Understanding Fabric Edge Routing Table

[techno.it]

Hi Folks

Hope can someone help me understanding and demistying Fabric routing table

VRF routing table shows only connected devices and routes for other IP pools in same VRF and other hand Global routing table maintain default and RLOC routes learned LISP.

From the edge device, I cannot ping any host outside fabric via source SVI IP address but from host connected to same fabric edge there is reachability. Is that normal?

Appreciate any explanation and clarification

[jalejand]

It is expected to have close to no information about remote networks on the VRF RIB on Edge nodes. On VRFs, traffic is forwarded using LISP Map-cache entries, rather than "show ip route vrf A x.x.x.x.", you can verify the forwarding decision with "show ip cef vrf A x.x.x.x", which is the result of "show lisp instance-id xxxx ipv4 map-cache x.x.x.x", where "xxxx" is the instance-id attached to the VRF.

It is normal for an Edge node to not ping any external destination using an overlay interface, this because of the anycast gateway mechanism. Your SVI IP address is replicated across all the fabric edge nodes and  fabric borders. The packet will be likely able to be forwarded out to its destination, but the return traffic will be processed and consumed by borders.

[techno.it]

@jalejand 

Thank for clarifying this. Another concern that only one of the borders can ping external host, is that the same reason as above mentioned above?

[jalejand]

Hi, that is correct, when you have more than one border, if you ping to an external destination sourcing a loopback (referencing an overlay IP Pool), it is possible that the return traffic is received on the adjacent border rathen than the original, where it will be consumed. It is expected.