环境：
操作系统Centos 6.5 X86_64(final)
Nginx-Master:192.168.2.32
Nginx-Backup:192.168.3.31
VIP:192.168.2.33
Web3:192.168.2.29
Web4:192.168.2.30
1、分别在Nginx-Master和Nginx-Backup上安装nginx
[root@Nginx-Master ~]# rpm -i http://nginx.org/packages/centos/6/noarch/RPMS/nginx-release-centos-6-0.el6.ngx.noarch.rpm
[root@Nginx-Backup ~]# rpm -i http://nginx.org/packages/centos/6/noarch/RPMS/nginx-release-centos-6-0.el6.ngx.noarch.rpm

[root@Nginx-Master ~]# yum install nginx

[root@Nginx-Backup ~]# yum install nginx

2、先配置Nginx-Master上的nginx，建立配置一个proxy.conf文件。并复制一份到Nginx-Backup上，删除两台机器上默认的default.conf,启动nginx服务；并将服务添加到开机启动。

[root@Nginx-Master ~]# vi /etc/nginx/conf.d/proxy.conf
    upstream web {
    ＃ip_hash;
        server 192.168.2.29:80;   #默认为rr轮询，如需解决session的问题采有哈希（ip_hash)模块。
        server 192.168.2.30:80;
}
        server {
        listen 80;
        index index.php index.html index.htm;
        location / {
        proxy_pass http://web;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
}
}
[root@Nginx-Master ~]# scp /etc/nginx/conf.d/proxy.conf root@192.168.2.31:/etc/nginx/conf.d/

[root@Nginx-Master ~]# service nginx start

Starting nginx:                                    [  OK  ]
[root@Nginx-Backup ~]# service nginx start
Starting nginx:                                    [  OK  ]
[root@Nginx-Master ~]# chkconfig nginx on
[root@Nginx-Backup ~]# chkconfig nginx on
3、分别在Nginx-Master和Nginx-Backup上安装keepalived并配置。
[root@Nginx-Master ~]# yum install keepalived
[root@Nginx-Backup ~]# yum install keepalived
[root@Nginx-Master ~]# vi /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 127.0.0.1
   smtp_connect_timeout 30
   router_id LVS_DEVEL
}
vrrp_script chk_nginx {                         #监控nginx服务进程脚本
    script "/root/nginx.sh"          
    interval 2
    weight 2
}
vrrp_instance VI_1 {
    state MASTER
    interface eth0
    virtual_router_id 51
    mcast_src_ip 192.168.2.32
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    track_script {                                    #调用监控脚本
        chk_nginx
}
    virtual_ipaddress {
        192.168.2.33
    }
}

[root@Nginx-Backup ~]# vi /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 127.0.0.1
   smtp_connect_timeout 30
   router_id LVS_DEVEL
}
vrrp_script chk_nginx {                        #监控nginx服务进程脚本
    script "/root/nginx.sh"       
    interval 2
    weight 2
}
vrrp_instance VI_1 {
    state BACKUP
    interface eth0
    virtual_router_id 51
    priority 90
    mcast_src_ip 192.168.2.31
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    track_script {                                  #调用监控脚本
        chk_nginx
    virtual_ipaddress {
        192.168.2.33
    }
}
[root@Nginx-Master ~]# service keepalived start
Starting keepalived:                                   [  OK  ]

[root@Nginx-Backup ~]# service keepalived start

Starting keepalived:                                   [  OK  ]

3.1.对keepalived的不足写的一个脚本，用来检测本机的nginx是否正常的运行，如果nginx挂掉试着重新启动，如果启动后又挂掉，那么就直接停止keepalived进程，keepalived将转移到另一台备用上，实现故障转移

