方形拓扑中采用CloudSec的多站点VXLAN故障排除

简介

本文档介绍以方形拓扑连接的边界网关之间的VXLAN多站点配置和CloudSec故障排除。

先决条件

要求

思科建议您熟悉以下主题：

• Nexus NXOS ©软件。
• VXLAN EVPN技术。
• BGP和OSPF路由协议。

使用的组件

本文档中的信息基于以下软件和硬件版本：

• Cisco Nexus 9000。
• NXOS 10.3(4a)版。

本文档中的信息都是基于特定实验室环境中的设备编写的。本文档中使用的所有设备最初均采用原始（默认）配置。如果您的网络处于活动状态，请确保您了解所有命令的潜在影响。

配置

网络图

VXLAN MultiSite，采用CloudSec方形拓扑

拓扑的详细信息

• 双站点多站点VXLAN EVPN交换矩阵。
• 两个站点都配置了vPC边界网关。
• 终端托管在VLAN 1100中。
• 每个站点的边界网关在SVI接口Vlan3600上彼此之间具有IPv4 iBGP邻居关系。
• 一个站点上的边界网关仅与另一个站点上的直接连接的边界网关具有eBGP IPv4邻居关系。
• 站点A的边界网关与站点B的边界网关具有eBGP L2VPN EVPN邻居关系。

寻址计划

表中的IP地址在配置过程中使用：

配置

• 请注意，本指南中仅显示与多站点相关的配置。对于完整配置，您可以使用VXLAN的思科官方文档指南Cisco Nexus 9000系列NX-OS VXLAN配置指南，版本10.3(x)

要启用CloudSec，必须在evpn multisite border-gateway下配置dci-advertise-pip  命令：

evpn multisite border-gateway 65001
  dci-advertise-pip

evpn multisite border-gateway 65002
  dci-advertise-pip

BGP配置

此配置特定于站点。 

router bgp 65001
  address-family ipv4 unicast
    maximum-paths 64
  address-family l2vpn evpn
    maximum-paths 64
    additional-paths send
    additional-paths receive

router bgp 65002
  address-family ipv4 unicast
    maximum-paths 64
  address-family l2vpn evpn
    maximum-paths 64
    additional-paths send
    additional-paths receive

• maximum-path 命令允许从邻居接收多个eBGP L2VPN EVPN路径。
• additional-path 命令指示BGP进程通告设备有能力发送/接收其他路径

对于边界网关上的所有L3VNI VRF，还必须配置多路径：

router bgp 65001
vrf tenant-1
    address-family ipv4 unicast
      maximum-paths 64
    address-family ipv6 unicast
      maximum-paths 64

router bgp 65002
vrf tenant-1
    address-family ipv4 unicast
      maximum-paths 64
    address-family ipv6 unicast
      maximum-paths 64

隧道加密配置

所有边界网关上的此配置必须相同：

key chain CloudSec_Key_Chain1 tunnel-encryption key 1000 key-octet-string Cl0udSec! cryptographic-algorithm AES_128_CMAC feature tunnel-encryption tunnel-encryption must-secure-policy tunnel-encryption source-interface loopback0 tunnel-encryption policy CloudSec_Policy1

此配置特定于站点。tunnel-encryption命令必须只应用于带有evpn multisite dci-tracking命令的接口。

tunnel-encryption peer-ip 192.168.13.2
  keychain CloudSec_Key_Chain1 policy CloudSec_Policy1
tunnel-encryption peer-ip 192.168.13.3
  keychain CloudSec_Key_Chain1 policy CloudSec_Policy1

interface Ethernet1/48
tunnel-encryption

tunnel-encryption peer-ip 192.168.3.2
  keychain CloudSec_Key_Chain1 policy CloudSec_Policy1
tunnel-encryption peer-ip 192.168.3.3
  keychain CloudSec_Key_Chain1 policy CloudSec_Policy1

interface Ethernet1/3
tunnel-encryption

启用隧道加密后，向邻居通告路由时，其他属性将添加到本地环回，并且所有eBGP IPv4单播邻居必须看到此属性：

SA-BGW-A# show ip bgp 192.168.2.3 BGP routing table information for VRF default, address family IPv4 Unicast BGP routing table entry for 192.168.2.3/32, version 1320 Paths: (2 available, best #1) Flags: (0x000002) (high32 00000000) on xmit-list, is not in urib Multipath: eBGP iBGP Advertised path-id 1 Path type: local, path is valid, is best path, no labeled nexthop AS-Path: NONE, path locally originated 0.0.0.0 (metric 0) from 0.0.0.0 (192.168.2.3) Origin IGP, MED not set, localpref 100, weight 32768 Tunnel Encapsulation attribute: Length 152 !---This is a new attribute Path type: redist, path is valid, not best reason: Locally originated, no labeled nexthop AS-Path: NONE, path locally originated 0.0.0.0 (metric 0) from 0.0.0.0 (192.168.2.3) Origin incomplete, MED 0, localpref 100, weight 32768 Path-id 1 advertised to peers: 10.12.10.2 192.168.4.2 SA-BGW-A# 

