MPLS VPN VRF route selection

[jain.manish94]

Hello Team, 

need your help to understand the route selection when we using MPLS and inside the MPLS there is OSPF protocol. I am getting confuse when doing traceroute from CE1 to CE2 another site and it is taking the path which i am not understanding. 

Please find my gns diagram. 

in this attached topology all are working fine there is no issue. but below is my question. 

when i am doing traceroute. from CE-1-GOA to CE-2-GOA Loopback ip address without source, it is taking other path but when using same traceroute with source it is taking another path. why here i am getting confused. 

CE1-GOA#traceroute 22.22.22.22
Type escape sequence to abort.
Tracing the route to 22.22.22.22
VRF info: (vrf in name/id, vrf out name/id)
1 192.168.2.254 120 msec 172 msec 216 msec
2 192.168.4.1 [MPLS: Labels 16/22 Exp 0] 784 msec 812 msec 900 msec
3 192.168.8.254 [MPLS: Label 22 Exp 0] 596 msec 728 msec 504 msec
4 192.168.8.1 688 msec 628 msec 628 msec

CE1-GOA#traceroute 22.22.22.22 source lo0
Type escape sequence to abort.
Tracing the route to 22.22.22.22
VRF info: (vrf in name/id, vrf out name/id)
1 192.168.2.254 228 msec 132 msec 212 msec
2 192.168.3.1 [MPLS: Labels 16/22 Exp 0] 752 msec 788 msec 588 msec
3 192.168.8.254 [MPLS: Label 22 Exp 0] 532 msec 484 msec 524 msec
4 192.168.8.1 576 msec 632 msec 896 msec

[MHM Cisco World]

I dont see any issue' 

Check

Show mpls forward 

See if you have load balance' what you see is that traffic is load balance between two label (via to link)

MHM

[jain.manish94]

did you check my gns design , i am not using any load balance. 

in my trace route output when using destination ip address without source lo0 it is taking another way. and when using source Lo0 it is taking another way. 

i just want to know why and how it is happening. 

[MHM Cisco World]

Can I see 

Show mpls forwarding 

In PE1 and PE2

MHM

[jain.manish94]

PE1#show mpls forwarding-table
Local Outgoing Prefix Bytes Label Outgoing Next Hop
Label Label or Tunnel Id Switched interface
16 16 20.20.20.20/32 0 Et3/0 192.168.3.1
16 20.20.20.20/32 0 Et3/5 192.168.4.1
17 Pop Label 4.4.4.4/32 0 Et3/0 192.168.3.1
18 Pop Label 192.168.6.0/24 0 Et3/5 192.168.4.1
19 Pop Label 192.168.5.0/24 0 Et3/0 192.168.3.1
20 Pop Label 5.5.5.5/32 0 Et3/5 192.168.4.1
21 No Label 2.2.2.2/32[V] 0 Et2/1 192.168.2.1
22 No Label 192.168.2.0/24[V] \
0 aggregate/GOA
23 No Label 1.1.1.1/32[V] 0 Et2/0 192.168.1.1

P2#show mpls forwarding-table
Local Outgoing Prefix Bytes Label Outgoing Next Hop
Label Label or Tunnel Id Switched interface
16 Pop Label 4.4.4.4/32 0 Et2/0 192.168.5.1
17 Pop Label 192.168.3.0/24 0 Et2/0 192.168.5.1
18 18 10.10.10.10/32 0 Et2/0 192.168.5.1
17 10.10.10.10/32 0 Et3/7 192.168.6.1
19 Pop Label 192.168.4.0/24 0 Et3/7 192.168.6.1
20 Pop Label 5.5.5.5/32 0 Et3/7 192.168.6.1
21 No Label 22.22.22.22/32[V] \
0 Et3/5 192.168.8.1
22 No Label 192.168.8.0/24[V] \
0 aggregate/GOA
23 No Label 11.11.11.11/32[V] \
0 Et3/0 192.168.7.1
P2#

[MHM Cisco World]

MPLS use two path
16 16 20.20.20.20/32 0 Et3/0 192.168.3.1
16 20.20.20.20/32 0 Et3/5 192.168.4.1

to be sure shut the 3/0 interface in PE1 and check the traffic, it must pass via 3/5 for two cases

MHM

[jain.manish94]

yes i shut the port e3/0 and now below are the output. 

