MySQL高可用之keepalived方案的示例分析

2025-06-24 08:03来源：互联网 [ 大中小 ]

实验环境
mysql master: 192.168.111.52
mysql slave: 192.168.111.53
keepalived vip : 192.168.111.60
、搭建过程
1. mysql双主的构建
① 互相 change master 即可，此处省略该过程，着重讲下keepalived

2. keepalived相关
①yum -y install keepalived 安装keepalived
②root@192.168.111.52:~# keepalived -v
Keepalived v1.2.7 (02/21,2013)

③ 编辑 /etc/keepalived/keepalived.conf

点击(此处)折叠或打开

root@192.168.111.52:keepalived# cat keepalived.conf

! Configuration File for keepalived

global_defs {##全局配置

notification_email {

horand_gc@163.com#####邮件接收者

}

notification_email_from dba@163.com#####邮件发送者

smtp_server smtp.163.com #####SMTP服务器

smtp_connect_timeout 30

router_id haMySQL #####routerID 同一组keepalived设置为相同

}

vrrp_script chk_mysql {##健康检测相关配置

script "/etc/keepalived/chk_mysql.sh" #####设置脚本或者直命令返回 0 则表示成功 >0 则表示失败，详情可以后面的脚本内容

interval 1#####检测间隔

weight 2 #####检测返回失败之后优先级会减少 2 (如果 master优先级100 ， slave 优先级 99 ,master检测失败则优先级为100-2 < 99 ，则slave会提升为主)

}

vrrp_instance VI_1 {

state MASTER ##### 设为master

interface eth2##### 网卡设定

virtual_router_id 51 ##### 针对该 instance的虚拟id , 同一组 instance设置相同

priority 100 ##### 优先级设定

advert_int 1 ##### 检测时间间隔

authentication {

auth_type PASS##### 同一组instance之间的认证方式为 PASS ,pass 为7777 ，必须相同(防止有用户恶意伪造 vrrp)

auth_pass 7777

}

virtual_ipaddress {

192.168.111.60##### 设置虚拟ip ,可以设置多个

}

track_script {

chk_mysql##### 表示该instance 使用chk_mysql进行相关检测

}

##### 以下配置在该实例转换为 master,slave，或者出错的时候执行的脚本(可以设置邮件通知，或者处理一些其他问题)

# notify_master "/etc/keepalived/change_master.sh"

# notify_slave "/etc/keepalived/change_slave.sh"

# notify_fault "/etc/keepalived/change_fault.sh"

}

root@192.168.111.53:keepalived# cat keepalived.conf

! Configuration File for keepalived

global_defs {

notification_email {

horand_gc@163.com

}

notification_email_from dba@163.com

smtp_server smtp.163.com

smtp_connect_timeout 30

router_id haMySQL

}

vrrp_script chk_mysql {

script "/etc/keepalived/chk_mysql.sh"

interval 1

weight 2

}

vrrp_instance VI_1 {

state BACKUP##### 该主机作为备机 BACKUP

interface eth2

virtual_router_id 51

priority 99##### 优先级设置小于 master

advert_int 1

authentication {

auth_type PASS

auth_pass 7777

}

virtual_ipaddress {

192.168.111.60

}

track_script {

chk_mysql

}

root@192.168.111.52:keepalived# cat chk_mysql.sh

#!/bin/bash

num=`ps -ef |grep mysqld | grep -v grep | wc -l`##### 查看mysqld进程数量，正常情况有一个root起的mysqld_safe守护进程，还有一个属于mysql用户的mysqld进程

[[ $num -eq 2 ]] && exit 0 || exit 1

3. 故障模拟
① 启动 keepalived

点击(此处)折叠或打开

root@192.168.111.52:keepalived# tail /var/log/messages

Apr 29 13:45:24 localhost Keepalived_healthcheckers[24184]: Registering Kernel netlink command channel

Apr 29 13:45:24 localhost Keepalived_healthcheckers[24184]: Opening file '/etc/keepalived/keepalived.conf'.

