
OPERATIONAL DEFECT DATABASE
...

...
On BGW, the overlay default route is pointing correctly to the shared border NVE (loopback 1): N9k-test# sh ip route 0.0.0.0 vrf VRF IP Route Table for VRF ?VRF? '*' denotes best ucast next-hop '**' denotes best mcast next-hop '[x/y]' denotes [preference/metric] '%' in via output denotes VRF 0.0.0.0/0, ubest/mbest: 1/0 *via 10.10.1.5%default, [20/0], 01:06:09, bgp-1.1, external, tag 65530 (evpn) segid: 50000 tunnelid: 0xada0104 encap: VXLAN The underlay next-hop 10.10.1.5 (shared border) is reachable over Eth1/5: N9k-test# sh ip route 10.10.1.5 IP Route Table for VRF "default" '*' denotes best ucast next-hop '**' denotes best mcast next-hop '[x/y]' denotes [preference/metric] '%' in via output denotes VRF 10.218.1.4/32, ubest/mbest: 1/0 *via 10.10.0.4, Eth1/5, [110/4], 01:07:18, ospf-UNDERLAY, intra The problem is that no ARP request is triggered on Eth1/5 by the switch connected to the shared border even if it is receiving VXLAN traffic from the client that needs to be forwarded. As a result, that VXLAN traffic towards the shared-border is sent to destination mac 0000.0000.0000 over Eth1/5, and then dropped on the shared border.
No ARP entry exists for underlay nexthop.
Issue a ping on the BGW to the underlay next hop to trigger an ARP request. (Static ARP does not work as a workaround.)
Click on a version to see all relevant bugs
Cisco Integration
Learn more about where this data comes from
Bug Scrub Advisor
Streamline upgrades with automated vendor bug scrubs
BugZero Enterprise
Wish you caught this bug sooner? Get proactive today.