[root@Nginx-Master ~]# vi /root/nginx.sh
#!/bin/bash
A=`ps -C nginx --no-header |wc -l`                      # 查看是否有 nginx进程 把值赋给变量A
if [ $A -eq 0 ];then                                                #如果没有进程值为零
                /usr/sbin/nginx
                sleep 1
                if [ `ps -C nginx --no-header |wc -l` -eq 0 ];then
                       /etc/init.d/keepalived stop             # 则停止keepalived 进程
                fi
fi

[root@Nginx-Master ~]# chmod 755 /root/nginx.sh

[root@Nginx-Master ~]# scp /root/nginx.sh root@192.168.2.31:/root/nginx.sh

3.2.重启keepalived服务，查看日志脚本是否正常

Nginx-Master日志
[root@Nginx-Master ~]# tail -f /var/log/messages
Sep  5 15:20:48 Nginx-Master Keepalived_vrrp[5136]: VRRP sockpool: [ifindex(2), proto(112), fd(10,11)]
Sep  5 15:20:48 Nginx-Master Keepalived_vrrp[5136]: VRRP_Script(chk_nginx) succeeded                                                                     ＃检查脚本成功，表示正常
Sep  5 15:20:48 Nginx-Master Keepalived_vrrp[5136]: VRRP_Instance(VI_1) Transition to MASTER STATE
Sep  5 15:20:49 Nginx-Master Keepalived_vrrp[5136]: VRRP_Instance(VI_1) Entering MASTER STATE
Sep  5 15:20:49 Nginx-Master Keepalived_vrrp[5136]: VRRP_Instance(VI_1) setting protocol VIPs.

Nginx-Backup日志

[root@Nginx-Backup ~]# tail -f /var/log/messages
Sep  5 15:20:41 Nginx-Backup Keepalived_vrrp[7670]: Using LinkWatch kernel netlink reflector...
Sep  5 15:20:41 Nginx-Backup Keepalived_vrrp[7670]: VRRP_Instance(VI_1) Entering BACKUP STATE
Sep  5 15:20:41 Nginx-Backup Keepalived_vrrp[7670]: VRRP sockpool: [ifindex(2), proto(112), fd(10,11)]
Sep  5 15:20:41 Nginx-Backup Keepalived_vrrp[7670]: VRRP_Script(chk_nginx) succeeded                                                                     #检查脚本成功，表示正常

[root@Nginx-Backup ~]# tail -f /var/log/messages

Sep  5 15:10:39 Nginx-Backup Keepalived_vrrp[5265]: Process [5913] didn't respond to SIGTERM
Sep  5 15:10:41 Nginx-Backup Keepalived_vrrp[5265]: Process [5923] didn't respond to SIGTERM                                                                        #如出现这种，看下脚本是否正确。
Sep  5 15:10:43 Nginx-Backup Keepalived_vrrp[5265]: Process [5933] didn't respond to SIGTERM

3.3.另外一种查看脚本是否生效方法,停止nginx服务，马上会自动启动   

[root@Nginx-Master ~]# service nginx status

nginx (pid  1114) is running...
[root@Nginx-Master ~]# service nginx stop
Stopping nginx:                                            [  OK  ]
[root@Nginx-Master ~]# service nginx status
nginx (pid  28751) is running...

4、测试VIP，检查是否能主备切换

4.1.当两台主机同时启动时，只有Nginx-Master服务器拥有VIP地址，备服务器没有。

[root@Nginx-Master ~]# ip addr
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:50:56:a6:00:13 brd ff:ff:ff:ff:ff:ff
    inet 192.168.2.32/24 brd 192.168.2.255 scope global eth0
    inet 192.168.2.33/32 scope global eth0                        #可以看到Nginx-Master服务器上拥有192.168.2.33这个VIP地址
    inet6 fe80::250:56ff:fea6:13/64 scope link
       valid_lft forever preferred_lft forever

[root@Nginx-Backup ~]# ip addr

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
    link/ether 08:00:27:77:d3:82 brd ff:ff:ff:ff:ff:ff
    inet 192.168.2.31/24 brd 192.168.2.255 scope global eth0
    inet6 fe80::a00:27ff:fe77:d382/64 scope link                #Nginx-Backup服务器上没有  
       valid_lft forever preferred_lft forever