对于路由类型2，还有新属性：

SA-BGW-A# show bgp l2vpn evpn 00ea.bd27.86ef BGP routing table information for VRF default, address family L2VPN EVPN Route Distinguisher: 65002:31100 BGP routing table entry for [2]:[0]:[0]:[48]:[00ea.bd27.86ef]:[0]:[0.0.0.0]/216, version 7092 Paths: (2 available, best #2) Flags: (0x000202) (high32 00000000) on xmit-list, is not in l2rib/evpn, is not in HW Multipath: eBGP iBGP Path type: external, path is valid, not best reason: Router Id, multipath, no labeled nexthop Imported to 1 destination(s) Imported paths list: L2-31100 AS-Path: 65002 , path sourced external to AS 192.168.13.3 (metric 0) from 192.168.12.4 (192.168.12.4) Origin IGP, MED 2000, localpref 100, weight 0 Received label 31100 Received path-id 1 Extcommunity: RT:65001:31100 ENCAP:8 ESI: 0300.0000.00fd.ea00.0309 Advertised path-id 1 Path type: external, path is valid, is best path, no labeled nexthop Imported to 1 destination(s) Imported paths list: L2-31100 AS-Path: 65002 , path sourced external to AS 192.168.13.2 (metric 0) from 192.168.12.3 (192.168.12.3) Origin IGP, MED 2000, localpref 100, weight 0 Received label 31100 Received path-id 1 Extcommunity: RT:65001:31100 ENCAP:8 ESI: 0300.0000.00fd.ea00.0309 Path-id 1 not advertised to any peer Route Distinguisher: 192.168.2.3:33867 (L2VNI 31100) BGP routing table entry for [2]:[0]:[0]:[48]:[00ea.bd27.86ef]:[0]:[0.0.0.0]/216, version 7112 Paths: (2 available, best #1) Flags: (0x000212) (high32 0x000400) on xmit-list, is in l2rib/evpn, is not in HW Multipath: eBGP iBGP Advertised path-id 1 Path type: external, path is valid, is best path, no labeled nexthop, in rib Imported from 65002:31100:[2]:[0]:[0]:[48]:[00ea.bd27.86ef]:[0]:[0.0.0.0]/216 AS-Path: 65002 , path sourced external to AS 192.168.13.2 (metric 0) from 192.168.12.3 (192.168.12.3) Origin IGP, MED 2000, localpref 100, weight 0 Received label 31100 Received path-id 1 Extcommunity: RT:65001:31100 ENCAP:8 ESI: 0300.0000.00fd.ea00.0309 Path type: external, path is valid, not best reason: Router Id, multipath, no labeled nexthop, in rib Imported from 65002:31100:[2]:[0]:[0]:[48]:[00ea.bd27.86ef]:[0]:[0.0.0.0]/216 AS-Path: 65002 , path sourced external to AS 192.168.13.3 (metric 0) from 192.168.12.4 (192.168.12.4) Origin IGP, MED 2000, localpref 100, weight 0 Received label 31100 Received path-id 1 Extcommunity: RT:65001:31100 ENCAP:8 ESI: 0300.0000.00fd.ea00.0309 !---Ethernet Segment Identifier (ESI) is also new attribute Path-id 1 (dual) advertised to peers: 192.168.2.2 SA-BGW-A# 

