Per VRF TE is not working

[arun.mallappally123]

  Hi All,

  I had configured the per vrf TE configurtion but it is not working.Tunnel is coming up but BGP next hop is not changing to the one which was specified under the VRF configuration mode (bgp next-hop loopback) .

 I attached the network diagram and configration.Please help.

Thanks,

Arun Mohan

[Muhammad Thanveer]

Hello Arun,

Is there any specific reason for not enabling MPLS labeled switch path on the interfaces of core devices, i.e.., R2 to R3 and R5 then from R4 to R3 and R5

Regards

Thanveer

[Muhammad Thanveer]

This link would be helpful for you

https://sites.google.com/site/amitsciscozone/home/important-tips/traffic-engineering/per-vrf-mpls-te-tunnels

[ashirkar]

Hi Thanveer,

Nice to see you again ,Hope you are doing well!! ,yesterday , i was just going through his config and wanted to put in GNS3 but some reason i couldn't able to do so,but as per my observation his PE config hits well as he is doing per vrf MPLS TE not sure may be issue could be with GNS too.let me set up and check whats wrong here :-)

Hi Arun,

Check this thing when you do per vrf MPLS TE

1) no need LDP as you are using rsvp to allocate labels

2) enable mpls traffic-engg globally 

3) enable  mpls TE on interface

4) enable TE under IGP and configure TE router ID 

5) configure RSVP under interface 

6) configure bgp next-hop to loopback you want under vrf

7) configure TE tunnel with IP address ,destination ip and path optioin etc but important is enable mpls on tunnel interface 

8) last you need to do is put traffic into tunnel ,configure static route to global process for BGP next hop ip address to tunnel

example 

If you 2 PE ,PE1 and PE2

and on PE2 ,under vrf you mention bgp next-hop lo100 & lo100 is 1.1.1.1

then on PE1 you need to configure static route for 1.1.1.1

ip route 1.1.1.1 255.255.255.255 tun 1

I hope it was clear to you 

Regards,

Ashish  

[Muhammad Thanveer]

Hi Ashirkar thanks for the reply, i am doing well and thanks for asking friend.

I hope same with you.

Actually after seeing his post I started studying per vrf TE.

Your post is quite helpful dear. i have checked his config as per your post. I only found that he has not put redistribute connected in R4's address-family ipv4 vrf A else everything has matched as you said. Is that can be a issue?

 address-family ipv4 vrf A
  no synchronization
  redistribute ospf 100 vrf A
 exit-address-family

[ashirkar]

Hi Thanveer,

The problem in his config is "next-hop self" command configure under vpnv4 address-family. He need to remove that command so that he can able to see next hop changed to what he configured in VRF ,also arun need to enable  "mpls ldp discovery targeted-hello accept" command on both PE to accept targeted hello which is disable default on IOS platform.

Regards,

Ashish Shirkar

[Muhammad Thanveer]

Yes Ashirkar, That would be the issue.

So we need to use targetted hellos as devices are not directly connected and moreover bgp next hop is not the TE tunnel ip too.

As the bgp next-hop is said to be self for vpnv4 neighbor, it is not going to take the next hop mentioned in the VRF.

Is this the case?

[ashirkar]

Yep Thanveer,

Generally you need to unable next-hop-self on your ASBR device which is connected to other AS doing inter-AS kind of thing within AS no need

Regards,

Ashish