4.2.当停止Nginx-Master服务器的keepalived服务，再查看下两台主机的VIP地址，发现VIP地址已从主服务器转移到了备服务器

[root@Nginx-Master ~]# service keepalived stop

Stopping keepalived:                  [  OK  ]

[root@Nginx-Master ~]# ip addr

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:50:56:a6:00:13 brd ff:ff:ff:ff:ff:ff
    inet 192.168.2.32/24 brd 192.168.2.255 scope global eth0
    inet6 fe80::250:56ff:fea6:13/64 scope link                #Nginx-Master服务器VIP地址已移除      
       valid_lft forever preferred_lft forever

[root@Nginx-Backup ~]# ip addr

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
    link/ether 08:00:27:77:d3:82 brd ff:ff:ff:ff:ff:ff
    inet 192.168.2.31/24 brd 192.168.2.255 scope global eth0
    inet 192.168.2.33/32 scope global eth0         #VIP地地已自动转移到Nginx-Backup服务器上            
    inet6 fe80::a00:27ff:fe77:d382/64 scope link
       valid_lft forever preferred_lft forever

5、日志查看主备切换过程

5.1.当停止Nginx-Master上的keepalived服务时

root@Nginx-Master ~]# service keepalived stop
Nginx-Master日志
[root@Nginx-Master ~]# tail -f /var/log/messages
Sep  4 18:04:06 Nginx-Master Keepalived[3278]: Stopping Keepalived v1.2.7 (02/21,2013)                                                                                  ＃主服务器已停掉
Sep  4 18:04:06 Nginx-Master Keepalived_vrrp[3281]: VRRP_Instance(VI_1) sending 0 priority    
Sep  4 18:04:06 Nginx-Master Keepalived_vrrp[3281]: VRRP_Instance(VI_1) removing protocol VIPs.

Nginx-Backup日志

[root@Nginx-Backup ~]# tail -f /var/log/messages
Sep  4 18:04:07 Nginx-Backup Keepalived_vrrp[1428]: VRRP_Instance(VI_1) Transition to MASTER STATE
Sep  4 18:04:08 Nginx-Backup Keepalived_vrrp[1428]: VRRP_Instance(VI_1) Entering MASTER STATE                                                                  #Nginx-Backup转为MASTER STATE
Sep  4 18:04:08 Nginx-Backup Keepalived_vrrp[1428]: VRRP_Instance(VI_1) setting protocol VIPs.
Sep  4 18:04:08 Nginx-Backup Keepalived_healthcheckers[1427]: Netlink reflector reports IP 192.168.2.33 added
Sep  4 18:04:08 Nginx-Backup Keepalived_vrrp[1428]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.2.33
Sep  4 18:04:13 Nginx-Backup Keepalived_vrrp[1428]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.2.33

5.2.当Nginx-Master的keepalived服务再次启动时

[root@Nginx-Master ~]# service keepalived restart
Stopping keepalived:                     [  OK  ]
Starting keepalived:                     [  OK  ]
Nginx-Master日志
[root@Nginx-Master ~]# tail -f /var/log/messages
Sep  4 18:06:47 Nginx-Master Keepalived_vrrp[3316]: VRRP_Instance(VI_1) Transition to MASTER STATE
Sep  4 18:06:48 Nginx-Master Keepalived_vrrp[3316]: VRRP_Instance(VI_1) Entering MASTER STATE                                                   #Nginx-Master转回MASTER STATE
Sep  4 18:06:48 Nginx-Master Keepalived_vrrp[3316]: VRRP_Instance(VI_1) setting protocol VIPs.
Sep  4 18:06:48 Nginx-Master Keepalived_healthcheckers[3315]: Netlink reflector reports IP 192.168.2.33 added
Sep  4 18:06:48 Nginx-Master Keepalived_vrrp[3316]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.2.33
Sep  4 18:06:53 Nginx-Master Keepalived_vrrp[3316]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.2.33

