Routing Protocols

[abdul.qadir5001]

I have an issue while configuring MPLS L3 VPN that customers of CBTSite1 on one PE and Customer of CBTSite2 on remote PE cannot communicate with each other.

One site one EIGRP is running and site 2 EBGP is running. Routes are reachable on respective PE but not able to reach on other PE. MPLS and OSPF is running on provider routers. 

Configurations are below. Kindly guide plz to resolve the issue. Thanks

[Georg Pauwen]

Hello,

can R1 and R5 reach each other ?

[abdul.qadir5001]

Yes R1 and R5 can ping each other but ACME site1 cannot ping ACME site2. VRF table on PE cannot contain routes of remote site routes. Kindly guide configurations and topology are provided earlier. Thanks

R1#ping 5.5.5.5 source loopback 0

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 5.5.5.5, timeout is 2 seconds:
Packet sent with a source address of 1.1.1.1
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 108/134/172 ms

R1#sh bgp vpnv4 unicast vrf 101:ACME

Route Distinguisher: 1.1.1.1:1 (default for vrf 101:ACME)
*> 172.16.101.0/30 0.0.0.0 0 32768 ?
*> 172.16.111.2/32 172.16.101.2 11 32768 ?

R1#sh bgp vpnv4 unicast vrf 102:CBT

Route Distinguisher: 1.1.1.1:2 (default for vrf 102:CBT)
*> 172.16.101.0/30 0.0.0.0 0 32768 ?
*> 172.16.222.2/32 172.16.101.2 409600 32768 ?

[Georg Pauwen]

Hello,

my guess is there is something wrong with the redistribution from VRF to global and/or vice versa. I need to lab this, it will be a while, unless somebody else comes up with a solution first, obviously. Will get back with you...

[abdul.qadir5001]

Dear Georg Pauwen,

Yes sure please. But i have also the file.

[Georg Pauwen]

Hello,

which program are you using ? Either way, post the file (if that is Packet Tracer, you need to zip it)...

[paul driver]

Hello

Your PE configuration doesn't looks correct- especially the redistribution between your PE and CE rtrs

Example:
router eigrp 1 <-------------
no auto-summary <--- summerization is enabled , 
redistribute bgp 65535 metric 1 1 1 1 1  <------------- no redistribution of bgp

router ospf 101 vrf 101:ACME
log-adjacency-changes
redistribute bgp 65535 subnets <------------- no redistribution of bgp

router bgp 65535

address-family ipv4 vrf 102:CBT
redistribute connected <-----------------not required
redistribute static <-----------------not required
redistribute eigrp 102 <---------- shouldn't this be redistribute 1 ?

address-family ipv4 vrf 101:ACME
redistribute connected <-----------------not required
redistribute static <-----------------not required
redistribute ospf 101 vrf 101:ACME1 <-------------shouldn't this be  redistribute ospf 1 ? 

[abdul.qadir5001]

Dear Paul Driver,

Extra configurations are removed now and redistribution enabled on both PEs.

Also  that 

router bgp 65535

address-family ipv4 vrf 102:CBT
redistribute eigrp 102 <---------- shouldn't this be redistribute 1 (102 is EIGRP AS number running on customer side)

address-family ipv4 vrf 101:ACME
redistribute ospf 101 vrf 101:ACME <-------------shouldn't this be redistribute ospf 1 (101 is OSPF AS number running on customer side)

so no need to change as numbers. But still after changing the config routes are not exchanging between customer site 1 and site 2. If you have any query further kindly let me know plz but kindly guide to resolve the issue.

Latest config are attached. Thanks

[paul driver · Accepted solution]

Hello
Can you please remove the redistribution from the opsf mpls backbone process, that was my mistake misread you OP also import RTs in both vrfs and test again

router ospf 1
log-adjacency-changes
redistribute bgp 65535 subnets
network 0.0.0.0 255.255.255.255 area 0

ip vrf 101:ACME
rd 1.1.1.1:1
route-target import 1.1.1.1:102

ip vrf 102:CBT
rd 1.1.1.1:2
route-target import 1.1.1.1:102

[abdul.qadir5001]

Dear Paul Driver,

Its awesome. Very nice. It worked.Thank you very uch. Now CBT site 1 and 2 can reach each other.

CBT2Site2#ping 172.16.101.2

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 172.16.101.2, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 148/168/188 ms

CBT2Site2#sh ip route

Gateway of last resort is not set

172.16.0.0/16 is variably subnetted, 2 subnets, 2 masks
B 172.16.222.2/32 [20/0] via 192.168.202.1, 00:21:24
B 172.16.101.0/30 [20/0] via 192.168.202.1, 00:21:24
192.168.202.0/30 is subnetted, 1 subnets
C 192.168.202.0 is directly connected, Ethernet2/1
192.168.222.0/32 is subnetted, 1 subnets
C 192.168.222.2 is directly connected, Loopback0

Q1) But some issue in ACME site 1 and site 2. They cannot ping each other.

Q2) Second Question is below:

ip vrf 101:ACME
rd 1.1.1.1:1
route-target import 1.1.1.1:102

ip vrf 102:CBT
rd 1.1.1.1:2
route-target import 1.1.1.1:102 (Why not we used  this 1.1.1.1:101 in CBT vrf )

Q3) Also if we use same RD and RT value then no need of extra command that you mentioned(route-target import 1.1.1.1:102). Why?

For example vrf "Atlas-Honda" RD and RT same value 38264:5142 we have no need of extra command.

ip vrf Atlas-Honda
rd 38264:5142
route-target export 38264:5142
route-target import 38264:5142

[paul driver · Accepted solution]

Hello

seems like it isn't my day for revising configuration looks like another typo in my part so apologies - i am also looking at his from my phone with no glasses ! which doesn't help !

Just make sure the RTs your importing have parity to what you want to receive from the other mpls sites that are exporting 

ip vrf 101:ACME

route-target import 1.1.1.1:101

ip vrf 102:CBT

route-target import 1.1.1.1:102

[abdul.qadir5001]

Now the issue is being resolved by importing RT. Thanks Paul Driver.