cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
3405
Views
0
Helpful
8
Replies

MPLS full mesh

abob21
Level 1
Level 1

Hi All,

 

Seek your advise on this. I'm exploring mpls full mesh design 4 sites (HQ, BR1, BR2 & BR3) with two PE routers. HQ & BR1 are connecting to PE1 with "vrf HQ" & "vrf BR1" and BR2, BR3 to PE2 with "vrf BR2" & "vrf BR3" likewise. All PE-CE use ospf and bi-redistribution is done. RD & RT export & import to all vrfs vice versa.

 

Each CE routers able to learn all routes from the rest of CEs but only can able to ping btw CE which sharing same PE router. Mean HQ & BR1 can ping each ip address and same to BR2 & BR3 but not btw HQ/BR1 <--> BR2/BR3

 

HQ-CE#sh ip route ospf

O E2 20.20.20.0 [110/1] via 10.10.10.2, 00:28:18, FastEthernet1/0
30.0.0.0/30 is subnetted, 1 subnets
O E2 30.30.30.0 [110/1] via 10.10.10.2, 00:24:18, FastEthernet1/0
40.0.0.0/30 is subnetted, 1 subnets
O E2 40.40.40.0 [110/1] via 10.10.10.2, 00:24:17, FastEthernet1/0

 

BR1-CE#sh ip route ospf

O E2 10.10.10.0 [110/1] via 20.20.20.2, 00:29:25, FastEthernet0/0
30.0.0.0/30 is subnetted, 1 subnets
O E2 30.30.30.0 [110/1] via 20.20.20.2, 00:24:25, FastEthernet0/0
40.0.0.0/30 is subnetted, 1 subnets
O E2 40.40.40.0 [110/1] via 20.20.20.2, 00:24:25, FastEthernet0/0

 

BR2-CE#sh ip route ospf
20.0.0.0/30 is subnetted, 1 subnets
O E2 20.20.20.0 [110/1] via 30.30.30.2, 00:28:45, FastEthernet0/0
40.0.0.0/30 is subnetted, 1 subnets
O E2 40.40.40.0 [110/1] via 30.30.30.2, 00:28:03, FastEthernet0/0
10.0.0.0/30 is subnetted, 1 subnets
O E2 10.10.10.0 [110/1] via 30.30.30.2, 00:28:45, FastEthernet0/0

 

BR3-CE#sh ip route ospf
4.0.0.0/32 is subnetted, 1 subnets
O 4.4.4.4 [110/11] via 192.168.4.252, 07:19:03, FastEthernet0/1
20.0.0.0/30 is subnetted, 1 subnets
O E2 20.20.20.0 [110/1] via 40.40.40.2, 00:28:10, FastEthernet0/0
10.0.0.0/30 is subnetted, 1 subnets
O E2 10.10.10.0 [110/1] via 40.40.40.2, 00:28:10, FastEthernet0/0
30.0.0.0/30 is subnetted, 1 subnets
O E2 30.30.30.0 [110/1] via 40.40.40.2, 00:27:56, FastEthernet0/0

 

1 Accepted Solution

Accepted Solutions

Your PEs loopback interface use a /24 subnet mask. You should use a /32 for it to work properly. Otherwise, OSPF advertises a /32 for the loopback and LDP advertises a /24 for the label binding and this will break things.

 

By the way, to test your mpls path between two PEs, you should use commands such as "ping mpls" or "traceroute mpls" rather than using ping or traceroute. These commands are better suites for this purpose.

 

Regards,

Harold Ritter
Sr Technical Leader
CCIE 4168 (R&S, SP)
harold@cisco.com
México móvil: +52 1 55 8312 4915
Cisco México
Paseo de la Reforma 222
Piso 19
Cuauhtémoc, Juárez
Ciudad de México, 06600
México

View solution in original post

8 Replies 8

Harold Ritter
Level 12
Level 12

Hi, 

 

Are LDP sessions are up between PE1, PE2 and Prouter?

 

Regards,

Harold Ritter
Sr Technical Leader
CCIE 4168 (R&S, SP)
harold@cisco.com
México móvil: +52 1 55 8312 4915
Cisco México
Paseo de la Reforma 222
Piso 19
Cuauhtémoc, Juárez
Ciudad de México, 06600
México

Hi Harold,

 

Thank you for your reply. LDP neighbor are up between P and PE1 /PE2. Not direct neighbor btw PE1 <>PE2. I have attached P, PE1 & PE2 configuration for more details. 

 

Best regards,

Abob

In addition, labels are seen when tracing from one PE to another

 

PE1#traceroute 7.7.7.7
Tracing the route to 7.7.7.7

1 12.12.12.1 [MPLS: Label 16 Exp 0] 60 msec 52 msec 12 msec
2 13.13.13.2 44 msec * 28 msec

 

PE2#traceroute 6.6.6.6

Tracing the route to 6.6.6.6

1 13.13.13.1 [MPLS: Label 17 Exp 0] 32 msec 20 msec 20 msec
2 12.12.12.2 16 msec * 32 msec

The configurations do not include "mpls ip" on the interfaces between PE1, Prouter and PE2. This command is required to enable LDP.

 

Regards,

Harold Ritter
Sr Technical Leader
CCIE 4168 (R&S, SP)
harold@cisco.com
México móvil: +52 1 55 8312 4915
Cisco México
Paseo de la Reforma 222
Piso 19
Cuauhtémoc, Juárez
Ciudad de México, 06600
México

Hi Harold,

 

I used "mpls ldp autoconfig" under IGP protocol on all P, PE1 & PE2 routers.

 

router ospf 1
mpls ldp autoconfig
log-adjacency-changes

 

Best regards,

Your PEs loopback interface use a /24 subnet mask. You should use a /32 for it to work properly. Otherwise, OSPF advertises a /32 for the loopback and LDP advertises a /24 for the label binding and this will break things.

 

By the way, to test your mpls path between two PEs, you should use commands such as "ping mpls" or "traceroute mpls" rather than using ping or traceroute. These commands are better suites for this purpose.

 

Regards,

Harold Ritter
Sr Technical Leader
CCIE 4168 (R&S, SP)
harold@cisco.com
México móvil: +52 1 55 8312 4915
Cisco México
Paseo de la Reforma 222
Piso 19
Cuauhtémoc, Juárez
Ciudad de México, 06600
México

Hi Harold,

 

Firstly Thank You!!

By changing ospf loopback interface to /32 finally paid off :) Now all my CE routers are able to ping each other.

 

I know ldp session use loopback interface as source but really didn't think /32 play as vital role. Got to learn something new from it & appreciated. Thanks for extended mpls ping & trace commands as well.

 

Best regards,

Abob

Glad it helped.

 

You could technically get it to work with a /24, by forcing ospf to advertise a /24 instead of the /32 by default for the loopback interface. You would do this by configuring "ip ospf network point-to-point" under the loopback interface.

 

But it is recommended to advertise a /32 in the context of mpls.

 

Regards,

Harold Ritter
Sr Technical Leader
CCIE 4168 (R&S, SP)
harold@cisco.com
México móvil: +52 1 55 8312 4915
Cisco México
Paseo de la Reforma 222
Piso 19
Cuauhtémoc, Juárez
Ciudad de México, 06600
México