Nginx-Backup日志

[root@Nginx-Backup ~]# tail -f /var/log/messages
Sep  4 18:06:47 Nginx-Backup Keepalived_vrrp[1428]: VRRP_Instance(VI_1) Received higher prio advert
Sep  4 18:06:47 Nginx-Backup Keepalived_vrrp[1428]: VRRP_Instance(VI_1) Entering BACKUP STATE                                                   #Nginx-Backup转回BACKUP STATE
Sep  4 18:06:47 Nginx-Backup Keepalived_vrrp[1428]: VRRP_Instance(VI_1) removing protocol VIPs.
Sep  4 18:06:47 Nginx-Backup Keepalived_healthcheckers[1427]: Netlink reflector reports IP 192.168.2.33 removed

6、安装后端两台httpd服务器，并添加内容测试

[root@web3 ~]# yum install httpd

[root@web3 ~]# vi /var/www/html/index.html

<h1>Welcome to web3(192.168.2.29)</h1>

[root@web3 ~]# service httpd start

[root@web4 ~]# yum install httpd

[root@web4 ~]# vi /var/www/html/index.html
<h1>Welcome to web4(192.168.2.30)</h1>

[root@web4 ~]# service httpd start

7.查看后端web服务器日志

7.1.后端web服务器日志，访问IP还是代理服务器的，不是真实客户端IP

[root@web3 ~]# tail -f /var/log/httpd/access_log
192.168.2.31 - - [05/Sep/2014:15:47:46 +0800] "GET / HTTP/1.0" 200 39 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.3.1.2000 Chrome/30.0.1599.101 Safari/537.36"
192.168.2.32 - - [05/Sep/2014:15:58:02 +0800] "GET / HTTP/1.0" 200 39 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.3.1.2000 Chrome/30.0.1599.101 Safari/537.36"

[root@web4 ~]# tail -f /var/log/httpd/access_log

192.168.2.31 - - [05/Sep/2014:15:48:47 +0800] "GET / HTTP/1.0" 200 39 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.3.1.2000 Chrome/30.0.1599.101 Safari/537.36"
192.168.2.32 - - [05/Sep/2014:15:59:55 +0800] "GET / HTTP/1.0" 200 39 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.3.1.2000 Chrome/30.0.1599.101 Safari/537.36"

7.2.虽然在配置proxy.conf时已配置了proxy_set_header X-Real-IP $remote_addr；但还需要在web端修改配围起文件显示

[root@web3 ~]# vi /etc/httpd/conf/httpd.conf
LogFormat "%h %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"" combined
改为:
LogFormat "%{X-Real-IP}i %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"" combined
web4也同样修改以上数据后重启

[root@web3 ~]# service httpd restart

[root@web4 ~]# service httpd restart

7.3.最后再查看后端web服务器日志为真实客户端IP

[root@web3 ~]# tail -f /var/log/httpd/access_log

192.168.2.200 - - [05/Sep/2014:16:58:36 +0800] "GET / HTTP/1.0" 200 39 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.3.1.2000 Chrome/30.0.1599.101 Safari/537.36"

[root@web4 ~]# tail -f /var/log/httpd/access_log

192.168.2.200 - - [05/Sep/2014:16:58:42 +0800] "GET / HTTP/1.0" 200 39 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.3.1.2000 Chrome/30.0.1599.101 Safari/537.36"

最后测试主keepalived挂掉，会转移到备用keepalived上，nginx负载均衡服务，keepalived会一直调用脚本检查机制，如果nginx服务挂掉或不能从新启动，都会停止掉keepalived并立即转移到备用上继续工作，后端web服务器出现问题，nginx负载均衡能自动切换后端有故障的web服务器。

参考：http://network.51cto.com/art/201007/209823.htm

          http://friendlinux.blog.51cto.com/6249249/1433295