验证

在启用cloudsec之前，最好检查一下没有它的设置是否运行正常：

SA-BGW-A(config)# show clock Warning: No NTP peer/server configured. Time may be out of sync. 10:02:01.016 UTC Fri Jul 19 2024 Time source is NTP SA-BGW-A(config)# show tunnel-encryption session Tunnel-Encryption Peer Policy Keychain RxStatus TxStatus ------------------------ ---------------------------------------- ---------------------------------------- ----------------- ----------------- =============================================== SA-HOST-A# show clock Warning: No NTP peer/server configured. Time may be out of sync. 10:02:21.592 UTC Fri Jul 19 2024 Time source is NTP SA-HOST-A# ping 10.100.20.10 count unlimited interval 1 PING 10.100.20.10 (10.100.20.10): 56 data bytes 64 bytes from 10.100.20.10: icmp_seq=0 ttl=254 time=1.583 ms 64 bytes from 10.100.20.10: icmp_seq=1 ttl=254 time=10.407 ms 64 bytes from 10.100.20.10: icmp_seq=2 ttl=254 time=1.37 ms 64 bytes from 10.100.20.10: icmp_seq=3 ttl=254 time=1.489 ms 64 bytes from 10.100.20.10: icmp_seq=4 ttl=254 time=6.685 ms 64 bytes from 10.100.20.10: icmp_seq=5 ttl=254 time=1.547 ms 64 bytes from 10.100.20.10: icmp_seq=6 ttl=254 time=1.859 ms 64 bytes from 10.100.20.10: icmp_seq=7 ttl=254 time=5.219 ms 64 bytes from 10.100.20.10: icmp_seq=8 ttl=254 time=1.337 ms 64 bytes from 10.100.20.10: icmp_seq=9 ttl=254 time=3.528 ms 64 bytes from 10.100.20.10: icmp_seq=10 ttl=254 time=4.057 ms

配置完cloudsec后，SA上的终端必须成功ping通站点B上的终端。但是，在某些情况下，ping操作可能不成功。这取决于本地设备选择用于发送cloudsec加密流量的cloudsec对等体。

SA-HOST-A# ping 10.100.20.10 PING 10.100.20.10 (10.100.20.10): 56 data bytes Request 0 timed out Request 1 timed out Request 2 timed out Request 3 timed out Request 4 timed out --- 10.100.20.10 ping statistics --- 5 packets transmitted, 0 packets received, 100.00% packet loss SA-HOST-A# 

故障排除

检查源终端上的本地ARP表：

SA-HOST-A# ping 10.100.20.10 count unlimited interval 1 Request 352 timed out Request 353 timed out Request 354 timed out 356 packets transmitted, 0 packets received, 100.00% packet loss SA-HOST-A# clear ip arp delete-force SA-HOST-A# show ip arp Flags: * - Adjacencies learnt on non-active FHRP router + - Adjacencies synced via CFSoE # - Adjacencies Throttled for Glean CP - Added via L2RIB, Control plane Adjacencies PS - Added via L2RIB, Peer Sync RO - Re-Originated Peer Sync Entry D - Static Adjacencies attached to down interface IP ARP Table for context default Total number of entries: 1 Address Age MAC Address Interface Flags 10.100.20.10 00:00:02 00ea.bd27.86ef Vlan1100 SA-HOST-A#    

此输出证明BUM流量正在传递且控制平面工作正常。下一步是检查隧道加密状态：

SA-BGW-A# show tunnel-encryption session Tunnel-Encryption Peer Policy Keychain RxStatus TxStatus ------------------------ ---------------------------------------- ---------------------------------------- ----------------- ----------------- 192.168.13.2 CloudSec_Policy1 CloudSec_Key_Chain1 Secure (AN: 0) Secure (AN: 0) 192.168.13.3 CloudSec_Policy1 CloudSec_Key_Chain1 Secure (AN: 0) Secure (AN: 0) SA-BGW-A# 

此输出显示CloudSec会话已建立。下一步可以在SA-HOST-A上运行无限制ping：

SA-HOST-A# ping 10.100.20.10 count unlimited interval 1

从此时起，您必须检查站点A上的设备，查看流量是否到达此设备。您可以通过ELAM在站点A路径上的所有设备上完成此任务。 in-select  将默认值6更改为9可根据内部报头进行匹配。您可以通过此链接了解有关ELAM的更多信息：Nexus 9000云扩展ASIC (Tahoe) NX-OS ELAM。

SA-LEAF-A上的ELAM

在生产网络中，存在多个SPINE设备。要了解流量发送到哪个主干，您必须首先在LEAF上获取ELAM。尽管使用了 in-select 9 这种方法，但在连接到源的枝叶上，必须使用外部ipv4报头，因为到达此枝叶的流量不是VXLAN加密的。在真实网络中，可能很难捕获到您生成的确切数据包。在这种情况下，您可以运行特定长度的ping，并使用Pkt len报头来识别数据包。默认情况下，icmp数据包的长度为64字节。加上20字节的IP报头，总而言之，为您提供了84字节的PKT Len： 

