原始出处：http://bbs.chinaunix.net/thread-4174822-1-1.html 

(出处: http://bbs.chinaunix.net/)


keepalived 主（192.168.1.3）

! Configuration File for keepalived

global_defs {
   notification_email {
     #admin@demo.com
   }
   #notification_email_from admin@demo.com
   #smtp_server 127.0.0.1
   #smtp_connect_timeout 30
   router_id LVS_DEVEL
}

vrrp_script chk_nginx {

script "/etc/keepalived/check_nginx.sh"

interval 2

weight 2

}

vrrp_instance VI_1 {
    state MASTER
    interface eth0
    virtual_router_id 51
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.1.2
    }
    track_script {
        chk_nginx
    }
}


keepalived 备（192.168.1.4）

! Configuration File for keepalived

global_defs {
   notification_email {
     #admin@michael-demo.com
   }
   #notification_email_from admin@michael-demo.com
   #smtp_server 127.0.0.1
   #smtp_connect_timeout 30
   router_id LVS_DEVEL
}

vrrp_script chk_nginx {

script "/etc/keepalived/check_nginx.sh"

interval 2

weight 2
}

vrrp_instance VI_1 {
    state       MASTER
    interface eth0
    virtual_router_id 51
    priority 98
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        221.1.220.206
    }
    track_script {
        chk_nginx
    }
}




现在问题是两台机器都有VIP .但是一台重启或宕机。另外一台能接管。备机state改为BACKUP模式也白搭！


keepalived 主日志：

Apr 17 07:30:04 localhost kernel: Bluetooth: RFCOMM socket layer initialized
Apr 17 07:30:04 localhost kernel: Bluetooth: RFCOMM TTY layer initialized
Apr 17 07:30:04 localhost kernel: Bluetooth: RFCOMM ver 1.8
Apr 17 07:30:04 localhost kernel: Bluetooth: HIDP (Human Interface Emulation) ver 1.1
Apr 17 07:36:39 localhost Keepalived_vrrp[2520]: VRRP_Script(chk_nginx) timed out
Apr 17 07:36:39 localhost Keepalived_vrrp[2520]: VRRP_Script(chk_nginx) succeeded
Apr 17 07:36:55 localhost Keepalived_vrrp[2520]: VRRP_Script(chk_nginx) timed out
Apr 17 07:36:55 localhost Keepalived_vrrp[2520]: VRRP_Script(chk_nginx) succeeded
Apr 17 07:54:13 localhost Keepalived_vrrp[2520]: VRRP_Script(chk_nginx) timed out
Apr 17 07:54:13 localhost Keepalived_vrrp[2520]: VRRP_Script(chk_nginx) succeeded


keepalived 备日志：

Apr 16 23:51:29  Keepalived_vrrp[4733]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]
Apr 16 23:51:29 Keepalived_vrrp[4733]: VRRP_Script(chk_nginx) succeeded
Apr 16 23:51:30  Keepalived_vrrp[4733]: VRRP_Instance(VI_1) Transition to MASTER STATE
Apr 16 23:51:31  Keepalived_vrrp[4733]: VRRP_Instance(VI_1) Entering MASTER STATE
Apr 16 23:51:31 Keepalived_vrrp[4733]: VRRP_Instance(VI_1) setting protocol VIPs.
Apr 16 23:51:31  Keepalived_healthcheckers[4732]: Netlink reflector reports IP 192.168.1.2 added
Apr 16 23:51:31 Keepalived_vrrp[4733]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.1.2
Apr 16 23:51:31  avahi-daemon[3246]: Registering new address record for 192.168.1.2 on eth0.
Apr 16 23:51:31  Keepalived_vrrp[4733]: Netlink reflector reports IP 19.168.1.2 added
Apr 16 23:51:36  Keepalived_vrrp[4733]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.1.2


解决方法吧，出现这问题的场景是在阿里VPS云服务器网络环境中，因为路由交换层禁用了ARP的广播限制，造成KEEPALIVE主备协议无法通过广播的方式进行通信，造成主备两台服务器都强占HAVIP地址，出现同时两台服务器都有VIP地址的情况出现，必须通过配置来指定IP的两台服务器间进行通讯（阿里说明文档中解释只能支持两台使用同一个HAVIP地址），基于以下方法可以的情况下，多备方式用同样的方式也应该可行 ，有需要的兄弟可以测试下多IP备的方式（正常情况需要主备一对主备就够了）。
在网卡配置后面需要加上以下配置：

    priority 100  ##主服务100 备服务90只要小于100就可以
    unicast_src_ip  192.168.1.21##（本地IP地址）
    unicast_peer {
                  192.168.1.22##（对端IP地址）此地址一定不能忘记
                       }

（对端IP地址）问题一直没注意。