描述一些前置知识：
BGP的全称是border gateway protocol，是用在AS之间的协议。
AS是Autonomous system的简称，可以理解为运行同一种选路策略、统一管理的一组路由器。
每个AS都有一个AS number，取值范围为 1-65535 ，其中1-64511是公有的，需要向IANA申请 ， 64512-65535是私有的可以自由使用，类似192.168的IP地址。
跨AS的是EBGP，AS内的是IBGP。
- BGP是基于tcp做的封包，可以理解为bgp的包头外面会由tcp封包，tcp外面由IP封包，然后然后是二层的包头。在网络中传输。
关于路由黑洞：
以2.1的拓扑图为例：
pc1的192.168.1.10对pc2的172.16.30.10做ping测试，就需要有路由信息。路由信息封装在bgp中， 从R4到R5的链路中，R1没有运行bgp，当bgp的数据包传输的时候，因为是基于tcp封装的，bgp信息依次被tcp、ip、以太网等包头封装，和普通的网络包一样，只要相邻的两台路由器可以互通，就可以把数据包从R4传送到R5；但是数据包从pc2返回pc1的时候，封包格式为icmp+ip，IP包头中会有src和dest的地址，这个dest ip就是pc1的192.168.1.10地址，这个数据包到了R3的后，R1是没有bgp协议的，也就学不到192.168.1.0的路由信息，所以传输会失败。这就是路由黑洞。
解决路由黑洞的方法有好几种，比如：
物理线路的full mesh，也就是说将as内的边缘路由器连线
将BGP重分布进OSPF
AS内所有路由器都运行bgp（本次实验采用的）
使用MPLS做tunnel
2. 测试过程本次实验是采用GNS3模拟器做的，采用了IOU设备的方式。
模拟了bgp的实现过程，顺便验证了一下BGP中的路由黑洞。
2.1 测试拓扑拓扑解释：R1、R2、R3为一个自治区域，号码为123，为ibgp，3台路由器之间运行ospf，R2和R3运行bgp，R1仅运行ospf
R4为一个自治区域，号码为4，和R123之间为ebgp
R5为一个自治区域，号码为5，和R123之间为ebgp
各自路由器的router id为路由器编号地址，比如R1的为1.1.1.1
2.2 R1、R2、R3配置IP，配置ospf
1、配置IP地址
R1配置IP：
R1#conf tEnter configuration commands, one per line. End with CNTL/Z.R1(config)#int s2/0R1(config-if)#ip add 12.1.1.2 255.255.255.0R1(config-if)#no shuR1(config-if)#exitR1(config)#int s2/1R1(config-if)#ip add 11.1.1.1 255.255.255.0R1(config-if)#no shuR2配置IP：
R2#conf tEnter configuration commands, one per line. End with CNTL/Z.R2(config)#int s2/0R2(config-if)#ip add 14.1.1.2 255.255.255.0R2(config-if)#no shuR2(config-if)#exitR2(config)#int s2/1R2(config-if)#ip add 12.1.1.1 255.255.255.0R2(config-if)#no shuR3配置IP：
R3#conf tEnter configuration commands, one per line. End with CNTL/Z.R3(config)#int s2/0R3(config-if)#ip add 11.1.1.2 255.255.255.0R3(config-if)#no shuR3(config-if)#exitR3(config)#int s2/1R3(config-if)#ip add 13.1.1.1 255.255.255.0R3(config-if)#no shu2、配置ospf，对于R2和R3来说，只配置AS内的接口，外联ebgp的口不配置到ospf中
R1配置
R1(config)#router ospf 110R1(config-router)#router-id 1.1.1.1R1(config-router)#net 12.1.1.2 0.0.0.0 area 0R1(config-router)#net 11.1.1.1 0.0.0.0 area 0R2配置
R2(config)#router ospf 110R2(config-router)#router-id 2.2.2.2R2(config-router)#net 12.1.1.1 0.0.0.0 area 0R3配
R3(config)#router ospf 110 R3(config-router)#router-id 3.3.3.3R3(config-router)#network 11.1.1.2 0.0.0.0 area 03、查看路由表信息
R2#show ip route Codes: L - local, C - connected, S - static, R - RIP, M - mobile, B - BGP D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2 E1 - OSPF external type 1, E2 - OSPF external type 2 i - IS-IS, su - IS-IS summary, L1 - IS-IS level-1, L2 - IS-IS level-2 ia - IS-IS inter area, * - candidate default, U - per-user static route o - ODR, P - periodic downloaded static route, H - NHRP, l - LISP a - application route + - replicated route, % - next hop overrideGateway of last resort is not set 11.0.0.0/24 is subnetted, 1 subnetsO 11.1.1.0 [110/128] via 12.1.1.2, 00:01:46, Serial2/1 12.0.0.0/8 is variably subnetted, 2 subnets, 2 masksC 12.1.1.0/24 is directly connected, Serial2/1L 12.1.1.1/32 is directly connected, Serial2/1查看ospf