SA-LEAF-A# debug platform internal tah elam SA-LEAF-A(TAH-elam)# trigger init in-select 9 Slot 1: param values: start asic 0, start slice 0, lu-a2d 1, in-select 9, out-select 0 SA-LEAF-A(TAH-elam-insel9)# set outer ipv4 src_ip 10.100.10.10 dst_ip 10.100.20.10 SA-LEAF-A(TAH-elam-insel9)# start SA-LEAF-A(TAH-elam-insel9)# report ELAM not triggered yet on slot - 1, asic - 0, slice - 0 SUGARBOWL ELAM REPORT SUMMARY slot - 1, asic - 0, slice - 1 ============================ Incoming Interface: Eth1/49 Src Idx : 0xc1, Src BD : 1100 Outgoing Interface Info: dmod 1, dpid 64 !---Note dpid value Dst Idx : 0xcd, Dst BD : 1100 Packet Type: IPv4 Outer Dst IPv4 address: 10.100.20.10 Outer Src IPv4 address: 10.100.10.10 Ver = 4, DSCP = 0, Don't Fragment = 0 Proto = 1, TTL = 255, More Fragments = 0 Hdr len = 20, Pkt len = 84, Checksum = 0xb4ae !---64 byte + 20 byte IP header Pkt len = 84 Inner Payload Type: CE L4 Protocol : 1 L4 info not available Drop Info: ---------- LUA: LUB: LUC: LUD: Final Drops: SA-LEAF-A(TAH-elam-insel9)# show system internal ethpm info all | i i "dpid=64" !---Put dpid value here IF_STATIC_INFO: port_name=Ethernet1/52,if_index:0x1a006600,ltl=5940,slot=0, nxos_port=204,dmod=1,dpid=64,unit=0,queue=65535,xbar_unitbmp=0x0,ns_pid=255,slice_num=1,port_on_slice=24,src_id=48 SA-LEAF-A(TAH-elam-insel9)# show cdp neighbors interface ethernet 1/52 Capability Codes: R - Router, T - Trans-Bridge, B - Source-Route-Bridge S - Switch, H - Host, I - IGMP, r - Repeater, V - VoIP-Phone, D - Remotely-Managed-Device, s - Supports-STP-Dispute Device-ID Local Intrfce Hldtme Capability Platform Port ID SA-SPINE-A(FDO242210CS) Eth1/52 130 R S s N9K-C93240YC-FX2 Eth1/52 Total entries displayed: 1 SA-LEAF-A(TAH-elam-insel9)# 

从该输出中，您可以看到流量到达SA-LEAF-A并从接口Ethernet1/52转发出去，该接口从拓扑连接到SA-SPINE-A。 

SA-SPINE-A上的ELAM

在SPINE上，Pkt Len值将更大，因为50字节的VXLAN报头也会添加。默认情况下，没有 vxlan-parse 或 feature nv overlay ，SPINE无法在内部报头上匹配。因此，您必须对SPINE使用 vxlan-parse enable 命令：

SA-SPINE-A(config-if)# debug platform internal tah elam SA-SPINE-A(TAH-elam)# trigger init in-select 9 Slot 1: param values: start asic 0, start slice 0, lu-a2d 1, in-select 9, out-select 0 SA-SPINE-A(TAH-elam)# vxlan-parse enable SA-SPINE-A(TAH-elam-insel9)# set inner ipv4 src_ip 10.100.10.10 dst_ip 10.100.20.10 SA-SPINE-A(TAH-elam-insel9)# start SA-SPINE-A(TAH-elam-insel9)# report ELAM not triggered yet on slot - 1, asic - 0, slice - 0 HEAVENLY ELAM REPORT SUMMARY slot - 1, asic - 0, slice - 1 ============================ Incoming Interface: Eth1/52 Src Idx : 0xcd, Src BD : 4153 Outgoing Interface Info: dmod 1, dpid 72 Dst Idx : 0xc5, Dst BD : 4151 Packet Type: IPv4 Outer Dst IPv4 address: 192.168.100.1 Outer Src IPv4 address: 192.168.3.1 Ver = 4, DSCP = 0, Don't Fragment = 0 Proto = 17, TTL = 255, More Fragments = 0 Hdr len = 20, Pkt len = 134, Checksum = 0x7d69 !---84 bytes + 50 bytes VXLAN header Pkt len = 134 Inner Payload Type: IPv4 Inner Dst IPv4 address: 10.100.20.10 Inner Src IPv4 address: 10.100.10.10 L4 Protocol : 17 L4 info not available Drop Info: ---------- LUA: LUB: LUC: LUD: Final Drops: SA-SPINE-A(TAH-elam-insel9)# show system internal ethpm info all | i i "dpid=72" IF_STATIC_INFO: port_name=Ethernet1/50,if_index:0x1a006200,ltl=5948,slot=0, nxos_port=196,dmod=1,dpid=72,unit=0,queue=65535,xbar_unitbmp=0x0,ns_pid=255,slice_num=1,port_on_slice=0,src_id=0 SA-SPINE-A(TAH-elam-insel9)# show cdp neighbors interface ethernet 1/50 Capability Codes: R - Router, T - Trans-Bridge, B - Source-Route-Bridge S - Switch, H - Host, I - IGMP, r - Repeater, V - VoIP-Phone, D - Remotely-Managed-Device, s - Supports-STP-Dispute Device-ID Local Intrfce Hldtme Capability Platform Port ID SA-BGW-A(FDO242210CX) Eth1/50 169 R S s N9K-C93240YC-FX2 Eth1/50 Total entries displayed: 1 SA-SPINE-A(TAH-elam-insel9)#   

