Why Cat6500 PE just can only show VPN label without IGP label ?

Answered Question
May 7th, 2008

Hi,there:

We have a couple Cat6500 as our MPLS/VPN P, PE routers. But from the following first command output, we just only can see one label (VPN lable) 339 imposed,can't see the IGP label imposed.

From the second command we can see both IGP and VPN labels (339,224 or 339,20)are imposed.

Is there anybody can point out why the first command just can only show the VPN label ?

We're using Sup720 supervisor and 12.2(18)SXF9 IOS.

---

6500PE#sh ip cef vrf XYZ 172.212.0.0 255.255.0.0 detail

172.212.0.0/16, version 321, epoch 0

0 packets, 0 bytes

tag information set, all rewrites owned

local tag: VPN-route-head

fast tag rewrite with

Recursive rewrite via 172.17.0.213/32, tags imposed {339}

via 172.17.0.213, 0 dependencies, recursive

next hop 192.168.131.128, TenGigabitEthernet2/5 via 172.17.0.213/32 (Default)

valid adjacency

tag rewrite with

Recursive rewrite via 172.17.0.213/32, tags imposed {339}

Recursive load sharing using 172.17.0.213/32.

-----

-----

-----

6500PE#sh mls cef vrf XYA 172.212.0.0

Codes: decap - Decapsulation, + - Push Label

Index Prefix Adjacency

108775 172.212.0.0/16 Te2/5 339(+),224(+) (Hash: 0001)

Te1/5 339(+),20(+) (Hash: 0002)

6500PE

I have this problem too.
0 votes
Correct Answer by mheusing about 8 years 7 months ago

Hi Jerry,

It is only possible to see both labels with "show ip cef ..." in one command, if there is no load sharing to the BGP next hop. If there is more than one equal cost path to the BGP next hop, you need two commands to see all involved labels, just as you have observed it.

In a 6500 your "show mls ... " command is the only way I know to see both labels in one output of a show commands.

In your case, "sh ip cef 172.17.0.213 255.255.255.255 detail" will only provide you with information about the LDP labels used to reach 172.17.0.213/32. The VPN labels have nothing to do with that.

In brief:

1) IF there is only one IGP path to the BGP next hop, then "show ip cef vrf ... " will show both labels, the LDP label and the VPN label.

2) IF there is more than one equal cost path to the BGP next hop, then "show ip cef vrf ... " will show the VPN label and point you to the BGP next hop: "recursive ...". In this case you need a second command "show ip cef detail" to see the LDP labels used.

As a quick check you could also use "traceroute vrf ... ", which should reveal the label stack used to forward packets towards the prefix.

Hope this helps! Please use the rating system.

Regards, Martin

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4 (1 ratings)
Loading.
mheusing Wed, 05/07/2008 - 23:40

Hi,

The observed output is due to load balancing in the MPLS core, i.e. there are several pathes to the BGP next hop known to CEF.

Have a look at the respective output statement:

--------

Recursive rewrite via 172.17.0.213/32, tags imposed {339}

Recursive load sharing using 172.17.0.213/32.

--------

To get the full label stack you need to check the CEF/LFIB entries for 172.17.0.213/32 with "show ip cef 172.17.0.213 255.255.255.255 detail" or with "show mpls forwarding-table", which will give you the values obtained by the output of your second command available on the 6500:

172.212.0.0/16

Te2/5 339(+),224(+) (Hash: 0001)

Te1/5 339(+),20(+) (Hash: 0002)

You should be able to see outgoing labels 20 and 224 for 172.17.0.213/32.

Hope this helps! Please use the rating system.

Regards, Martin

jerrytozhang Fri, 05/09/2008 - 06:43

Hi,Martin:

Thanks for your reply.

I tried these two commands, but just only could get IGP labels(20 and 224), no way to see the VPN label(339).

Any idea ?

Jerry

--

--

6500PE#sh ip cef 172.17.0.213 255.255.255.255 detail

172.17.0.213/32, version 1080, epoch 0

0 packets, 0 bytes

tag information set, shared, all rewrites owned

local tag: 171

via 192.168.130.128, TenGigabitEthernet1/5, 11 dependencies

traffic share 1

next hop 192.168.130.128, TenGigabitEthernet1/5

valid adjacency

tag rewrite with Te1/5, 192.168.130.128, tags imposed: {20}

via 192.168.131.128, TenGigabitEthernet2/5, 43 dependencies

traffic share 1

next hop 192.168.131.128, TenGigabitEthernet2/5

valid adjacency

tag rewrite with Te2/5, 192.168.131.128, tags imposed: {224}

0 packets, 0 bytes switched through the prefix

tmstats: external 0 packets, 0 bytes

internal 0 packets, 0 bytes

6500PE#sh

--

--

6500PE#show mpls forwarding-table 172.17.0.213 32 detail

Local Outgoing Prefix Bytes tag Outgoing Next Hop

tag tag or VC or Tunnel Id switched interface

171 20 172.17.0.213/32 0 Te1/5 192.168.130.128

MAC/Encaps=14/18, MRU=1548, Tag Stack{20}

001A6C9F0A44001CF9B3D0008847 00014000

No output feature configured

Per-destination load-sharing, slots: 0 2 4 6 8 10 12 14

224 172.17.0.213/32 0 Te2/5 192.168.131.128

MAC/Encaps=14/18, MRU=1548, Tag Stack{224}

001A6C9F12BC001CF9B3D0008847 000E0000

No output feature configured

Per-destination load-sharing, slots: 1 3 5 7 9 11 13 15

6500PE#

Correct Answer
mheusing Tue, 05/13/2008 - 20:46

Hi Jerry,

It is only possible to see both labels with "show ip cef ..." in one command, if there is no load sharing to the BGP next hop. If there is more than one equal cost path to the BGP next hop, you need two commands to see all involved labels, just as you have observed it.

In a 6500 your "show mls ... " command is the only way I know to see both labels in one output of a show commands.

In your case, "sh ip cef 172.17.0.213 255.255.255.255 detail" will only provide you with information about the LDP labels used to reach 172.17.0.213/32. The VPN labels have nothing to do with that.

In brief:

1) IF there is only one IGP path to the BGP next hop, then "show ip cef vrf ... " will show both labels, the LDP label and the VPN label.

2) IF there is more than one equal cost path to the BGP next hop, then "show ip cef vrf ... " will show the VPN label and point you to the BGP next hop: "recursive ...". In this case you need a second command "show ip cef detail" to see the LDP labels used.

As a quick check you could also use "traceroute vrf ... ", which should reveal the label stack used to forward packets towards the prefix.

Hope this helps! Please use the rating system.

Regards, Martin

Actions

This Discussion