R2#show ip ospf database OSPF Router with ID (2.2.2.2) (Process ID 110) Router Link States (Area 0)Link ID ADV Router Age Seq# Checksum Link count1.1.1.1 1.1.1.1 178 0x80000005 0x00A804 42.2.2.2 2.2.2.2 205 0x80000004 0x0083EC 23.3.3.3 3.3.3.3 179 0x80000004 0x002742 2这部分分为ibgp和ebgp的
2.3.1 为as123配置ibgp1、在R2配置bgp协议
R2(config)#router bgp 123R2(config-router)#bgp router-id 2.2.2.2R2(config-router)#neighbor 11.1.1.2 remote-as 123 //为bgp指定邻居，这里指定的是本as内R3的s2/0R2(config-router)#*Feb 29 15:43:36.408: %BGP-5-ADJCHANGE: neighbor 11.1.1.2 Up //等待片刻，可以看到邻居建立R2(config-router)#neighbor 14.1.1.1 remote-as 4 //这里是指定ebgp对端as4的邻居关系2、在R3配置bgp协议
R3(config)#router bgp 123R3(config-router)#bgp router-id 3.3.3.3R3(config-router)#neighbor 12.1.1.1 remote-as 123R3(config-router)#neighbor 13.1.1.2 remote-as 53、查看bgp信息，因为as4和5的IP和bgp还没配置，所以邻居部分的显示不正常
R2执行查看
R2#show ip bgp summary BGP router identifier 2.2.2.2, local AS number 123BGP table version is 1, main routing table version 1Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd11.1.1.2 4 123 9 9 1 0 0 00:04:47 014.1.1.1 4 4 0 0 1 0 0 never IdleR3执行查看
R3#show ip bgp summary BGP router identifier 3.3.3.3, local AS number 123BGP table version is 1, main routing table version 1Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd12.1.1.1 4 123 9 10 1 0 0 00:05:28 013.1.1.2 4 5 0 0 1 0 0 never Idle1、首先配置R4和5的IP地址
R4的地址
R4#conf tEnter configuration commands, one per line. End with CNTL/Z.R4(config)#int s2/0 R4(config-if)#ip add 14.1.1.1 255.255.255.0R4(config-if)#no shuR4(config-if)#exitR4(config)#int e0/0R4(config-if)#ip add 192.168.1.1 255.255.255.0 R4(config-if)#no shuR5的地址
R5#conf tEnter configuration commands, one per line. End with CNTL/Z.R5(config)#int s2/0R5(config-if)#ip add 13.1.1.2 255.255.255.0R5(config-if)#no shuR5(config-if)#exitR5(config)#int e0/0R5(config-if)#ip add 172.16.30.1 255.255.255.0R5(config-if)#no shu2、为R4配置bgp
R4(config)#router bgp 4R4(config-router)#bgp router-id 4.4.4.4R4(config-router)#neighbor 14.1.1.2 remote-as 123 //指定as123中R2的s2/0为邻居R4(config-router)#*Feb 29 15:54:02.133: %BGP-5-ADJCHANGE: neighbor 14.1.1.2 Up //邻居建立提示3、为R5配置bgp
R5(config-router)#bgp routerR5(config-router)#bgp router-id 5.5.5.5R5(config-router)#neighbor 13.1.1.1 remote-as 1234、查看bgp信息
R4#show ip bgp summary BGP router identifier 4.4.4.4, local AS number 4BGP table version is 1, main routing table version 1Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd14.1.1.2 4 123 6 6 1 0 0 00:02:38 0此时再看R2或者R3的bgp信息，邻居关系就都正常了