SA-SPINE-A根据输出向SA-BGW-A发送流量。

SA-BGW-A上的ELAM

SA-BGW-A(TAH-elam-insel9)# set inner ipv4 src_ip 10.100.10.10 dst_ip 10.100.20.10 SA-BGW-A(TAH-elam-insel9)# start SA-BGW-A(TAH-elam-insel9)# report ELAM not triggered yet on slot - 1, asic - 0, slice - 0 HEAVENLY ELAM REPORT SUMMARY slot - 1, asic - 0, slice - 1 ============================ Incoming Interface: Eth1/50 Src Idx : 0xc5, Src BD : 1100 Outgoing Interface Info: dmod 1, dpid 48 Dst Idx : 0xbd, Dst BD : 1100 Packet Type: IPv4 Outer Dst IPv4 address: 192.168.100.1 Outer Src IPv4 address: 192.168.3.1 Ver = 4, DSCP = 0, Don't Fragment = 0 Proto = 17, TTL = 254, More Fragments = 0 Hdr len = 20, Pkt len = 134, Checksum = 0x7e69 Inner Payload Type: IPv4 Inner Dst IPv4 address: 10.100.20.10 Inner Src IPv4 address: 10.100.10.10 L4 Protocol : 17 L4 info not available Drop Info: ---------- LUA: LUB: LUC: LUD: Final Drops: SA-BGW-A(TAH-elam-insel9)# show system internal ethpm info all | i i "dpid=48" IF_STATIC_INFO: port_name=Ethernet1/48,if_index:0x1a005e00,ltl=5956,slot=0, nxos_port=188,dmod=1,dpid=48,unit=0,queue=65535,xbar_unitbmp=0x0,ns_pid=255,slice_num=0,port_on_slice=48,src_id=96 SA-BGW-A(TAH-elam-insel9)# show cdp neighbors interface ethernet 1/48 Capability Codes: R - Router, T - Trans-Bridge, B - Source-Route-Bridge S - Switch, H - Host, I - IGMP, r - Repeater, V - VoIP-Phone, D - Remotely-Managed-Device, s - Supports-STP-Dispute Device-ID Local Intrfce Hldtme Capability Platform Port ID SB-BGW-A(FDO2452070B) Eth1/48 122 R S s N9K-C93216TC-FX2 Eth1/3 Total entries displayed: 1 SA-BGW-A(TAH-elam-insel9)# 

根据SA-BGW-A的输出，流量从Ethernet1/48流向SB-BGW-A。下一步是检查SB-BGW-A：

SB-BGW-A# debug platform internal tah elam SB-BGW-A(TAH-elam)# trigger init in-select 9 Slot 1: param values: start asic 0, start slice 0, lu-a2d 1, in-select 9, out-select 0 SB-BGW-A(TAH-elam-insel9)# set inner ipv4 src_ip 10.100.10.10 dst_ip 10.100.20.10 SB-BGW-A(TAH-elam-insel9)# start SB-BGW-A(TAH-elam-insel9)# report ELAM not triggered yet on slot - 1, asic - 0, slice - 0 ELAM not triggered yet on slot - 1, asic - 0, slice - 1 !---Reset the previous filter and start again just in case if packet was not captured. SB-BGW-A(TAH-elam-insel9)# reset SB-BGW-A(TAH-elam-insel9)# set inner ipv4 src_ip 10.100.10.10 dst_ip 10.100.20.10 SB-BGW-A(TAH-elam-insel9)# start SB-BGW-A(TAH-elam-insel9)# report ELAM not triggered yet on slot - 1, asic - 0, slice - 0 ELAM not triggered yet on slot - 1, asic - 0, slice - 1 SB-BGW-A(TAH-elam-insel9)# 

