OSPF与BGP协议联动

Posted 2022-02-02 szc425

一、OSPF与BGP协议联动分析

免责声明：上图来源于网络，如有版权问题请联系作者删除！

• 实验环境：4台路由器底层运行ospf并建立IBGP邻居；RouterB、RouterC配置为RR路由反射器

• 在有备份链路的情况下，BGP在链路回切时，由于BGP路由收敛速度滞后于ospf路由收敛速度，从而造成流量丢失

• 如上图所示，四台设备RouterA、RouterB、RouterC、RouterD之间运行ospf协议，并建立IBGP连接。RouterC为RouterB的备份设备，当网络环境稳定时，BGP与ospf在设备上是完全收敛的，默认RouterA访问10.3.1.0/30优先从B转发（A→B→D→E），且RouterA收到10.3.1.0/30的下一跳为RouterD的环回接口（RR路由反射器不修改下一跳）

1、问题

• 正常情况下，从RouterA到10.3.1.0/30的流量会优先从RouterB转发。当RouterB发生故障后（比如关机重启了），流量切换到RouterC转发，注意不是立刻切回到C，因为ospf hello包的死亡超时时间为40s，40s之后RouterA才会感知到与RouterB的邻居关系中断了，流量转而从RouterC转发。RouterB故障恢复以后，重新与RouterA建立起ospf邻居关系，流量回切到RouterB，此时会有流量丢失

2、原因分析

• RouterB故障恢复以后，RouterB与RouterA、RouterD的ospf邻居立刻建立，RouterA访问 RouterD环回接口的路由可能从RouterB转发

• 在流量回切到RouterB的过程中，IGP收敛速度比BGP快。RouterA去往10.3.1.0/30的下一跳为RouterD的环回接口（RR路由反射器），RouterA去往RouterD环回接口优先从RouterB转发，但RouterB的BGP路由还没有完成收敛（RouterB还没有学习到去往10.3.1.0/30的BGP路由），数据包在RouterB上丢弃，导致路由转发黑洞

3、解决办法

• 核心思想：在RouterB完成BGP收敛之前，将ospf的metric改大（LSA-1的metric值改大），让RouterA去往RouterD别选择我作为下一跳，即RouterA访问RouterD环回接口时，因为RouterA去往10.3.1.0/30的下一跳为RouterD的环回接口

• RouterB(config)#router ospf 100
  RouterB(config-router)#max-metric router-lsa on-starup wait-for-bgp //等BGP协议完成收敛，ospf的metric值由最大恢复正常
    OR
  RouterB(config-router)#max-metric router-lsa on-starup 10 //ospf协议起来10min之内将metric值设置为最大（目的是让邻居路由别把我当做下一跳），这段时间希望BGP协议完成收敛；10min之后，ospf的metric值由最大恢复正常

• 这样配置以后，流量回切到RouterB时，RouterA去往RouterD环回时不会把RouterB当做下一跳

 

二、实验验证

（1）实验环境：R1配置lo100： 172.16.1.1/32，不宣告进ospf，宣告进AS 10，R6配置lo100： 192.168.1.1/32，不宣告进ospf，宣告进AS 20。默认172.16.1.1访问192.168.1.1的转发路径：R1-R2-R4-R5-R6，可通过抓包软件看到效果，即172.16.1.1访问192.168.1.1时，在R1的f0/0、R3的f0/0接口抓包，看不到ICMP流量，而在R1的f1/0、R2的f1/0接口抓包能看到ICMP流量。

（2）实验步骤：172.16.1.1访问192.168.1.1ping上10000个包同时在R1的f1/0接口抓包，中途将R2设备关机再开机观察效果（R2断电之前一定要保存配置）。