R2#show ip bgp summary BGP router identifier 2.2.2.2, local AS number 123BGP table version is 1, main routing table version 1Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd11.1.1.2 4 123 19 18 1 0 0 00:13:40 014.1.1.1 4 4 7 7 1 0 0 00:03:14 0这里是要将pc1的192.168.1.0/24这个网络宣告到bgp中，步骤如下
1、在R4上宣告路由
R4(config)#router bgp 4R4(config-router)#network 192.168.1.0 mask 255.255.255.0 //只能宣告网段，指定掩码查看bgp 路由信息
R4#show ip bgp BGP table version is 2, local router ID is 4.4.4.4Status codes: s suppressed, d damped, h history, * valid, > best, i - internal, r RIB-failure, S Stale, m multipath, b backup-path, f RT-Filter, x best-external, a additional-path, c RIB-compressed, Origin codes: i - IGP, e - EGP, ? - incompleteRPKI validation codes: V valid, I invalid, N Not found Network Next Hop Metric LocPrf Weight Path *> 192.168.1.0 0.0.0.0 0 32768 i本地产生，下一跳为0.0.0.0
2、此时到R2上查看bgp信息，发现路由信息已经过来了，且可以看到192.168.1.0网段的下一跳为14.1.1.1（R4的S2/0，出口接口）
R2#show ip bgpBGP table version is 2, local router ID is 2.2.2.2Status codes: s suppressed, d damped, h history, * valid, > best, i - internal, r RIB-failure, S Stale, m multipath, b backup-path, f RT-Filter, x best-external, a additional-path, c RIB-compressed, Origin codes: i - IGP, e - EGP, ? - incompleteRPKI validation codes: V valid, I invalid, N Not found Network Next Hop Metric LocPrf Weight Path *> 192.168.1.0 14.1.1.1 0 0 4 iR2上也能看到路由信息，可以看到192.168.1.0是通过bgp学习到的，metric是20，通过ebgp学习到的路由条目metric都是20，如果是ibgp则为200
R2#show ip route Codes: L - local, C - connected, S - static, R - RIP, M - mobile, B - BGP D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2 E1 - OSPF external type 1, E2 - OSPF external type 2 i - IS-IS, su - IS-IS summary, L1 - IS-IS level-1, L2 - IS-IS level-2 ia - IS-IS inter area, * - candidate default, U - per-user static route o - ODR, P - periodic downloaded static route, H - NHRP, l - LISP a - application route + - replicated route, % - next hop overrideGateway of last resort is not set 11.0.0.0/24 is subnetted, 1 subnetsO 11.1.1.0 [110/128] via 12.1.1.2, 00:31:18, Serial2/1 12.0.0.0/8 is variably subnetted, 2 subnets, 2 masksC 12.1.1.0/24 is directly connected, Serial2/1L 12.1.1.1/32 is directly connected, Serial2/1 14.0.0.0/8 is variably subnetted, 2 subnets, 2 masksC 14.1.1.0/24 is directly connected, Serial2/0L 14.1.1.2/32 is directly connected, Serial2/0B 192.168.1.0/24 [20/0] via 14.1.1.1, 00:05:353、再看R3上的bgp信息，发现下一跳还是R4上的s2/0，而且路由条目前面没有>，不是最优路由，说明没有把信息传过来，ibgp传输路由信息的时候不会传输下一跳的地址
