08-06-2016 05:29 AM
Hi Everyone,
I am working with a Service provider.
Within my network i have observed a very wired scenario as MPLS L2VPN traffic not load balance between cisco IOS-XR.
TOPO:
CISCO7613 <----[10G]------> ASR9K(1) <------[10G]---------> ASR9K(2) <--------[3X1G links]----------> ASR9K(3)
The L2VPN is created between CISCO7613 & ASR9K(3) routers.
There are 3 links of 1Gig capacity between ASR9K(2) and ASR9K(3).
But out of these 3 links only one link is getting highly utilized.
The config of L2VPN is show below::
On CISCO7613
Current configuration : 307 bytes
!
interface GigabitEthernet10/10
no ip address
speed nonegotiate
mpls propagate-cos
no keepalive
service instance 2010 ethernet
encapsulation default
xconnect 202.123.37.49 2010 encapsulation mpls pw-class TEST
mtu 1500
!
end
pseudowire-class TEST
encapsulation mpls
control-word
!
#sh mpls l2transport vc 2010
Local intf Local circuit Dest address VC ID Status
------------- -------------------------- --------------- ---------- ----------
Gi10/10 Ethernet:2010 202.123.37.49 2010 UP
On ASR9K(3)::
interface GigabitEthernet0/0/1/10
l2transport
!
l2vpn
nsr
load-balancing flow src-dst-ip
pw-class MPLS
encapsulation mpls
control-word
transport-mode ethernet
!
xconnect group PRIME-CEN
p2p PRIME
interface GigabitEthernet0/0/1/10
neighbor ipv4 202.123.37.229 pw-id 2010
pw-class MPLS
!
#sh l2vpn xconnect
----------------------------------------------------------------------------------------
PRIME PRIME UP Gi0/0/1/10 UP 202.123.37.229 2010 UP
----------------------------------------------------------------------------------------
Those 3 links connected btw ASR9K(2) and ASR9K(3) are having mpls ldp relationship.
#sh mpls ldp neighbor XXX.123.37.49
Peer LDP Identifier: XXX.123.37.49:0
TCP connection: XXX.123.37.49:646 - XXX.123.47.223:65196
Graceful Restart: Yes (Reconnect Timeout: 120 sec, Recovery: 0 sec)
Session Holdtime: 180 sec
State: Oper; Msgs sent/rcvd: 94055/94036; Downstream-Unsolicited
Up time: 8w0d
LDP Discovery Sources:
GigabitEthernet0/4/1/5
Targeted Hello (XXX.123.47.223 -> XXX.123.37.49, active/passive)
GigabitEthernet0/1/0/16
GigabitEthernet0/4/1/6
Addresses bound to this peer:
172.16.252.49 182.79.204.98 182.79.204.114 182.79.211.42
182.79.214.33 182.79.215.10 182.79.215.94 ... so on ..
Out of these interfaces the Traffic is only increasing on interface GigabitEthernet0/4/1/5, while other interfaces are very less utilized.
I have already tried using control word as you can see above. Is there anything i am missing. Kindly suggest what we can do in this scenario.
09-12-2016 05:38 AM
Take a look at the FAT-PW session in this document, it may be helpful.
https://supportforums.cisco.com/document/111291/asr9000xr-load-balancing-architecture-and-characteristics#alt
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: