Variance in command in EIGRP

Unanswered Question
Feb 12th, 2007

Question for the EIGRP gurus out there.. ;-)

I am working on a unequal path load balancing config for a customer consisting of 4 routers ..

I am trying to load balance destinations Networks on all 4 routers so looking at the "show ip ei topology all" reveals two listing to all destination Networks ..which is what I want. However, only one route ever makes it into the routing table even with using the "variance" command...

I know i havent pasted any configs but, does anything spring to mind at first bat?

There dont seem to be any feasbily condition issues either.. any ideas appreciated..

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 3.5 (2 ratings)
Loading.
Edison Ortiz Mon, 02/12/2007 - 13:36

Without seeing configs, the only thing I can suggest is to keep increasing the variance until the lesser route shows up.

mheusinger Mon, 02/12/2007 - 13:37

Have you set variance to a value high enough? It works like a multiplicator, i.e. it is a factor multiplied to the best metric at the time. Any entry into the routing table must have a metric better than the result.

Without more details this would be my only advice.

regards, Martin

sundar.palaniappan Mon, 02/12/2007 - 14:04

As Martin indicated the variance (value) you are using is probably not high enough to install the alternate route in the routing table. An example would make it easy to understand. Let's say you have the following route with 2 paths in the topology table. EIGRP, by default, would install only the best route, with metric of 5, in the routing table. However, if you set the variance to 3 (5*3), the next best route, with a metric of 14, would be installed as well. The goal here is to set the variance (#) at least to a certain level that would accommodate the alternate route(s).

172.16.1.0/24

path 1 - metric 5

path 2 - metric 14

router eigrp 1

variance 3

HTH

Sundar

Richard Burts Mon, 02/12/2007 - 14:17

Olumide has not given us much to work with. One thing that does come to mind for me when someone mentions difficulty with variance in EIGRP is that the inferior route must qualify as a feasible successor for variance to be able to put it into the routing table. Olumide says that he believes that there are no feasibility issues. But I would like him to confirm that the second route does qualify as a feasible successor. Posting the entries from the EIGRP topology table would be even better.

HTH

Rick

olumidek Tue, 02/13/2007 - 06:31

Thank you all for your responses..

I now have a snipet of the configs..This is identical for all 4 routers..

router eigrp 10

variance 4

redistribute static metric 100000 100 255 1 1500

network 10.192.2.132 0.0.0.0

network 10.192.2.249 0.0.0.0

no auto-summary

eigrp router-id 1.1.1.1

!

ip route 0.0.0.0 0.0.0.0 FastEthernet1/1 10.192.2.129

sh ip ei topology all

P 0.0.0.0/0, 1 successors, FD is 51200, serno 86

via Rstatic (51200/0)

via 10.192.2.133 (53760/51200), FastEthernet1/1

P 10.192.2.248/30, 1 successors, FD is 28160, serno 63

via Connected, FastEthernet1/0

P 10.192.2.252/30, 1 successors, FD is 30720, serno 114

via 10.192.2.133 (30720/28160), FastEthernet1/1

via 10.192.2.250 (33280/30720), FastEthernet1/0

P 10.192.2.128/27, 1 successors, FD is 28160, serno 2

via Connected, FastEthernet1/1

P 100.10.0.0/16, 1 successors, FD is 30720, serno 108

via 10.192.2.250 (30720/28160), FastEthernet1/0

via 10.192.2.133 (33280/30720), FastEthernet1/1

P 101.10.0.0/16, 1 successors, FD is 53760, serno 109

via 10.192.2.250 (53760/51200), FastEthernet1/0

via 10.192.2.133 (56320/53760), FastEthernet1/1

r1#sho ip route ei

100.0.0.0/16 is subnetted, 1 subnets

D 100.10.0.0 [90/30720] via 10.192.2.250, 00:00:34, FastEthernet1/0

101.0.0.0/16 is subnetted, 1 subnets

D EX 101.10.0.0 [170/53760] via 10.192.2.250, 00:00:34, FastEthernet1/0

10.0.0.0/8 is variably subnetted, 3 subnets, 2 masks

D 10.192.2.252/30 [90/30720] via 10.192.2.133, 00:00:34, FastEthernet1/1

r1#

see anything strange?

Many Thanks..

Richard Burts Tue, 02/13/2007 - 08:16

Olumide

Thanks for posting the additional information. It does demonstrate that the issue is what I talked about in my previous post - it is a question of feasible distance and feasible successor. The information from the EIGRP topology table shows that each of the destination routes has a successor and no feasible successor. For variance to be able to put the second route into the routing table the advertised distance for the second route must be less than the feasible distance for the primary route. And they are not less for your routes. Using the first set of routes as an example:

P 10.192.2.252/30, 1 successors, FD is 30720, serno 114

via 10.192.2.133 (30720/28160), FastEthernet1/1

via 10.192.2.250 (33280/30720), FastEthernet1/0

we see that the feasible distance is 30720 and the advertised distance for the second route is 30720. Because the advertised distance is not less than the feasible distance then variance can not put the second route into the routing table.

I am not sure what you are indicating when you ask: see anything strange? Perhaps you can clarify what you are asking?

HTH

Rick

olumidek Wed, 02/14/2007 - 04:15

Rick.. I now see the error of my ways and why i tend to stay clear of EIGRP ;-)

Guess Id better start playing with the metrics to try and get this load-balancing scenario to work ;-)

many thanks to all for your responses!

Actions

This Discussion