我们网站的XX服务随着业务的增长我们需要配置跨IDC容灾情况，为了处理突发事件，避免影响网站的正常访问。自己在网上看了很多处理方法，结合自己的实际情况，我自己也在在局域网的测试服务器上演练了一把。

测试结果表现良好，现在把我在局域网中的部署整理如下：

思路：
当服务器A 发生故障，服务器B可以迅速接管服务器A的任务，不影响用户的正常访问。
当服务器A 故障恢复，服务器A可以马上接管服务器B的任务，服务器B恢复备机状态。

做法：
两台服务器上配置一个虚拟IP地址，主服务器先绑定虚拟ip地址，当发生故障时，备机自动接管虚拟ip地址，刷新网关路由地址。当主机恢复后，备机释放虚拟ip地址，主机再次主动接管虚拟ip地址，刷新网关路由地址。

结构图：


主服务器：ip 192.168.190.199
备服务器: ip 192.168.190.208

vip：192.168.190.88
gateWay=192.168.190.254
netMask=255.255.255.0
bcast=192.168.190.255

步骤：
1、配置虚拟ip：192.168.190.88 到 www.test.com 绑定虚拟ip

/sbin/ifconfig eth0:1 192.168.190.88 broadcast 192.168.190.255  netmask 255.255.255.0  up/sbin/route add -host 192.168.190.88 dev eth0:1

/sbin/ifconfig eth0:1 192.168.190.88 broadcast 192.168.190.255  netmask 255.255.255.0  up/sbin/route add -host 192.168.190.88 dev eth0:1

2、刷新网关路由

/sbin/arping -i eth0 -s 192.168.190.88  192.168.190.254 > /dev/null 2>&1

/sbin/arping -i eth0 -s 192.168.190.88  192.168.190.254 > /dev/null 2>&1

3、当发生故障时，主机192.168.190.199 ，释放虚拟ip192.168.190.88，备机192.168.190.208接管 虚拟ip192.168.190.88

/sbin/ifconfig eth0:1 192.168.190.88  broadcast  192.168.190.255   netmask 255.255.255.0 down        /sbin/arping -i eth0 -s 192.168.190.88  192.168.190.254 > /dev/null 2>&1

/sbin/ifconfig eth0:1 192.168.190.88  broadcast  192.168.190.255   netmask 255.255.255.0 down        /sbin/arping -i eth0 -s 192.168.190.88  192.168.190.254 > /dev/null 2>&1

3、这时 www.test.com 解析到了备机服务器192.168.190.208

4、重启备机web服务器

5、若主机192.168.190.199服务恢复正常，备机192.168.190.208释放虚拟IP,主机
绑定虚拟IP 192.168.190.88  
备机释放虚拟ip：

/sbin/ifconfig eth0:1 192.168.190.88  broadcast  192.168.190.255   netmask 255.255.255.0 down        /sbin/arping -i eth0 -s 192.168.190.88  192.168.190.254 > /dev/null 2>&1

/sbin/ifconfig eth0:1 192.168.190.88  broadcast  192.168.190.255   netmask 255.255.255.0 down        /sbin/arping -i eth0 -s 192.168.190.88  192.168.190.254 > /dev/null 2>&1

6、重启启动主备机web服务

主机192.168.190.199
autoSwitchMain.sh切换脚本

#!/bin/sh##############################################################desc:服务器宕机自动切换服务#author:gaozhonghui#mail:toptreegzh@163.com#date:20121101#############################################################vip=192.168.190.88gateWay=192.168.190.254netMask=255.255.255.0bcast=192.168.190.255function_bind_vip1(){/sbin/ifconfig eth0:1 ${vip} broadcast ${bcast}  netmask ${netMask} up/sbin/route add -host ${vip} dev eth0:1}function_remove_vip1(){/sbin/ifconfig eth0:1 ${vip} broadcast ${bcast}  netmask ${netMask down}function_vip_arping1(){/sbin/arping -i eth0 -s ${vip} ${gateWay} > /dev/null 2>&1}function_restart_nginx(){/web/webserver/nginx/sbin/nginx -s reload}bind_time_vip="N"while truedo httpCode_rip1=`/usr/bin/curl -o /dev/null -s -w %{http_code} http://192.168.190.199`  if [ x${httpCode_rip1} == "x200" ]; then  if [ ${bind_time_vip} == "N" ];  then   function_bind_vip1   function_vip_arping1   bind_time_vip="Y"  fi  function_vip_arping1 else  if [ ${bind_time_vip} == "Y" ]  then   function_remove_vip1   bind_time_vip="N"  fi fi sleep 10done

