cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
843
Views
0
Helpful
2
Replies

l3vpn with RR on IOS (cannot ping)

suren130
Level 1
Level 1
I have the following L3VPN topology (https://imgur.com/a/B1bd94Z). VRF CUST-A is configured on the PE routers. I also have mp-bgp neighborship from R1/R2 towards RR (R3).
 
I am not able to ping from 9.9.9.9 towards 10.10.10.10
 
  1. R9 can see the loopback of R10 (10.10.10.10)
  2. ping 10.10.10.10 source 9.9.9.9 doesn't work
  3. sh ip route vrf CUST-A : https://pastebin.com/F79pF6KD
  4. How does R1 know that it should use the global routing table when it looks up the next hop 2.2.2.2 (loopback of R2) ?

The control plane seems to be fine as 10.10.10.10 can be seen from R9:

https://pastebin.com/EpKsyR5h

2 Replies 2

ngkin2010
Level 7
Level 7

Hi,

 

  1. R9 can see the loopback of R10 (10.10.10.10)

With the given information, please review the following possible cause of the issue:

 

 (i). MPLS LDP is not working correctly, please have a check on MPLS LDP neighborship (show mpls ldp neighbor) on P and PE equipment. 

 

(ii). You have show that CE1 have learnt 10.10.10.10/32, but how about CE2? Did CE2 correctly learnt 9.9.9.9/32?

 

  1. ping 10.10.10.10 source 9.9.9.9 doesn't work

Trace route would give you more information for troubleshoot.

 

  1. sh ip route vrf CUST-A : https://pastebin.com/F79pF6KD

Please check the 9.9.9.9/32 had learnt on PE2 and CE2 as well.

 

  1. How does R1 know that it should use the global routing table when it looks up the next hop 2.2.2.2 (loopback of R2) ?

 

Your PE2 (R2) had learnt the route and send to RR with next-hop-self (which should be default in VPNv4). Thus, the next-hop is overrided with its update-source address (which is the loopback 2.2.2.2). RR will not change the next-hop and send to PE1 (R1), so R1 knows should be use 2.2.2.2 as next-hop for prefix 10.10.10.10/32.

CCIE28
Level 1
Level 1
Can you run a traceroute from both sides, and confirm that labels are being imposed? Also can you show mpls forwarding on all 4 core routers?
As for the next hop, this is the only case of "automatic route leaking" if you want to call it. The router will always use the default route table to lookup the bgp next hop