Apr 29 13:45:24 localhost Keepalived_healthcheckers[24184]: Configuration is using : 7417 Bytes

Apr 29 13:45:24 localhost Keepalived_vrrp[24185]: Opening file '/etc/keepalived/keepalived.conf'.

Apr 29 13:45:24 localhost Keepalived_vrrp[24185]: Configuration is using : 65552 Bytes

Apr 29 13:45:24 localhost Keepalived_vrrp[24185]: Using LinkWatch kernel netlink reflector...

Apr 29 13:45:24 localhost Keepalived_healthcheckers[24184]: Using LinkWatch kernel netlink reflector...

Apr 29 13:45:25 localhost Keepalived_vrrp[24185]: VRRP_Script(chk_mysql) succeeded

Apr 29 13:45:25 localhost Keepalived_vrrp[24185]: VRRP_Instance(VI_1) Transition to MASTER STATE

Apr 29 13:45:26 localhost Keepalived_vrrp[24185]: VRRP_Instance(VI_1) Entering MASTER STATE

root@192.168.111.53:keepalived# tailf /var/log/messages

Apr 29 13:49:27 localhost Keepalived_healthcheckers[32569]: Registering Kernel netlink reflector

Apr 29 13:49:27 localhost Keepalived_healthcheckers[32569]: Registering Kernel netlink command channel

Apr 29 13:49:27 localhost Keepalived_vrrp[32570]: Opening file '/etc/keepalived/keepalived.conf'.

Apr 29 13:49:27 localhost Keepalived_vrrp[32570]: Configuration is using : 65550 Bytes

Apr 29 13:49:27 localhost Keepalived_vrrp[32570]: Using LinkWatch kernel netlink reflector...

Apr 29 13:49:27 localhost Keepalived_healthcheckers[32569]: Opening file '/etc/keepalived/keepalived.conf'.

Apr 29 13:49:27 localhost Keepalived_healthcheckers[32569]: Configuration is using : 7415 Bytes

Apr 29 13:49:27 localhost Keepalived_vrrp[32570]: VRRP_Instance(VI_1) Entering BACKUP STATE

Apr 29 13:49:27 localhost Keepalived_healthcheckers[32569]: Using LinkWatch kernel netlink reflector...

Apr 29 13:49:27 localhost Keepalived_vrrp[32570]: VRRP_Script(chk_mysql) succeeded

ip a 可以查看到 vip 192.168.111.60 在 192.168.111.52(master)上

② 关闭 111.52上面的mysql

点击(此处)折叠或打开

root@192.168.111.52:keepalived# mysqladmin shutdown

root@192.168.111.52:keepalived# tailf /var/log/messages

Apr 29 14:19:30 localhost Keepalived_vrrp[24862]: VRRP_Script(chk_mysql) failed

Apr 29 14:19:32 localhost Keepalived_vrrp[24862]: VRRP_Instance(VI_1) Received higher prio advert

Apr 29 14:19:32 localhost Keepalived_vrrp[24862]: VRRP_Instance(VI_1) Entering BACKUP STATE

root@192.168.111.53:keepalived# tailf /var/log/messages

Apr 29 14:19:55 localhost Keepalived_vrrp[566]: VRRP_Instance(VI_1) forcing a new MASTER election

Apr 29 14:19:56 localhost Keepalived_vrrp[566]: VRRP_Instance(VI_1) Transition to MASTER STATE

Apr 29 14:19:57 localhost Keepalived_vrrp[566]: VRRP_Instance(VI_1) Entering MASTER STATE

ip a 可以查看到 vip 192.168.111.60 在 192.168.111.53(master)上

③ 启动111.52上面的mysql

点击(此处)折叠或打开

root@192.168.111.52:keepalived# mysqld_safe & ;tailf /var/log/messages

Apr 29 14:24:21 localhost Keepalived_vrrp[24862]: VRRP_Script(chk_mysql) succeeded

Apr 29 14:24:22 localhost Keepalived_vrrp[24862]: VRRP_Instance(VI_1) forcing a new MASTER election

Apr 29 14:24:23 localhost Keepalived_vrrp[24862]: VRRP_Instance(VI_1) Transition to MASTER STATE