#!/bin/sh##############################################################desc:服务器宕机自动切换服务#author:gaozhonghui#mail:toptreegzh@163.com#date:20121101#############################################################vip=192.168.190.88gateWay=192.168.190.254netMask=255.255.255.0bcast=192.168.190.255function_bind_vip1(){/sbin/ifconfig eth0:1 ${vip} broadcast ${bcast}  netmask ${netMask} up/sbin/route add -host ${vip} dev eth0:1}function_remove_vip1(){/sbin/ifconfig eth0:1 ${vip} broadcast ${bcast}  netmask ${netMask down}function_vip_arping1(){/sbin/arping -i eth0 -s ${vip} ${gateWay} > /dev/null 2>&1}function_restart_nginx(){/web/webserver/nginx/sbin/nginx -s reload}bind_time_vip="N"while truedohttpCode_rip1=`/usr/bin/curl -o /dev/null -s -w %{http_code} http://192.168.190.199`if [ x${httpCode_rip1} == "x200" ];thenif [ ${bind_time_vip} == "N" ];thenfunction_bind_vip1function_vip_arping1bind_time_vip="Y"fifunction_vip_arping1elseif [ ${bind_time_vip} == "Y" ]thenfunction_remove_vip1bind_time_vip="N"fifisleep 10done

然后linux 启动守候进程
/usr/bin/nohup /bin/sh  /home/Gzh/shell/ autoSwitchMain.sh   2>&1 > /dev/null &

备机192.168.190.208:
autoSwitchSlave.sh

#!/bin/sh##############################################################desc:服务器宕机自动切换服务#author:gaozhonghui#mail:toptreegzh@163.com#date:20121101#############################################################vip=192.168.190.88gateWay=192.168.190.254netMask=255.255.255.0bcast=192.168.190.255function_bind_vip1(){/sbin/ifconfig eth0:1 ${vip} broadcast ${bcast}  netmask ${netMask} up/sbin/route add -host ${vip} dev eth0:1}function_remove_vip1(){/sbin/ifconfig eth0:1 ${vip} broadcast ${bcast}  netmask ${netMask} down}function_vip_arping1(){/sbin/arping -i eth0 -s ${vip} ${gateWay} > /dev/null 2>&1}function_restart_nginx(){/web/webserver/nginx/sbin/nginx -s reload}bind_time_vip="N"while truedo        httpCode_rip1=`/usr/bin/curl -o /dev/null -s -w %{http_code} http://192.                                                168.190.199`        if [ x${httpCode_rip1} == "x200" ];        then                if [ ${bind_time_vip} == "Y" ];                then                        function_remove_vip1                        bind_time_vip="N"                fi                function_vip_arping1        else                if [ ${bind_time_vip} == "N" ]                then                        function_bind_vip1                        function_vip_arping1                        bind_time_vip="Y"                fi        fi        sleep 10done

#!/bin/sh##############################################################desc:服务器宕机自动切换服务#author:gaozhonghui#mail:toptreegzh@163.com#date:20121101#############################################################vip=192.168.190.88gateWay=192.168.190.254netMask=255.255.255.0bcast=192.168.190.255function_bind_vip1(){/sbin/ifconfig eth0:1 ${vip} broadcast ${bcast}  netmask ${netMask} up/sbin/route add -host ${vip} dev eth0:1}function_remove_vip1(){/sbin/ifconfig eth0:1 ${vip} broadcast ${bcast}  netmask ${netMask} down}function_vip_arping1(){/sbin/arping -i eth0 -s ${vip} ${gateWay} > /dev/null 2>&1}function_restart_nginx(){/web/webserver/nginx/sbin/nginx -s reload}bind_time_vip="N"while truedo        httpCode_rip1=`/usr/bin/curl -o /dev/null -s -w %{http_code} http://192.                                                168.190.199`        if [ x${httpCode_rip1} == "x200" ];        then                if [ ${bind_time_vip} == "Y" ];                then                        function_remove_vip1                        bind_time_vip="N"                fi                function_vip_arping1        else                if [ ${bind_time_vip} == "N" ]                then                        function_bind_vip1                        function_vip_arping1                        bind_time_vip="Y"                fi        fi        sleep 10done

启动守候进程
/usr/bin/nohup /bin/sh  /home/Gzh/shell/autoSwitchSlave.sh   2>&1 > /dev/null &

参考资料：
http://blog.s135.com/post/379/

http://toptree.iteye.com/blog/1735324