In the EOS user manual section about EVPN, there is configuration example portion that i dont understand.
Can you please explain the purpose of that interface?
Hi Leo, this interface is configured to assign an IP address on the switch to the VRF, this is then used as the source IP for an pings originating from switch when verifying/testing connectivity in the VRF.
I guessed that but why to create that vlan instead of any vlan inside the same VRF? Is it an undocumented best practice or so?
Actually, i think i found my answer. That interface is for the iBGP IP-VRF session between leaves of the same MLAG.
Expanding a bit on what Alex mentioned.
When Overlay anycast gateway is defined using “ip address virtual” option, there’s no unique address per SVI. So, it’s possible that the response to any traffic such as ping, trace route etc., sourced from the VTEP lands on a “wrong” VTEP since they all share a common anycast gateway IP. To workaround this, EOS provides SNAT feature (enabled by default) that picks the highest IP in the VRF to translate the source IP of the packets sourced from the switch.
Vlan1111 serves this purpose in ‘tenant-a’ VRF. You’ll have one such IP in each non-default Overlay VRF.
Post your Answer
You must be logged in to post an answer.