R3#show ip bgpBGP table version is 1, local router ID is 3.3.3.3Status codes: s suppressed, d damped, h history, * valid, > best, i - internal, r RIB-failure, S Stale, m multipath, b backup-path, f RT-Filter, x best-external, a additional-path, c RIB-compressed, Origin codes: i - IGP, e - EGP, ? - incompleteRPKI validation codes: V valid, I invalid, N Not found Network Next Hop Metric LocPrf Weight Path * i 192.168.1.0 14.1.1.1 0 100 0 4 i需要在R2路由上添加next-hop-self，才会把下一跳的信息传递过来。这个参数只用在as边缘路由器向内部传ibgp信息时。
R2(config)#router bgp 123R2(config-router)#neighbor 11.1.1.2 next-hop-self此时再看，发现下一跳的地址已经更新过来了，12.1.1.1正是R2的s2/1出口
R3#show ip bgpBGP table version is 2, local router ID is 3.3.3.3Status codes: s suppressed, d damped, h history, * valid, > best, i - internal, r RIB-failure, S Stale, m multipath, b backup-path, f RT-Filter, x best-external, a additional-path, c RIB-compressed, Origin codes: i - IGP, e - EGP, ? - incompleteRPKI validation codes: V valid, I invalid, N Not found Network Next Hop Metric LocPrf Weight Path *>i 192.168.1.0 12.1.1.1 0 100 0 4 i再看R3的路由信息，也就有bgp学习到的条目了，当然因为是ibgp传输，所有metric值为200
R3#show ip routeCodes: L - local, C - connected, S - static, R - RIP, M - mobile, B - BGP D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2 E1 - OSPF external type 1, E2 - OSPF external type 2 i - IS-IS, su - IS-IS summary, L1 - IS-IS level-1, L2 - IS-IS level-2 ia - IS-IS inter area, * - candidate default, U - per-user static route o - ODR, P - periodic downloaded static route, H - NHRP, l - LISP a - application route + - replicated route, % - next hop overrideGateway of last resort is not set 11.0.0.0/8 is variably subnetted, 2 subnets, 2 masksC 11.1.1.0/24 is directly connected, Serial2/0L 11.1.1.2/32 is directly connected, Serial2/0 12.0.0.0/24 is subnetted, 1 subnetsO 12.1.1.0 [110/128] via 11.1.1.1, 00:40:23, Serial2/0 13.0.0.0/8 is variably subnetted, 2 subnets, 2 masksC 13.1.1.0/24 is directly connected, Serial2/1L 13.1.1.1/32 is directly connected, Serial2/1B 192.168.1.0/24 [200/0] via 12.1.1.1, 00:04:57同理，R3上也需要为R2来设置next-hop-self，否则R5向R4方向的路由也会出现同样的问题
R3(config)#router bgp 123R3(config-router)#neighbor 12.1.1.1 next-hop-self过程和上一章节类似
1、直接在R5上宣告
R5(config)#router bgp 5 R5(config-router)#network 172.16.30.0 mask 255.255.255.02、在R4上已经可以看到正常的路由条目了
R4#show bgpBGP table version is 3, local router ID is 4.4.4.4Status codes: s suppressed, d damped, h history, * valid, > best, i - internal, r RIB-failure, S Stale, m multipath, b backup-path, f RT-Filter, x best-external, a additional-path, c RIB-compressed, O��ȡ,��ȡrigin codes: i - IGP, e - EGP, ? - incompleteRPKI validation codes: V valid, I invalid, N Not found Network Next Hop Metric LocPrf Weight Path *> 172.16.30.0/24 14.1.1.2 0 123 5 i *> 192.168.1.0 0.0.0.0 0 32768 i1、pc机配置IP地址
在pc1上配置192.168.1.10
在pc2上配置172.16.30.10
