基于Ldirectord的高可用性群集实现http服务

Ldirectord是个服务用于实现后方真实的服务器的健康状态探测，其配置文件中定义虚拟服务规则、探测页面、哪种服务，当后方的服务器失效，在节点上的虚拟服务规则将会失效。

实验的ip地址规划： 案例ip地址规划 虚拟ip地址为192.168.20.100 节点一的eth0的ip地址为192.168.20.10，eth1ip地址为192.168.30.10 节点二的eth0的ip地址为192.168.20.20，eth1ip地址为192.168.30.20 Realserver1的ip地址为192.168.20.30 Realserver2的ip地址为192.168.20.40 实验的拓扑图：实现http服务

在节点上的相关配置 打开配置文件Ldirectord.cf checkinterval=1 #fallback=127.0.0.1:80 autoreload=yes #logfile="/var/log/ldirectord.log" #logfile="local0" #emailalert="admin@x.y.z" #emailalertfreq=3600 #emailalertstatus=all quiescent=no当检测到后方的服务器失效时，节点上虚拟规则将会失效

#Sampleforanhttpvirtualservice virtual=192.168.20.100:80 real=192.168.20.30:80gate后方真正的服务器 real=192.168.20.40:80gate service=http request=".test.html"#ldirectord发送http请求请求后方服务器的测试页面的内容，.test.html是测试页面 receive="ok"#请求返回的内容 scheduler=rr#虚拟服务的算法为轮训机制 #persistent=600 #netmask=255.255.255.255 protocol=tcp#基于tcp的四层协议 checktype=negotiate checkport=80#检查端口 打开ha.cf这个文件，ha.cf是高可用性群集的配置文件 bcasteth1#Linux广播地址为eth：0这个端口，当主节点失效时通过这个端口把资源（vip、heartbeat服务控制的其他服务、其他服务的配置文件）传到备份节点上。 #bcasteth1eth2#Linux #nodekathy nodenode1.a.com nodenode2.a.com

打开resource文件定义资源（虚拟服务ip端口服务配置文件） #node-nameresource1resource2...resourceN node1.a.com192.168.20.100/24/eth0ldirectord::ldirectord.cf 打开authkeys文件 authkeys auth3#节点之间传递资源时要进行验证，MD5身份验证验证内容是RedHat #1crc #2sha1HI! 3md5redha

[root@node1ha.d]#serviceheartbeatstart logdisalreadyrunning StartingHigh-Availabilityservices: 2013/10/04_18:08:24INFO:RunningOK 2013/10/04_18:08:24CRITICAL:Resource192.168.20.100/24/eth0isactive,andshouldnotbe! 2013/10/04_18:08:24CRITICAL:Non-idleresourcescanaffectdataintegrity!

[root@node1ha.d]#ifconfig eth0Linkencap:EthernetHWaddr00:0C:29:FA:2D:49 inetaddr:192.168.20.10Bcast:192.168.20.255Mask:255.255.255.0 inet6addr:fe80::20c:29ff:fefa:2d49/64Scope:Link UPBROADCASTRUNNINGMULTICASTMTU:1500Metric:1 RXpackets:10176errors:0dropped:0overruns:0frame:0 TXpackets:7179errors:0dropped:0overruns:0carrier:0 collisions:0txqueuelen:1000 RXbytes:3123678(2.9MiB)TXbytes:962204(939.6KiB) Interrupt:19Baseaddress:0x2000

eth0:0Linkencap:EthernetHWaddr00:0C:29:FA:2D:49 inetaddr:192.168.20.100Bcast:192.168.20.255Mask:255.255.255.0 UPBROADCASTRUNNINGMULTICASTMTU:1500Metric:1 Interrupt:19Baseaddress:0x2000

当节点二的heartbeat服务起来之后将抢夺资源 [root@node2ha.d]#ifconfig eth0Linkencap:EthernetHWaddr00:0C:29:D2:75:D0 inetaddr:192.168.20.20Bcast:192.168.20.255Mask:255.255.255.0 inet6addr:fe80::20c:29ff:fed2:75d0/64Scope:Link UPBROADCASTRUNNINGMULTICASTMTU:1500Metric:1 RXpackets:6379errors:0dropped:0overruns:0frame:0 TXpackets:4892errors:0dropped:0overruns:0carrier:0 collisions:0txqueuelen:1000 RXbytes:3005372(2.8MiB)TXbytes:712083(695.3KiB) Interrupt:19Baseaddress:0x2000