根据SB-BGW-A的输出，ELAM甚至未触发。这意味着SB-BGW-B正在接收数据包，无法正确解密和解析这些数据包，或者根本无法接收这些数据包。要了解cloudsec流量发生的情况，可以再次在SB-BGW-A上运行ELAM，但是触发过滤器必须设置为用于cloudsec的外部IP地址，因为无法查看cloudsec加密传输数据包的内部报头。从先前的输出可以了解到，SA-BGW-A处理了流量，这意味着SA-BGW-A使用cloudsec加密流量。因此，您可以使用SA-BGW-A的NVE IP作为ELAM的触发过滤器。根据前面的输出，VXLAN加密ICMP数据包长度为134字节。加上摘要中的32字节的cloudsec报头，可得出166字节：

SB-BGW-A(TAH-elam-insel9)# reset SB-BGW-A(TAH-elam-insel9)# set outer ipv4 src_ip 192.168.3.2 SB-BGW-A(TAH-elam-insel9)# start SB-BGW-A(TAH-elam-insel9)# report ELAM not triggered yet on slot - 1, asic - 0, slice - 0 HEAVENLY ELAM REPORT SUMMARY slot - 1, asic - 0, slice - 1 ============================ Incoming Interface: Eth1/3 Src Idx : 0x9, Src BD : 4108 Outgoing Interface Info: dmod 1, dpid 130 Dst Idx : 0xd, Dst BD : 4109 Packet Type: IPv4 Outer Dst IPv4 address: 192.168.13.3 !---NVE IP address of SB-BGW-B Outer Src IPv4 address: 192.168.3.2 Ver = 4, DSCP = 0, Don't Fragment = 0 Proto = 17, TTL = 254, More Fragments = 0 Hdr len = 20, Pkt len = 166, Checksum = 0xd546 !---134 byte VXLAN packet + 32 byte cloudsec header Pkt len = 166 Inner Payload Type: CE L4 Protocol : 17 L4 info not available Drop Info: ---------- LUA: LUB: LUC: LUD: Final Drops: !---To reach SB-BGW-B NVE IP traffic was sent out of Ethernet1/4 which is connected to SB-SPINE-A SB-BGW-A(TAH-elam-insel9)# show system internal ethpm info all | i i "dpid=130" IF_STATIC_INFO: port_name=Ethernet1/4,if_index:0x1a000600,ltl=6132,slot=0, nxos_port=12,dmod=1,dpid=130,unit=0,queue=65535,xbar_unitbmp=0x0,ns_pid=255,slice_num=1,port_on_slice=58,src_id=116 
SB-BGW-A(TAH-elam-insel9)# show cdp neighbors interface ethernet 1/4 Capability Codes: R - Router, T - Trans-Bridge, B - Source-Route-Bridge S - Switch, H - Host, I - IGMP, r - Repeater, V - VoIP-Phone, D - Remotely-Managed-Device, s - Supports-STP-Dispute Device-ID Local Intrfce Hldtme Capability Platform Port ID SB-SPINE-A(FDO22302CJ0) Eth1/4 131 R S s N9K-C9236C Eth1/34 Total entries displayed: 1 SB-BGW-A(TAH-elam-insel9)# show ip route 192.168.13.3 IP Route Table for VRF "default" '*' denotes best ucast next-hop '**' denotes best mcast next-hop '[x/y]' denotes [preference/metric] '%<string>' in via output denotes VRF <string> 192.168.13.3/32, ubest/mbest: 1/0 *via 192.168.11.5, Eth1/4, [110/6], 00:56:13, ospf-UNDERLAY, intra via 192.168.14.2, [200/0], 01:13:46, bgp-65002, internal, tag 65002 !---The device still have a route for SB-BGW-B NVE IP via SVI  
SB-BGW-A(TAH-elam-insel9)# show ip route 192.168.14.2 IP Route Table for VRF "default" '*' denotes best ucast next-hop '**' denotes best mcast next-hop '[x/y]' denotes [preference/metric] '%<string>' in via output denotes VRF <string> 192.168.14.2/32, ubest/mbest: 1/0, attached *via 192.168.14.2, Vlan3600, [250/0], 01:15:05, am SB-BGW-A(TAH-elam-insel9)# show ip arp 192.168.14.2 Flags: * - Adjacencies learnt on non-active FHRP router + - Adjacencies synced via CFSoE # - Adjacencies Throttled for Glean CP - Added via L2RIB, Control plane Adjacencies PS - Added via L2RIB, Peer Sync RO - Re-Originated Peer Sync Entry D - Static Adjacencies attached to down interface IP ARP Table Total number of entries: 1 Address Age MAC Address Interface Flags 192.168.14.2 00:00:13 ecce.1324.c803 Vlan3600 SB-BGW-A(TAH-elam-insel9)# show mac address-table address ecce.1324.c803 Legend: * - primary entry, G - Gateway MAC, (R) - Routed MAC, O - Overlay MAC age - seconds since last seen,+ - primary entry using vPC Peer-Link, (T) - True, (F) - False, C - ControlPlane MAC, ~ - vsan, (NA)- Not Applicable VLAN MAC Address Type age Secure NTFY Ports ---------+-----------------+--------+---------+------+----+------------------ G 3600 ecce.1324.c803 static - F F vPC Peer-Link(R) SB-BGW-A(TAH-elam-insel9)#  

