Load Balancing in MPLS VPN BGP and EIGRP SOO
I have two questions regarding MPLS VPN
# QUESTION1: With Route Reflector in ISP Core and "maximum-paths" BGP configured at PE, I wish to achieve load balancing within the Core. However, I note the following:
## With "PE1#show ip cef vrf" command I can see that there are two paths to the customer CE2 network, but only one is being used as seen by traceroute from CE1 or with "PE1#show ip vrf exact-route". Why?
# QUESTION2: I have configured EIGRP SOO on PE2 and PE3 to avoid routing loop, but after using all the possible debug commands I don't see any "loop detected" message at PE2/PE3/CE2?
## I am sure EIGRP SOO is preventing the routing loop as I see that without SOO "PE#debug ip bgp vpnv4 unicast updates" shows a lot of messages indicating there is loop, but after configuring SOO, the network becomes stable and I don't see any BGP update messages
## But my only question is why i don't see any "loop detected" debug message
## I am using the debug commands on PE/CE: debug ip bgp vpnv4 unicast updates, debug ip eigrp vrf *
It will be great if you can help me out....
I have configured the network in GNS3 with IOS "Version 12.2(33)SRC6"
I have the network with following setup:
# All the interfaces have subnet of format: 10.0.$LocalRouterNo$RemoteRouterNo.$LocalRouterNo/24
# All routers have loopback lo0: $LocalRouterNo.0.0.1/32
# Customer Site1 CE1:
## R1/CE1 (AS1) - R2/PE1 (AS234) run EBGP
## R1/CE1 is advertising 1.0.0.1/32
## R2/PE1 has fa1/0 in VRF 1 with RD/RT
### RD 234:2
### RT import/export 234:2
### RT import 234:3
### RT import 234:4
# Customer Site2 CE2:
## R5/CE2 - R3/PE2 and R5/CE2 - R4/PE3 are in EIGRP AS345
## R5/CE2 is advertising 5.0.0.1/32
## R3/PE2 has fa1/1 in VRF 5 with RD/RT
### RD 234:3
### RT import/export 234:3
### RT import 234:2
### RT import 234:4
## R4/PE3 has fa1/1 in VRF 5 with RD/RT
### RD 234:4
### RT import/export 234:4
### RT import 234:2
### RT import 234:3
# ISP AS234
## R2/PE1, R3/PE2, R4/PE3 are in AS234
## R2/PE1 is RR Server for R3/PE2 & R4/PE3 Clients
## BGP is redistributed into EIGRP on R3 and R4
## EIGRP is redistributed into BGP with specific metric on R3 and R4
# Interfaces of PE at PE2/R3-CE2 is configured with sitemap 3 that sets SOO to 1:3
# Interfaces of PE at PE3/R4-CE2 is configured with sitemap 4 that sets SOO to 1:4
Configurations and the network GNS3 file have been attached or at https://www.dropbox.com/s/r998iffgg7luqo3/Lab11.zip
If this link is broken contact me please
Sample snippets showing load balancing not working:
<source>
! Two paths
R2#show ip cef vrf 1 5.0.0.1/32 detail
5.0.0.1/32, epoch 0, per-destination sharing
recursive via 3.0.0.1 label 22
nexthop 10.0.23.3 FastEthernet1/1
recursive via 4.0.0.1 label 20
nexthop 10.0.24.4 FastEthernet2/0
R2#
! But only one is used
R2#show ip cef vrf 1 exact-route 1.0.0.1 5.0.0.1
1.0.0.1 -> 5.0.0.1 => label 22 label implicit-null TAG adj out of FastEthernet1/1, addr 10.0.23.3
R2#
R1#traceroute 5.0.0.1 source lo0
Type escape sequence to abort.
Tracing the route to 5.0.0.1
1 10.0.12.2 76 msec 124 msec 252 msec
2 10.0.35.3 [AS 234] [MPLS: Label 22 Exp 0] 156 msec 72 msec 72 msec
3 10.0.35.5 [AS 234] 352 msec * 308 msec
R1#
</source>