• 在R2上没配置R2(config-router)#max-metric router-lsa on-starup wait-for-bgp之前，效果如下：

  R1#ping 192.168.1.1 source 172.16.1.1 repeat 10000

  Type escape sequence to abort.
  Sending 10000, 100-byte ICMP Echos to 192.168.1.1, timeout is 2 seconds:
  Packet sent with a source address of 172.16.1.1
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!................ //此时将R2设备断电
  *Mar 1 00:30:45.159: %OSPF-5-ADJCHG: Process 100, Nbr 2.2.2.2 on FastEthernet1/0 from FULL to DOWN, Neighbor Down:
  Dead timer expired..... //这段时间产生丢包是因为ospf hello超时时间为40s，40s之后R1才感知到与R2的ospf邻居关系中断了
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! //这段时间流量转发正常，是因为去往192.168.1.1的流量转而从R3转发了
  *Mar 1 00:31:21.539: %OSPF-5-ADJCHG: Process 100, Nbr 2.2.2.2 on FastEthernet1/0 from LOADING to FULL, Loading
  Done!!!!!!!!!!!!!!!!!!!!!!!!!!!!............. //将R2设备重启，这段时间又产生丢包是因为R2与R1、R1的ospf建立成功，R1去往R4的数据包又通过R2转发，但R2的BGP还没有完成收敛，BGP数据库中没有去往192.168.1.1的路由，形成路由转发黑洞
  *Mar 1 00:31:39.963: %BGP-5-ADJCHANGE: neighbor 2.2.2.2 Down Peer closed the session.......
  *Mar 1 00:31:48.627: %BGP-5-ADJCHANGE: neighbor 2.2.2.2 Up ..........!!!!!!!!!!!!!!!!!!!!!!!
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!   //当R2的BGP数据库中有192.168.1.1路由时，流量又可以通了

• 在R2上没配置R2(config-router)#max-metric router-lsa on-starup wait-for-bgp之后，效果如下：

  R1#ping 192.168.1.1 source 172.16.1.1 repeat 10000

  Type escape sequence to abort.
  Sending 10000, 100-byte ICMP Echos to 192.168.1.1, timeout is 2 seconds:
  Packet sent with a source address of 172.16.1.1 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!........
  ..............    //只在R1感知与R2邻居关系中断的这段时间产生丢包
  *Mar 1 00:36:56.427: %OSPF-5-ADJCHG: Process 100, Nbr 2.2.2.2 on FastEthernet1/0 from FULL to DOWN, Neighbor Down:
  Dead timer expired....!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  !!!!!!!!!!!!!!!!!
  *Mar 1 00:37:32.987: %OSPF-5-ADJCHG: Process 100, Nbr 2.2.2.2 on FastEthernet1/0 from LOADING to FULL, Loading
  Done!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  *Mar 1 00:37:49.443: %BGP-5-ADJCHANGE: neighbor 2.2.2.2 Down Peer closed the session!!!!!!!!!!!!
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  *Mar 1 00:37:56.319: %BGP-5-ADJCHANGE: neighbor 2.2.2.2 Up !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

• R2重启完成初期，路由表中没有BGP路由，只有O路由且metric值最大，目的让R1去往4.4.4.4别以我为下一跳

  R2#sho ip route   //只截取一部分
      34.0.0.0/24 is subnetted, 1 subnets
  O 34.1.1.0 [110/65536] via 24.1.1.4, 00:00:01, FastEthernet1/0
      1.0.0.0/32 is subnetted, 1 subnets
  O 1.1.1.1 [110/65536] via 12.1.1.1, 00:00:01, FastEthernet0/0
      4.0.0.0/32 is subnetted, 1 subnets
  O 4.4.4.4 [110/65536] via 24.1.1.4, 00:00:03, FastEthernet1/0
      13.0.0.0/24 is subnetted, 1 subnets
  O 13.1.1.0 [110/65536] via 12.1.1.1, 00:00:03, FastEthernet0/0

 

三、结语

此实验一定要配合抓包软件观察效果，这样才能有深刻的印象，感兴趣的朋友可以自己动手试验一下啊~文中若有不足，请在评论区留言~~