eth0:0Linkencap:EthernetHWaddr00:0C:29:D2:75:D0 inetaddr:192.168.20.100Bcast:192.168.20.255Mask:255.255.255.0 UPBROADCASTRUNNINGMULTICASTMTU:1500Metric:1 Interrupt:19Baseaddress:0x2000 节点一执行hb_takeover抢夺资源 [root@node1heartbeat]#ifconfig eth0Linkencap:EthernetHWaddr00:0C:29:FA:2D:49 inetaddr:192.168.20.10Bcast:192.168.20.255Mask:255.255.255.0 inet6addr:fe80::20c:29ff:fefa:2d49/64Scope:Link UPBROADCASTRUNNINGMULTICASTMTU:1500Metric:1 RXpackets:10759errors:0dropped:0overruns:0frame:0 TXpackets:7779errors:0dropped:0overruns:0carrier:0 collisions:0txqueuelen:1000 RXbytes:3175621(3.0MiB)TXbytes:1028476(1004.3KiB) Interrupt:19Baseaddress:0x2000

eth0:0Linkencap:EthernetHWaddr00:0C:29:FA:2D:49 inetaddr:192.168.20.100Bcast:192.168.20.255Mask:255.255.255.0 UPBROADCASTRUNNINGMULTICASTMTU:1500Metric:1 Interrupt:19Baseaddress:0x2000 在节点一上查看虚拟服务规则表 [root@node1heartbeat]#ipvsadm-l IPVirtualServerversion1.2.1(size=4096) ProtLocalAddress:PortSchedulerFlags ->RemoteAddress:PortForwardWeightActiveConnInActConn TCP192.168.20.100:httprr ->192.168.20.30:httpRoute100 ->192.168.20.40:httpRoute100

节点二上没有虚拟ip root@node2ha.d]#ifconfig eth0Linkencap:EthernetHWaddr00:0C:29:D2:75:D0 inetaddr:192.168.20.20Bcast:192.168.20.255Mask:255.255.255.0 inet6addr:fe80::20c:29ff:fed2:75d0/64Scope:Link UPBROADCASTRUNNINGMULTICASTMTU:1500Metric:1 RXpackets:6417errors:0dropped:0overruns:0frame:0 TXpackets:4923errors:0dropped:0overruns:0carrier:0 collisions:0txqueuelen:1000 RXbytes:3009061(2.8MiB)TXbytes:716759(699.9KiB) Interrupt:19Baseaddress:0x2000

eth1Linkencap:EthernetHWaddr00:0C:29:CA:A4:3C inetaddr:192.168.30.20Bcast:192.168.30.255Mask:255.255.255.0 inet6addr:fe80::20c:29ff:feca:a43c/64Scope:Link UPBROADCASTRUNNINGMULTICASTMTU:1500Metric:1 RXpackets:4112errors:0dropped:0overruns:0frame:0 TXpackets:3482errors:0dropped:0overruns:0carrier:0 collisions:0txqueuelen:1000 RXbytes:1018506(994.6KiB)TXbytes:863519(843.2KiB) Interrupt:19Baseaddress:0x2080

[root@node2ha.d]#ipvsadm-l IPVirtualServerversion1.2.1(size=4096) ProtLocalAddress:PortSchedulerFlags ->RemoteAddress:PortForwardWeightActiveConnInActConn

在客户端上访问http://192.168.20.100/ 得到的页面是后方真实的web服务器1上的内容Server1 刷新页面 得到的页面是后方真实的web服务器2上的内容Server2 server2

在节点一上查看虚拟服务规则表 [root@node1heartbeat]#ipvsadm-l IPVirtualServerversion1.2.1(size=4096) ProtLocalAddress:PortSchedulerFlags ->RemoteAddress:PortForwardWeightActiveConnInActConn TCP192.168.20.100:httprr ->192.168.20.30:httpRoute102 ->192.168.20.40:httpRoute102

当server1失效时 查看虚拟规则表 [root@node1heartbeat]#ipvsadm-l IPVirtualServerversion1.2.1(size=4096) ProtLocalAddress:PortSchedulerFlags ->RemoteAddress:PortForwardWeightActiveConnInActConn TCP192.168.20.100:httprr ->192.168.20.40:httpRoute101

再次访问http://192.168.20.100/ 得到的页面是后方真实的web服务器2上的内容 Server2 刷新还是server2