Apr 29 14:24:24 localhost Keepalived_vrrp[24862]: VRRP_Instance(VI_1) Entering MASTER STATE

root@192.168.111.53:keepalived# tailf /var/log/messages

Apr 29 14:24:45 localhost Keepalived_vrrp[566]: VRRP_Instance(VI_1) Received higher prio advert

Apr 29 14:24:45 localhost Keepalived_vrrp[566]: VRRP_Instance(VI_1) Entering BACKUP STATE

ip a 可以查看到 vip 192.168.111.60 在 192.168.111.52(master)上，也就是说111.52会持续通过track_script的脚本检查，若成功的话会恢复原来的优先级100 ，便把vip抢过来了(若不希望优先级高的直接上来直接抢占vip的话需要再instance 里面配置 nopreempt,backup无需设置)

虚拟server
以上是实验是通过 vrrp_script以及trace_script 实现优先级变换来实现故障转移的，现在看下通过虚拟server怎么实现mysql的高可用
① 配置

点击(此处)折叠或打开

root@192.168.111.52:keepalived# cat /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {

notification_email {

tab@taomee.com

}

notification_email_from dba@taomee.com

smtp_server smtp.shidc.taomee.com

smtp_connect_timeout 30

router_id haMySQL

}

vrrp_script chk_mysql {

script "/etc/keepalived/chk_mysql.sh"

interval 1

weight 2

}

vrrp_instance VI_1 {

state MASTER

interface eth2

virtual_router_id 51

priority 100

nopreempt

advert_int 1

authentication {

auth_type PASS

auth_pass 7777

}

virtual_ipaddress {

192.168.111.60

}

#track_script {

#chk_mysql

}

virtual_server 192.168.111.60 3306 {

delay_loop 6

persistence_timeout 300

protocol TCP

real_server 192.168.111.52 3306 {

weight 1

notify_down /etc/keepalived/kill_self.sh

TCP_CHECK {

tcp_port 3306

connect_timeout 3

}

root@192.168.111.53:keepalived# cat /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {

notification_email {

tab@taomee.com

}

notification_email_from dba@taomee.com

smtp_server smtp.shidc.taomee.com

smtp_connect_timeout 30

router_id haMySQL

}

vrrp_script chk_mysql {

script "/etc/keepalived/chk_mysql.sh"

interval 1

weight 2

}

vrrp_instance VI_1 {

state BACKUP

interface eth2

virtual_router_id 51

priority 99

advert_int 1

authentication {

auth_type PASS

auth_pass 7777

}

virtual_ipaddress {

192.168.111.60

}

#track_script {#########这里先注释掉通过追踪脚本的检查

#chk_mysql

}

virtual_server 192.168.111.60 3306 {

delay_loop 6

persistence_timeout 300

protocol TCP

real_server 192.168.111.53 3306 {### 真实服务

weight 1#### 权重，用来多真实服务均衡使用

notify_down /etc/keepalived/kill_self.sh####在检查该服务不可用时执行该脚本(用来杀死 keepalived 实现 vip 飘逸)

TCP_CHECK {

tcp_port 3306 #### 检查端口继承real_server192.168.111.53 3306 {###真实服务

connect_timeout 3 #### tcp超时时间

}

root@192.168.111.53:keepalived#cat/etc/keepalived/kill_self.sh

#!/bin/bash

killall keepalived

②启动keepalived

点击(此处)折叠或打开

root@192.168.111.52:keepalived# keepalived

root@192.168.111.52:keepalived# tailf /var/log/messages

Apr 29 14:48:22 localhost Keepalived_vrrp[20482]: Opening file '/etc/keepalived/keepalived.conf'.

Apr 29 14:48:22 localhost Keepalived_vrrp[20482]: Configuration is using : 64590 Bytes

Apr 29 14:48:22 localhost Keepalived_vrrp[20482]: Using LinkWatch kernel netlink reflector...

Apr 29 14:48:22 localhost Keepalived_healthcheckers[20481]: IPVS: Scheduler not found

Apr 29 14:48:22 localhost kernel: IPVS: Scheduler module ip_vs_ not found

Apr 29 14:48:22 localhost Keepalived_healthcheckers[20481]: IPVS: Service not defined

Apr 29 14:48:22 localhost Keepalived_healthcheckers[20481]: Using LinkWatch kernel netlink reflector...