从该输出中，您可以看到，根据路由表，Cloudsec流量通过接口Ethernet1/4转发到SB-BGW-B。根据Cisco Nexus 9000系列NX-OS VXLAN配置指南，版本10.3(x)准则和限制：

• 流向交换机的CloudSec流量必须通过DCI上行链路进入交换机。

根据同一指南的vPC边界网关对Cloudsec的支持部分，如果vPC BGW了解对等vPC BGW的PIP地址并在DCI端进行通告，则来自两个vPC BGW的BGP路径属性将相同。因此，DCI中间节点最终可以从不拥有PIP地址的vPC BGW中选择路径。在此场景中，MCT链路用于来自远程站点的加密流量。但是，在这种情况下，使用的是指向SPINE的接口，尽管如此，BGW也通过BackUp SVI具有OSPF邻接关系。

SB-BGW-A(TAH-elam-insel9)# show ip ospf neighbors OSPF Process ID UNDERLAY VRF default Total number of neighbors: 2 Neighbor ID Pri State Up Time Address Interface 192.168.12.4 1 FULL/ - 01:33:11 192.168.14.2 Vlan3600 192.168.12.2 1 FULL/ - 01:33:12 192.168.11.5 Eth1/4 SB-BGW-A(TAH-elam-insel9)#

问题的原因和修复

原因是SVI接口的OSPF开销。默认情况下，在NXOS上，自动成本参考带宽为40G。SVI接口的带宽为1Gbps，而物理接口的带宽为10Gbps：

SB-BGW-A(TAH-elam-insel9)# show ip ospf interface brief OSPF Process ID UNDERLAY VRF default Total number of interface: 5 Interface ID Area Cost State Neighbors Status Vlan3600 3 0.0.0.0 40 P2P 1 up <Output omitted> Eth1/4 5 0.0.0.0 1 P2P 1 up 