CE1-GOA#traceroute 22.22.22.22 source lo0
Type escape sequence to abort.
Tracing the route to 22.22.22.22
VRF info: (vrf in name/id, vrf out name/id)
1 192.168.2.254 184 msec 104 msec 176 msec
2 192.168.4.1 [MPLS: Labels 16/21 Exp 0] 580 msec 624 msec 808 msec
3 192.168.8.254 [MPLS: Label 21 Exp 0] 716 msec 648 msec 364 msec
4 192.168.8.1 500 msec 560 msec 396 msec
CE1-GOA#traceroute 22.22.22.22
Type escape sequence to abort.
Tracing the route to 22.22.22.22
VRF info: (vrf in name/id, vrf out name/id)
1 192.168.2.254 128 msec 172 msec 144 msec
2 192.168.4.1 [MPLS: Labels 16/21 Exp 0] 420 msec 596 msec 612 msec
3 192.168.8.254 [MPLS: Label 21 Exp 0] 580 msec 508 msec 420 msec
4 192.168.8.1 748 msec 660 msec 812 msec

but still same question load balance is fine it is happening. 

but still confusion why when using source lo0 it is taking one path and without source lo0 it is taking another path to to understand this and clear my doubts. 

[jain.manish94]

see after turning on the e3/0 again same result.

CE1-GOA#traceroute 22.22.22.22
Type escape sequence to abort.
Tracing the route to 22.22.22.22
VRF info: (vrf in name/id, vrf out name/id)
1 192.168.2.254 188 msec 268 msec 176 msec
2 192.168.4.1 [MPLS: Labels 16/21 Exp 0] 732 msec 756 msec 580 msec
3 192.168.8.254 [MPLS: Label 21 Exp 0] 552 msec 396 msec 592 msec
4 192.168.8.1 832 msec 616 msec 692 msec
CE1-GOA#traceroute 22.22.22.22 source lo0
Type escape sequence to abort.
Tracing the route to 22.22.22.22
VRF info: (vrf in name/id, vrf out name/id)
1 192.168.2.254 100 msec 204 msec 204 msec
2 192.168.3.1 [MPLS: Labels 16/21 Exp 0] 580 msec 620 msec 644 msec
3 192.168.8.254 [MPLS: Label 21 Exp 0] 496 msec 648 msec 652 msec
4 192.168.8.1 744 msec 852 msec 692 msec

if this is using lable that also same in both the cases. but with source using 3.1 and without source lo0 using 4.1 , this is not clear and how it is happening , how to verify this so that i can clear my doubts. 

[MHM Cisco World]

the hash my friend use in load balance effect the selection of path. hash use Source/destination IP and some time L4 port to select the path, here you change the source of traffic one time you use LO and other the router by default use the link between CE and PE as source and that effect link selection. 

and for it is fine or not. sure it fine and healthy in network to load traffic between multi link (labels)

MHM

[jain.manish94]

Can I put my site A in ospf 100 and site B in ospf for same customer?

I did the same but the communication is not happening between my CE2 and
CE2-1.

Will I need to do any extra commands.

[jain.manish94]

Means to say 

I configured ce1 goa ospf 100

And CE2 Goa ospf 200

And now the communication is not happening between.

Will I need to run any commands or this will not work

[MHM Cisco World]

Different process not effect' in end mpls ibgp will advertise route as lsa5.

Did you clear bgp after you change the ospf process?

MHM

[jain.manish94]

On pe1 -- router bgp 500 -- redis ospf 100 

On PE2  -- router bgp 500 -- redi ospf 200

I did run these commands

Still communicating is not happening.

Any specific commands I need to run.

[MHM Cisco World]

Check below

MHM 

[jain.manish94]

No when I am simply distributed it is not working 

Iike this redistrubte ospf into bgp 

But but 

When using 

redistrubte ospf into bgp  match internal external 

It is start to work.

Not understanding why it is happening