Haproxy+keepalived高可用集群，允许keepalived宕机，允许后端真实服务器宕机，但是不允许haproxy宕机，

所以下面就是解决方案

keepalived配置高可用检测脚本 ，master和backup都要添加

配置脚本

# vim /etc/keepalived/check_haproxy.sh

#!/bin/bash

echo "check haproxy"

HA= ` ps -C haproxy --no-header |wc -l `

if [ $HA -eq 0 ];then

#判断haproxy的进程是存储，0无haproxy进程，则haproxy宕机

systemctl stop keepalived

fi

并赋予它执行权限

[root@master ~]# chmod +x /etc/keepalived/check_haproxy.sh

修改主备keepalived配置（主备都要修改)  master和backup

重启服务

[root@master ~]# systemctl restart haproxy.service

[root@backup ~]# systemctl restart haproxy.service

测试模拟haproxy宕机，测试是否可以高可用

master ~]# systemctl stop haproxy

master ~]# systemctl status keepalived

keepalived检测脚本，检测到haproxy宕机，直接将keepalived服务停止

查看vip是否漂移到备用节点

[root@backup keepalived]# ip a

for ((i=1;i<=10;i++));do curl 192.168.10.100;done

服务依然可用

测试haproxy的web_ui ，依然可用

启动主服务的haproxy和keepalived

systemctl start haproxy keepalived

systemctl status haproxy keepalived

vip节点又漂移回来了