在这种情况下，对SVI成本进行管理更改可以解决此问题。必须在所有边界网关上进行调整。

SB-BGW-A(config)# int vlan 3600 SB-BGW-A(config-if)# ip ospf cost 1 SB-BGW-A(config-if)# sh ip route 192.168.13.3 IP Route Table for VRF "default" '*' denotes best ucast next-hop '**' denotes best mcast next-hop '[x/y]' denotes [preference/metric] '%<string>' in via output denotes VRF <string> 192.168.13.3/32, ubest/mbest: 1/0 *via 192.168.14.2, Vlan3600, [110/2], 00:00:08, ospf-UNDERLAY, intra via 192.168.14.2, [200/0], 01:34:07, bgp-65002, internal, tag 65002 SB-BGW-A(config-if)#   
!---The ping is started to work immediately
Request 1204 timed out Request 1205 timed out Request 1206 timed out 64 bytes from 10.100.20.10: icmp_seq=1207 ttl=254 time=1.476 ms 64 bytes from 10.100.20.10: icmp_seq=1208 ttl=254 time=5.371 ms 64 bytes from 10.100.20.10: icmp_seq=1209 ttl=254 time=5.972 ms 64 bytes from 10.100.20.10: icmp_seq=1210 ttl=254 time=1.466 ms 64 bytes from 10.100.20.10: icmp_seq=1211 ttl=254 time=2.972 ms 64 bytes from 10.100.20.10: icmp_seq=1212 ttl=254 time=4.582 ms 64 bytes from 10.100.20.10: icmp_seq=1213 ttl=254 time=1.434 ms 64 bytes from 10.100.20.10: icmp_seq=1214 ttl=254 time=4.486 ms 64 bytes from 10.100.20.10: icmp_seq=1215 ttl=254 time=2.743 ms 64 bytes from 10.100.20.10: icmp_seq=1216 ttl=254 time=1.469 ms 64 bytes from 10.100.20.10: icmp_seq=1217 ttl=254 time=7.322 ms 64 bytes from 10.100.20.10: icmp_seq=1218 ttl=254 time=1.532 ms 64 bytes from 10.100.20.10: icmp_seq=1219 ttl=254 time=1.438 ms 64 bytes from 10.100.20.10: icmp_seq=1220 ttl=254 time=7.122 ms 64 bytes from 10.100.20.10: icmp_seq=1221 ttl=254 time=1.344 ms 64 bytes from 10.100.20.10: icmp_seq=1222 ttl=254 time=1.63 ms 64 bytes from 10.100.20.10: icmp_seq=1223 ttl=254 time=6.133 ms 64 bytes from 10.100.20.10: icmp_seq=1224 ttl=254 time=1.455 ms 64 bytes from 10.100.20.10: icmp_seq=1225 ttl=254 time=3.221 ms 64 bytes from 10.100.20.10: icmp_seq=1226 ttl=254 time=4.435 ms 64 bytes from 10.100.20.10: icmp_seq=1227 ttl=254 time=1.463 ms 64 bytes from 10.100.20.10: icmp_seq=1228 ttl=254 time=5.14 ms 64 bytes from 10.100.20.10: icmp_seq=1229 ttl=254 time=2.796 ms 64 bytes from 10.100.20.10: icmp_seq=1230 ttl=254 time=1.49 ms 64 bytes from 10.100.20.10: icmp_seq=1231 ttl=254 time=6.707 ms 64 bytes from 10.100.20.10: icmp_seq=1232 ttl=254 time=1.447 ms 64 bytes from 10.100.20.10: icmp_seq=1233 ttl=254 time=1.285 ms 64 bytes from 10.100.20.10: icmp_seq=1234 ttl=254 time=7.097 ms 64 bytes from 10.100.20.10: icmp_seq=1235 ttl=254 time=1.295 ms 64 bytes from 10.100.20.10: icmp_seq=1236 ttl=254 time=0.916 ms 64 bytes from 10.100.20.10: icmp_seq=1237 ttl=254 time=6.24 ms 64 bytes from 10.100.20.10: icmp_seq=1238 ttl=254 time=1.439 ms 64 bytes from 10.100.20.10: icmp_seq=1239 ttl=254 time=2.739 ms 64 bytes from 10.100.20.10: icmp_seq=1240 ttl=254 time=4.477 ms 64 bytes from 10.100.20.10: icmp_seq=1241 ttl=254 time=1.431 ms 64 bytes from 10.100.20.10: icmp_seq=1242 ttl=254 time=5.372 ms 64 bytes from 10.100.20.10: icmp_seq=1243 ttl=254 time=3.119 ms 64 bytes from 10.100.20.10: icmp_seq=1244 ttl=254 time=1.504 ms 64 bytes from 10.100.20.10: icmp_seq=1245 ttl=254 time=6.909 ms 64 bytes from 10.100.20.10: icmp_seq=1246 ttl=254 time=1.498 ms 64 bytes from 10.100.20.10: icmp_seq=1247 ttl=254 time=1.454 ms 64 bytes from 10.100.20.10: icmp_seq=1248 ttl=254 time=6.701 ms 64 bytes from 10.100.20.10: icmp_seq=1249 ttl=254 time=1.441 ms 64 bytes from 10.100.20.10: icmp_seq=1250 ttl=254 time=1.888 ms 64 bytes from 10.100.20.10: icmp_seq=1251 ttl=254 time=6.052 ms 64 bytes from 10.100.20.10: icmp_seq=1252 ttl=254 time=1.469 ms 64 bytes from 10.100.20.10: icmp_seq=1253 ttl=254 time=3.61 ms 64 bytes from 10.100.20.10: icmp_seq=1254 ttl=254 time=4.213 ms 64 bytes from 10.100.20.10: icmp_seq=1255 ttl=254 time=1.276 ms 64 bytes from 10.100.20.10: icmp_seq=1256 ttl=254 time=5.712 ms 64 bytes from 10.100.20.10: icmp_seq=1257 ttl=254 time=2.299 ms 64 bytes from 10.100.20.10: icmp_seq=1258 ttl=254 time=1.417 ms 64 bytes from 10.100.20.10: icmp_seq=1259 ttl=254 time=7.159 ms 64 bytes from 10.100.20.10: icmp_seq=1260 ttl=254 time=1.538 ms 64 bytes from 10.100.20.10: icmp_seq=1261 ttl=254 time=1.629 ms 64 bytes from 10.100.20.10: icmp_seq=1262 ttl=254 time=7.892 ms 64 bytes from 10.100.20.10: icmp_seq=1263 ttl=254 time=1.495 ms 64 bytes from 10.100.20.10: icmp_seq=1264 ttl=254 time=2.792 ms ^C --- 10.100.20.10 ping statistics --- 1265 packets transmitted, 58 packets received, 95.42% packet loss round-trip min/avg/max = 0.916/3.31/7.892 ms SA-HOST-A#