keepalived检测Nginx高可用

node1 192.168.136.55

node2 192.168.136.56

两台机器都安装dnf install keepalived nginx

[root@node1 ~]# echo "web test page, ip is `hostname -I`." > /usr/share/nginx/html/index.html 
[root@node2 ~]# echo "web test page, ip is `hostname -I`." > /usr/share/nginx/html/index.html

主的配置

[root@node1 ~]# vim /etc/keepalived/keepalived.conf 
global_defs {router_id Nginx1
}vrrp_instance Nginx {state MASTERinterface ens160virtual_router_id 51priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.136.100}
}

从的配置

[root@node2 ~]# vim /etc/keepalived/keepalived.conf 
global_defs {router_id Nginx2
}vrrp_instance Nginx {state BACKUPinterface ens160virtual_router_id 51priority 80advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.136.100}
}

[root@node1 ~]# systemctl start keepalived.service 
[root@node2 ~]# systemctl start keepalived.service

测试访问虚拟IP 访问的是master

当master的keepalived停掉

访问的是backup

健康检查Nginx的状态

目前 Keepalived 仅能通过主机是否宕机来进行业务切换，如果仅 Nginx 业务出现故障是无法切换的，需在 Keepalive 中添加健康检查来达到检测 Nginx 是否可用的目的。

[root@openEuler-node1 ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {router_id Nginx1
}vrrp_script chk_ngx{script "killall -0 nginx"interval 1
}vrrp_instance Nginx {state MASTERinterface ens160virtual_router_id 51priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}track_script {chk_ngx}virtual_ipaddress {192.168.136.100}
}

node2上修改类似 killall -0 nginx 一刀切直接切换

[root@openEuler-node2 ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalivedglobal_defs {router_id Nginx2
}vrrp_script chk_ngx{script "killall -0 nginx"interval 1
}vrrp_instance Nginx {state BACKUPinterface ens160virtual_router_id 51priority 80advert_int 1authentication {auth_type PASSauth_pass 1111}track_script {chk_ngx}virtual_ipaddress {192.168.136.100}
}

当MATSER的nginx停掉之后

当检测到nginx异常，然后停止keepalive，VIP漂移

1、检测到web服务异常

2、重启web

3、启动不了，再VIP漂移

更改配置文件

node1 和Node2一样

vim /etc/keepalived/chk_ngx.sh
#!/bin.bash
systemctl status nginx | grep "active (running)" > /dev/null
if [ $? -ne 0 ]; thensystemctl restart nginx &> /dev/nullsleep 1systemctl status nginx | grep "active (running)" > /dev/nullif [ $? -ne 0 ]; thensystemctl stop keepalivedelseexitfi
fi
[root@openEuler-node1 ~]# chmod +x /etc/keepalived/chk_ngx.sh

检验1 将node1上nginx stop了