cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
296
Views
0
Helpful
3
Replies

Load balacing among multiple equal paths from Catalyst 4000

asrkumar
Level 1
Level 1

Hi,

We have a two catalyst 4000 IOS ver 12.1(8a) connected to two 7204 router. The two routers and switches are connected in a full mesh configuration. To reach a destination x.x.x.x I can see four equal cost paths in the switch routing table. Now how can I do a load sharing/balacing among these four paths so that all the four paths are utilised efficiently. I am running OSPF in my catalyst switch. The two routers are running EIGRP.

Does it help if I configure CEF?

Thanks and Best Regards,

Raj.

3 Replies 3

Gilles Dufour
Cisco Employee
Cisco Employee

if you see 4 routes for each destination it means the router is already loadbalancing the traffic.

Enabling CEF will just help you select what type of loadbalancing you want.

We have per-packet or per-destination.

The command is 'ip load [per-packet|per-destination]'

Gilles.

Hi Giles,

Thanks for the info. Just to clarify some points.

I had that in my mind what you said "ip load [per-packet|per-destination]". I read a cisco doc which says perpacket load balancing is not possible in L3 switching. So does catalyst 4000 support per packet load balancing.

The reason I ask this because we have two links connected to two routers for destination X. I find all the traffic is passing via link-1. But link-2 is under utilized, I want to share the load across the two links.

I have captured the route for one particular subnet from both the switches.

Switch-1>sh ip rou 10.192.0.0 255.255.0.0

Routing entry for 10.192.0.0/16

Known via "ospf 123", distance 110, metric 20

Tag 888, type extern 2, forward metric 1

Last update from 10.197.1.18 on Vlan6, 4d08h ago

Routing Descriptor Blocks:

10.197.1.10, from 192.168.197.18, 4d08h ago, via Vlan5

Route metric is 20, traffic share count is 1

* 10.197.1.18, from 192.168.197.18, 4d08h ago, via Vlan6

Route metric is 20, traffic share count is 1

10.197.1.9, from 192.168.201.13, 4d08h ago, via Vlan5

Route metric is 20, traffic share count is 1

10.197.1.17, from 192.168.201.13, 4d08h ago, via Vlan6

Route metric is 20, traffic share count is 1

switch-2>sh ip rou 10.192.0.0 255.255.0.0

Routing entry for 10.192.0.0/16

Known via "ospf 123", distance 110, metric 20

Tag 888, type extern 2, forward metric 1

Last update from 10.197.1.18 on Vlan6, 4d08h ago

Routing Descriptor Blocks:

10.197.1.10, from 192.168.197.18, 4d08h ago, via Vlan5

Route metric is 20, traffic share count is 1

* 10.197.1.18, from 192.168.197.18, 4d08h ago, via Vlan6

Route metric is 20, traffic share count is 1

10.197.1.9, from 192.168.201.13, 4d08h ago, via Vlan5

Route metric is 20, traffic share count is 1

10.197.1.17, from 192.168.201.13, 4d08h ago, via Vlan6

Route metric is 20, traffic share count is 1

From the route statements we see both the active routes point to the Vlan6. Is it possible to have each switches goes to different vlans (5 &6).

If I configure Ip load per packet does it help? I guess I can configure this only on Vlan interface level. So do I have to configure for all Vlan's.

Thanks for your comments.

Best Regards,

Raj.

We are having the same problem with our 4006 switches w/ the layer 3 card. We see 4 equal cost routes(2 ethernets per router), however one router gets considerably more traffic than the other. We are currently doing process switching with the no ip route-cache.

Getting Started

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: