cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
644
Views
5
Helpful
4
Replies

Why a particular route path was taken

wilson_1234_2
Level 3
Level 3

We have HQ site and DR connected via DS3 as shown in the frawing

There are three VLANs bridged across the link, one of the VLANs has a 16 bit subnet mask.

Recently we had a problem with one of the hosts in HQ (10.1.11.1/16, after being replaced) was taking a different path to get to a remote host (shown as hanging off of MPLS cloud).

Normally, the host in HQ would go default gateway 10.1.1.1 and the next hop would be the ethernet interface of the local router in vlan 7 (10.7.7.254), then to MPLS cloud, then to remote host.

What was happening was the host went to the default gateway, but the next hop was the ethernet interface of the remote router in DR, vlan 8 (10.8.8.254), then to MPLS cloud, then to remote host.

At first I thought the default gateway had been incorrectly configured with the BVI in that site, but what it turned out to be was the subnet mask on host 10.1.11.1 was only 8 bits.

Once the mask was changed to 16 bit, the route path was correct.

What was causing the DR BVI to be more desirable as a next hop with an 8 bit mask?

4 Replies 4

jcozzupoli
Level 1
Level 1

Hi,

For what you're describing to happen, there must be a static route for the 10/8 network to be routed out the far end router instead of the closer rouetr and then into the MPLS cloud! Also, if there is no static route doing this, the far router must be advertising a route for the 10/8 to follow over the 10/16. With these types of issues, its always best to check all routers involved and think "why" would this path be taken over the other, and "who" is telling my host to go that way?

HTH

Joe.

There are no static routes, and both routers are advertising the same routes no matter what mask the host trying to get to those routes has configured.

Hello Richard,

a device with a wrong shorter subnet mask thinks that all ip addresses in net 10/8 are in a single broadcast domain and sends out an ARP request for that IP address.

If proxy ARP is enabled on your routers the first to answer is used.

Check with sh ip int bviX if proxy ARP is enabled on your routers

with the correct mask the host makes a single ARP request for its configured default gateway and gets an answer from that router only

Hope to help

Giuseppe

Thanks Giuseppe,

You are absolutely correct.

I knew that the ARP requests would go out in the 10/8 subnet, but could not reason what the deciding factor in the remote router answering consitently first was.

Proxy arp is enabled on both interfaces, local proxy arp is disabled.

I am sitting on the same core switch and the ethernet interface connecteing the bridge group of the local router is on the same switch, how can the remote router across a DS3 answer first?

Edit:

Not sure if I made that clear, on both routers:

proxy arp is enabled

local proxy arp is disabled

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:

Review Cisco Networking products for a $25 gift card