Apr 29 14:48:22 localhost Keepalived_healthcheckers[20481]: Activating healthchecker for service [192.168.111.52]:3306

Apr 29 14:48:23 localhost Keepalived_vrrp[20482]: VRRP_Instance(VI_1) Transition to MASTER STATE

Apr 29 14:48:24 localhost Keepalived_vrrp[20482]: VRRP_Instance(VI_1) Entering MASTER STATE

root@192.168.111.53:keepalived# keepalived

root@192.168.111.53:keepalived# tailf /var/log/messages

Apr 29 14:48:51 localhost Keepalived_healthcheckers[25092]: Configuration is using : 11673 Bytes

Apr 29 14:48:51 localhost Keepalived_vrrp[25093]: Opening file '/etc/keepalived/keepalived.conf'.

Apr 29 14:48:51 localhost Keepalived_vrrp[25093]: Configuration is using : 64568 Bytes

Apr 29 14:48:51 localhost Keepalived_vrrp[25093]: Using LinkWatch kernel netlink reflector...

Apr 29 14:48:51 localhost Keepalived_vrrp[25093]: VRRP_Instance(VI_1) Entering BACKUP STATE

Apr 29 14:48:51 localhost kernel: IPVS: Scheduler module ip_vs_ not found

Apr 29 14:48:51 localhost Keepalived_healthcheckers[25092]: IPVS: Scheduler not found

Apr 29 14:48:51 localhost Keepalived_healthcheckers[25092]: IPVS: Service not defined

Apr 29 14:48:51 localhost Keepalived_healthcheckers[25092]: Using LinkWatch kernel netlink reflector...

Apr 29 14:48:51 localhost Keepalived_healthcheckers[25092]: Activating healthchecker for service [192.168.111.53]:3306

此时 ip a 命令可以查看虚拟ip 111.60 在111.52(master) 上

③关闭 111.52上的mysql

点击(此处)折叠或打开

root@192.168.111.52:keepalived# mysqladmin shutdown

2017-04-29T07:07:38.121123Z mysqld_safe mysqld from pid file /opt/mysql/mysqld.pid ended

[1]+ Donemysqld_safe

root@192.168.111.52:keepalived# tailf /var/log/messages

Apr 29 15:07:31 localhost Keepalived_healthcheckers[23405]: TCP connection to [192.168.111.52]:3306 failed !!!

Apr 29 15:07:31 localhost Keepalived_healthcheckers[23405]: Removing service [192.168.111.52]:3306 from VS [192.168.111.60]:3306

Apr 29 15:07:31 localhost Keepalived_healthcheckers[23405]: IPVS: Service not defined

Apr 29 15:07:31 localhost Keepalived_healthcheckers[23405]: Executing [/etc/keepalived/kill_self.sh] for service [192.168.111.52]:3306 in VS [192.168.111.60]:3306

Apr 29 15:07:31 localhost Keepalived_healthcheckers[23405]: Lost quorum 1-0=1 > 0 for VS [192.168.111.60]:3306

Apr 29 15:07:31 localhost Keepalived_healthcheckers[23405]: Remote SMTP server [0.0.0.0]:25 connected.

Apr 29 15:07:31 localhost Keepalived[23404]: Stopping Keepalived v1.2.7 (02/21,2013)

Apr 29 15:07:31 localhost Keepalived_healthcheckers[23405]: IPVS: No such service

Apr 29 15:07:31 localhost Keepalived_vrrp[23406]: VRRP_Instance(VI_1) sending 0 priority

root@192.168.111.53:keepalived# tailf /var/log/messages

Apr 29 15:07:32 localhost Keepalived_vrrp[26815]: VRRP_Instance(VI_1) Transition to MASTER STATE

Apr 29 15:07:33 localhost Keepalived_vrrp[26815]: VRRP_Instance(VI_1) Entering MASTER STATE

此时 ip a 命令可以看到虚拟ip 111.60在111.53(新master)上