Web集群案例实战 -- Keepalived 实现 web 服务高可用

前言

本环境是基于 Centos 7.8 系统构建 Keepalived 学习环境
具体构建，请参考 Keepalived 环境部署

环境准备

role	host	ip	keepalive-version	httpd-version
MASTER	node01	192.168.5.11	keepalived-2.0.12	httpd-2.4.6
BACKUP	node02	192.168.5.12	keepalived-2.0.12	httpd-2.4.6
Client	node03	192.168.5.13	----	----

部署 http服务

---node01
[root@node01 ~]# yum install httpd -y
[root@node01 ~]# echo "`hostname -I` httpd test page..." > /var/www/html/index.html
[root@node01 ~]# systemctl start httpd

---node02
[root@node02 ~]# yum install httpd -y
[root@node02 ~]# echo "`hostname -I` httpd test page..." > /var/www/html/index.html
[root@node02 ~]# systemctl start httpd

配置 MASTER

[root@node01 ~]# vim /etc/keepalived/keepalived.conf 
! 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 192.168.5.10
   smtp_connect_timeout 30
   router_id LVS_DEVEL1
}

vrrp_instance VI_1 {
    state MASTER
    interface ens33
    virtual_router_id 51
    priority 200
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.5.20
    }
}

[root@node01 ~]# systemctl restart keepalived.service 

配置 BACKUP

[root@node02 ~]# vim /etc/keepalived/keepalived.conf 
! 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 192.168.5.10
   smtp_connect_timeout 30
   router_id LVS_DEVEL2
}

vrrp_instance VI_1 {
    state BACKUP
    interface ens33
    virtual_router_id 51
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.5.20
    }
}

[root@node02 ~]# systemctl restart keepalived.service 

客户端通过 VIP 访问 web 服务

node01 发生故障后

[root@node01 ~]# systemctl stop keepalived.service 
[root@node01 ~]# ps -ef | grep keepalived
root      32598  31955  0 10:19 pts/0    00:00:00 grep --color=auto keepalived

此时 node02 成为 MASTER VIP 漂移到了node02！！！

当node01 web 服务发生故障后

[root@node01 ~]# systemctl start keepalived.service 
[root@node01 ~]# ps -ef | grep keepalived
root      32611      1  0 10:21 ?        00:00:00 /usr/local/keepalived/sbin/keepalived -D
root      32612  32611  0 10:21 ?        00:00:00 /usr/local/keepalived/sbin/keepalived -D
root      32614  31955  0 10:21 pts/0    00:00:00 grep --color=auto keepalived
[root@node01 ~]# systemctl stop httpd
[root@node01 ~]# ps -ef | grep httpd
root      32625  31955  0 10:21 pts/0    00:00:00 grep --color=auto httpd

此时 VIP 在node01 上，且node01 web 服务发生故障…

主：keepalived 仅在LVS中，对后端服务有健康检查的功能

我们可以通过脚本的方式，keepalive 对web服务进行健康检查！以确保服务的高可用性

---node01
[root@node01 ~]# vim /etc/keepalived/check_web_server_keepalive.sh
#!/bin/bash

while true
do
   http_pid=$(ps -C httpd --no-header | wc -l)
   echo $http_pid
   if [ ${http_pid} -eq 0 ]
   then
     systemctl start httpd
     sleep 4
   fi
   
   http_pid=$(ps -C httpd --no-header | wc -l)
   if [ ${http_pid} -eq 0 ]
   then 
     systemctl stop keepalived
   fi
   sleep 3

done

[root@node01 ~]# chmod +x /etc/keepalived/check_web_server_keepalive.sh
[root@node01 ~]# cd /etc/keepalived/
[root@node01 keepalived]# nohup ./check_web_server_keepalive.sh &


---node02
[root@node02 ~]# vim /etc/keepalived/check_web_server_keepalive.sh
#!/bin/bash

while true
do
   http_pid=$(ps -C httpd --no-header | wc -l)
   echo $http_pid
   if [ ${http_pid} -eq 0 ]
   then
     systemctl start httpd
     sleep 4
   fi
   
   http_pid=$(ps -C httpd --no-header | wc -l)
   if [ ${http_pid} -eq 0 ]
   then 
     systemctl stop keepalived
   fi
   sleep 3

done

[root@node02 ~]# chmod +x /etc/keepalived/check_web_server_keepalive.sh
[root@node02 ~]# cd /etc/keepalived/
[root@node02 keepalived]# nohup ./check_web_server_keepalive.sh &

测试
node02 停止httpd

[root@node01 ~]# systemctl stop httpd

查看node01 web 服务状态

[root@node01 ~]# systemctl is-active httpd
active

node01 web 服务正常响应！！！

停止node01 keepalived 关闭node02 web 服务

[root@node01 ~]# systemctl stop keepalived
[root@node01 ~]# systemctl is-active keepalived
unknown
[root@node02 keepalived]# systemctl stop httpd

node01 web 服务正常响应！！！

总结

---

Keepalive 默认故障切换，为抢占式（优先级大小），为了保证线上业务的稳定，也可以通过nopreempt 项配置非抢占式切换！（但是所有节点服务器服务设置成MASTER）