cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
775
Views
8
Helpful
9
Replies

eigrp not re-routing to alternate path

ricschaaf
Level 1
Level 1

Hello Experts,

I am trying to set up a ring in a lab but have one

issue with eigrp. (diagram attached)

When I pull any one of the links, the routing table

updates no problem to re-route traffic the other way.

But when I pull the power on one of the routers, it

doesn't re-route. For example, if I power down 152_Low

and look at the route table on THCC_1, I lose the

152 network.

ANy ideas? Thank you

Before power down of 152_Low

Gateway of last resort is not set

172.23.0.0/16 is variably subnetted, 2 subnets, 2 masks

C 172.23.31.0/26 is directly connected, Vlan1

D 172.23.30.0/24 [90/3328] via 172.23.31.10, 02:41:29, Vlan1

10.0.0.0/25 is subnetted, 6 subnets

D 10.90.152.128 [90/1792] via 10.90.152.2, 00:05:22, Port-channel1

D 10.90.153.128 [90/2304] via 10.90.152.2, 00:05:12, Port-channel1

D 10.90.154.128 [90/2560] via 10.90.152.2, 00:05:12, Port-channel1

C 10.90.152.0 is directly connected, Port-channel1

D 10.90.153.0 [90/2048] via 10.90.152.2, 00:05:12, Port-channel1

D 10.90.154.0 [90/2816] via 10.90.152.2, 00:05:12, Port-channel1

After power down 152_Low

Gateway of last resort is not set

172.23.0.0/16 is variably subnetted, 2 subnets, 2 masks

C 172.23.31.0/26 is directly connected, Vlan1

D 172.23.30.0/24 [90/3328] via 172.23.31.10, 02:43:01, Vlan1

10.0.0.0/25 is subnetted, 4 subnets

D 10.90.153.128 [90/3584] via 172.23.31.24, 00:00:04, Vlan1

D 10.90.154.128 [90/3328] via 172.23.31.24, 00:00:04, Vlan1

D 10.90.153.0 [90/3840] via 172.23.31.24, 00:00:03, Vlan1

D 10.90.154.0 [90/3072] via 172.23.31.24, 00:00:04, Vlan1

I can post configs of any of the routers.

These are all 3560's running in layer 3 mode.

9 Replies 9

chrihussey
VIP Alumni
VIP Alumni

By turning off the power to 152_Low, does this in turn cause all interfaces, even those that may be on THCC_1 and 152_High to drop also? If that is the case, then the network(s) should no longer be active and should be removed from the routing table.

I don't know if I'm missing something but that should be the case.

The interfaces on THCC_1 stay up (except for

the one to 152_low of course). 152_High becomes unreachable also when I power down 152_Low. Not sure why that is...

Ideally I'd like to have each routing table

hold two routes to each network in the table.

The only router that sort of does does it right now is 153_Low.

Gateway of last resort is not set

172.23.0.0/16 is variably subnetted, 2 subnets, 2 masks

D 172.23.31.0/26 [90/3584] via 10.90.153.130, 00:08:54, Port-channel1

[90/3584] via 10.90.153.1, 00:08:54, Vlan1

D 172.23.30.0/24 [90/4096] via 10.90.153.130, 00:08:54, Port-channel1

[90/4096] via 10.90.153.1, 00:08:54, Vlan1

10.0.0.0/25 is subnetted, 6 subnets

D 10.90.152.128 [90/3072] via 10.90.153.1, 00:08:56, Vlan1

C 10.90.153.128 is directly connected, Port-channel1

D 10.90.154.128 [90/1792] via 10.90.153.130, 02:14:18, Port-channel1

D 10.90.152.0 [90/3328] via 10.90.153.1, 00:08:54, Vlan1

C 10.90.153.0 is directly connected, Vlan1

D 10.90.154.0 [90/2048] via 10.90.153.130, 02:06:04, Port-channel1

If by powering off 152_Low all interfaces associated with the 152 network are in a down state on the connected devices then that network should not be in the routing table as it is no longer accessible.

If that is the case then all is working as it should. If this is not the case, then I'm missing something something that you need to explain.

It appears EIGRP does reconverge after 152_Low is powered down. Routing table on THCC_1 shows that network 10.90.153.128/25, which is connected to 152_High, is originally routed through 152_Low via port-channel1 gets rerouted via THCC_2(vlan1) when 152_Low was powered down.

Bear in mind EIGRP by default only installs the route with the best metric in the routing table and feasible successor routes remain in the topology table until the successor route becomes unavailable. If your concern is about seeing multiple routes to a network then you can tweak the metrics and or use 'variance' command to achieve that.

HTH

Sundar

I lose ping to the 152_High net when I power down 152_Low. I don't know why. I would think that I could still reach it by re-routing the other way.

I also wonder why some routers keep two routes to some nets in the table while others do not.

All routers have auto-sum off, and I currently don't have ip default-network set anywhere right now.

Thanks for your help!

I'm curious why in this configuration all routes would not show a feasible successor in the topology tables seeing as there's two ways to get to each destination. Only a few of them show some FS. The only one that shows an FS for every route is 153_Low.

EIGRP topology table will retain only the routes that meet the feasiblity condition. Is it possible the EIGRP metrics are way off between the successor route and feasible successor route leading to the feasibility condition not being met. If that's the case then you may have to tweak the bandwidth/delay or use offset lists to influence the metric.

HTH

Sundar

Looking at the network topology I don't know why feasibility condition would not be met. They are all high speed etherchannels around the ring.

Do you think this has something to do with split horizon and poison return not allowing feasible succesors (the other route around the ring) from being listed in all the tables?

Thanks, Ric

vinayrajkp
Level 1
Level 1

Notice that route to 10.90.153.0 is there on the THCC_1 even after power down of 152_Low.This proves that connectivity to 152_high is still there.

Most probably (like mentioned by chrihussey) the interface 10.90.152.130 must be going down when you power off 152_Low. You can confirm this by pinging 10.92.152.130 from 152_high itself. If you cannot ping then atleast layer 3 interface 10.90.152.130 goes down when you power it off.

-vinay

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:

Innovations in Cisco Full Stack Observability - A new webinar from Cisco