2911 Routing Problem

[mjones]

Hi all I have a problem I am stumped on. I have a 2911 router with the following setup:

Interface G0/0 points to the lan

172.17.96.250 255.255.255.0

Int G0/1 points to the wan

172.16.96.250 255.255.255.0

On the Lan side i have two servers both with 2 network cards...one is 172.17.96.3 and 10.144.0.1 255.248.0.0 (on its own lan side)

The other server is 172.17.96.1 and 10.96.0.1 255.248.0.0 on is Lan side.

My routes are:

0.0.0.0 0.0.0.0 172.16.96.251 (which is a cisco switch owned by my service provider that is directly connected to my router)

10.144.0.0 255.248.0.0 172.17.96.3

10.96.0.0 255.248.0.0 172.17.96.1

I have several other sites setup the same way. For example a router 172.17.16.250 contected to a server with 172.17.16.1 and 10.16.0.1 on the inside of that server.

Here is my issue:

From my router in question if i trace ip to 10.16.0.1 my first hop is my default route like I would expect (and almost all my sites work) but I have 3 in particular that dont send to the default route. 10.64.0.1, 10.80.0.1, and 10.88.0.1 (and any addresses that would fall under those subnets). They just dont send and I am not sure why. My service provider can ping those sites from their switch.

Any help will be greatly appreciated.

mark

[Reza Sharifi]

Mark,

Why do your server's NICs are in two different subnerts?  Usually if you have 2 NICs, you bundle them together using Etherchannel with LACP all in one subnet.  Also, can you provide a diagram on how everything connects together?

HTH

[mjones]

This is setup in a lan wan configuration. Originally the 10.x.x.x was nat'd but later switched to routed.

[cadet alain]

Hi,

can you do this:

logging buffered 7

logging buffered 10000

no service timestamp debug

access-list 199 permit icmp any any

do debug ip pack detail 199

do debug arp

do sh log

and post output.

Regards.

Alain

[ebarticel · Accepted solution]

Hi Mark,

I see a subnetting pattern for your 10.x.x.x network that is using a /13 subnet mask and those works they way you expect.

For the 3 subnets you mention 10.64.0.0, 10.80.0.0, 10.88.0.0 that you have problems, one is using a /12 mask and other 2 the regular /13 mask. I will check the "ip route" statements on those routers if they are using the correct mask.

Hope this helps

Eugen

[mjones]

This is now solved. Eugen put me on the right track but here is what was wrong. I looked at my router configs and server setups over and over and could not find anything wrong. I did realize i had overlooked 10.72.0.0 and it too was not reachable. After some more digging I discovered they were not reachable from multiple sites. I called my service provider back and after a some work on their part they fixed it. They told me originally we were not supposed to use the 10.x.x.x range but had agreed to let us use it, however they hadn't completely changed their acl's so somewhere in the configs somewhere (they didn't pinpoint it for me) those ranges were being discarded.

Thanks to all that replied

Mark