Ubuntu14.04 + Keepalived1.2.13

172.18.8.74 (master，vip:172.18.8.100)
172.18.8.73 (slave,vip:172.18.8.100)

通过keepalived 实现redis主从备份零切换

keepalived.conf配置文件，如下：

! Configuration File for keepalived



global_defs {

   router_id ubuntu-lvs-74

}



vrrp_script chk_redis {

    script "/etc/keepalived/scripts/redis_check.sh"

    interval 2

}



vrrp_instance VI_1 {

    state MASTER

    interface eth0

    virtual_router_id 51

    priority 101 #  备机 优先级为100



    authentication {

        auth_type PASS

        auth_pass redis

    }



    track_script {

       chk_redis

    }



    virtual_ipaddress {

       172.18.8.100/32

    }



    notify_master /etc/keepalived/scripts/redis_master.sh

    notify_backup /etc/keepalived/scripts/redis_backup.sh

    notify_fault  /etc/keepalived/scripts/redis_fault.sh

    notify_stop   /etc/keepalived/scripts/redis_stop.sh



}



virtual_server 172.18.8.100 6379 {

    delay_loop 1

    lb_algo rr

    lb_kind DR #NAT

    #nat_mask 255.255.255.255

    persistence_timeout 50

    protocol TCP



    real_server 172.18.8.74 6379 {

        weight 1

        TCP_CHECK {

          connect_timeout 3

          nb_get_retry 3

          delay_before_retry 3

          connect_port 6379

        }

    }



    real_server 172.18.8.73 6379 {

        weight 1

        TCP_CHECK {

          connect_timeout 3

          nb_get_retry 3

          delay_before_retry 3

          connect_port 6379

        }

    }

}

正确的结果应该是：
当前启动主keepalived时候，虚拟主机IP应该被主机监听绑定，备机不绑定虚拟IP；当主机keepalived宕掉时，备机进行绑定VIP

1、为什么我keepalived主备都会监听虚拟IP？
#sudo ip a
都是输出：[code]root@ubuntu-lvs-74:/opt/lvs# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default 
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
   inet 172.18.8.100/32 brd 172.18.8.100 scope global lo:0
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host [/code]/etc/sysctl.conf 设置了：
net.ipv4.ip_nonlocal_bind = 1

VIP设置如下,ifconfig：

root@ubuntu-lvs-74:/opt/lvs# ifconfig

eth0      Link encap:Ethernet  HWaddr 52:54:00:ee:21:49  

          inet addr:172.18.8.74  Bcast:172.18.8.127  Mask:255.255.255.128

          inet6 addr: fe80::5054:ff:feee:2149/64 Scope:Link

          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

          RX packets:3038779 errors:0 dropped:1356 overruns:0 frame:0

          TX packets:564682 errors:0 dropped:0 overruns:0 carrier:0

          collisions:0 txqueuelen:1000 

          RX bytes:155424872 (155.4 MB)  TX bytes:39263783 (39.2 MB)



lo        Link encap:Local Loopback  

          inet addr:127.0.0.1  Mask:255.0.0.0

          inet6 addr: ::1/128 Scope:Host

          UP LOOPBACK RUNNING  MTU:65536  Metric:1

          RX packets:703904 errors:0 dropped:0 overruns:0 frame:0

          TX packets:703904 errors:0 dropped:0 overruns:0 carrier:0

          collisions:0 txqueuelen:0 

          RX bytes:39418624 (39.4 MB)  TX bytes:39418624 (39.4 MB)



lo:0      Link encap:Local Loopback  

          inet addr:172.18.8.100  Mask:255.255.255.255

          UP LOOPBACK RUNNING  MTU:65536  Metric:1

请教，这是哪里出了问题了？

2、keepalived启动后，出现VRRP_Instance(VI_1) Now in FAULT state

Dec  6 21:01:43 ubuntu-lvs-74 Keepalived_healthcheckers[9801]: Netlink reflector reports IP fe80::5054:ff:feee:2149 added

Dec  6 21:01:43 ubuntu-lvs-74 Keepalived_healthcheckers[9801]: Registering Kernel netlink reflector

Dec  6 21:01:43 ubuntu-lvs-74 Keepalived_healthcheckers[9801]: Registering Kernel netlink command channel

Dec  6 21:01:43 ubuntu-lvs-74 Keepalived_healthcheckers[9801]: Opening file '/etc/keepalived/keepalived.conf'.

Dec  6 21:01:43 ubuntu-lvs-74 Keepalived_healthcheckers[9801]: Configuration is using : 13613 Bytes

Dec  6 21:01:43 ubuntu-lvs-74 Keepalived_healthcheckers[9801]: Using LinkWatch kernel netlink reflector...

Dec  6 21:01:43 ubuntu-lvs-74 Keepalived_healthcheckers[9801]: Activating healthchecker for service [172.18.8.74]:6379

Dec  6 21:01:43 ubuntu-lvs-74 Keepalived_healthcheckers[9801]: Activating healthchecker for service [172.18.8.73]:6379

Dec  6 21:01:44 ubuntu-lvs-74 Keepalived_vrrp[9802]: VRRP_Instance(VI_1) Now in FAULT state

Dec  6 21:01:44 ubuntu-lvs-74 Keepalived_vrrp[9802]: Opening script file /etc/keepalived/scripts/redis_fault.sh

route信息
Destination     Gateway         Genmask              Flags  Metric Ref    Use Iface
default            172.18.8.1      0.0.0.0                UG     0      0        0    eth0
172.18.8.0      *                    255.255.255.128 U       0      0        0    eth0
172.18.8.100  *                    255.255.255.255 UH     0      0        0     lo

root@ubuntu-lvs-74:/opt/lvs# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn

TCP  172.18.8.100:80 rr

  -> 172.18.8.73:80               Route   1      0          0         

  -> 172.18.8.74:80               Route   1      0          0         

TCP  172.18.8.100:6379 rr persistent 50

  -> 172.18.8.73:6379             Route   1      0          0         

  -> 172.18.8.74:6379             Route   1      0          0

VIP:172.18.8.100 ，在73,74之间可以互相ping通

请教，VRRP_Instance(VI_1) Now in FAULT state 可能是哪方面原因造成的？