2、在路由器R1上打开debug
R1#debug ip packet
IP packet debugging is on
3、执行ping测试，失败
wangql> ping 172.16.30.10172.16.30.10 icmp_seq=1 timeout172.16.30.10 icmp_seq=2 timeout172.16.30.10 icmp_seq=3 timeout172.16.30.10 icmp_seq=4 timeout172.16.30.10 icmp_seq=5 timeout4、观察R1的输出，提示源192.168.1.10到目的172.16.30.10不可达
*Feb 29 16:22:51.401: IP: s=192.168.1.10 (Serial2/0), d=172.16.30.10, len 84, input feature, MCI Check(99), rtype 0, forus FALSE, sendself FALSE, mtu 0, fwdchk FALSE*Feb 29 16:22:51.401: IP: s=192.168.1.10 (Serial2/0), d=172.16.30.10, len 84, unroutable因为R1没开启bgp，只有ospf，所有没有这两个网段的路由
R1#show ip route Codes: L - local, C - connected, S - static, R - RIP, M - mobile, B - BGP D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2 E1 - OSPF external type 1, E2 - OSPF external type 2 i - IS-IS, su - IS-IS summary, L1 - IS-IS level-1, L2 - IS-IS level-2 ia - IS-IS inter area, * - candidate default, U - per-user static route o - ODR, P - periodic downloaded static route, H - NHRP, l - LISP a - application route + - replicated route, % - next hop overrideGateway of last resort is not set 11.0.0.0/8 is variably subnetted, 2 subnets, 2 masksC 11.1.1.0/24 is directly connected, Serial2/1L 11.1.1.1/32 is directly connected, Serial2/1 12.0.0.0/8 is variably subnetted, 2 subnets, 2 masksC 12.1.1.0/24 is directly connected, Serial2/0L 12.1.1.2/32 is directly connected, Serial2/01、R1启用bgp，并添加对端端口的邻居
R1(config)#router bgp 123R1(config-router)#bgp router-id 1.1.1.1R1(config-router)#neighbor 12.1.1.1 remote-as 123R1(config-router)#neighbor 11.1.1.2 remote-as 1232 R2和R3两天边缘路由器需要对R1添加邻居信息和next-hop-self，否则无法传递bgp路由的下一跳参数R2
R2(config)#router bgp 123R2(config-router)#neighbor 12.1.1.2 remote-as 123R2(config-router)#neighbor 12.1.1.2 next-hop-selfR3
R3(config)#router bgp 123R3(config-router)#neighbor 11.1.1.1 remote-as 123R3(config-router)#neighbor 11.1.1.1 next-hop-self3、再次测试pc1和pc2，网络通畅了
wangql> ping 172.16.30.1084 bytes from 172.16.30.10 icmp_seq=1 ttl=59 time=36.557 ms84 bytes from 172.16.30.10 icmp_seq=2 ttl=59 time=36.735 ms84 bytes from 172.16.30.10 icmp_seq=3 ttl=59 time=36.005 ms84 bytes from 172.16.30.10 icmp_seq=4 ttl=59 time=33.740 mswangql> trace 172.16.30.10trace to 172.16.30.10, 8 hops max, press Ctrl+C to stop 1 192.168.1.1 0.290 ms 0.141 ms 0.163 ms 2 14.1.1.2 9.157 ms 9.058 ms 8.496 ms 3 12.1.1.2 18.042 ms 17.537 ms 17.808 ms 4 11.1.1.2 26.682 ms 26.471 ms 26.210 ms 5 13.1.1.2 36.360 ms 35.922 ms 34.192 ms 6 *172.16.30.10 37.326 ms (ICMP type:3, code:3, Destination port unreachable)通过上面的实验，总结如下：
BGP不同于那些IGP的协议，配置的时候需要手工指定邻居关系。
在IBGP的配置过程中，除了显式指定邻居关系，还需要手工配置 next-hop-self，否则传递的BGP路由信息中下一跳的信息不会发送改变，导致路由信息无效
- 在做网络规划的时候，要注意拓扑的合理性，避免出现路由黑洞。
如有任何问题，可点击文末阅读原文，到社区原文下评论交流