.0.255 scope global eth1

    inet6 fe80::20c:29ff:fec5:225f/64 scope link 

       valid_lft forever preferred_lft forever

  此VIP用ifconfig看不到，但是真实存在的。

####check_mysql.sh 需要在两台机器上都存在

  ###############keepalived+mha+check_mysql运行原理#################

  实现VIP飘移，需要将node1的Mysql服务停掉，此时keepalived中运行的脚本会检测mysql，检测到mysql挂了之后，会pkill keepalived

  当BACKUP-keepalived检测到主keepalived挂掉之后，会自动启动VIP，而MHA检测到master的mysql挂掉之后，也会将Master移至BAKCUP上

  此时所有的权限都移至Mysql-backup上，并且他已经成为MASTER，接替宕机的master所有工作，从而实现高可用。

本文出自 “Fate” 博客，请务必保留此出处http://czybl.blog.51cto.com/4283444/1431693

keepalived结合MHA实现mysql高可用,古老的榕树,5-wow.com