Intra-area ISIS routes showing up as L2

Unanswered Question
Dec 11th, 2008

I've got Routers F, G, and H on a token-ring segment (10.4.4.0/24), over which they all participate in IS-IS. Router G also has a serial link to Router D over 215.0.0.12/252.

Each router has a NET of

49.0001.000x.0000.0000.00

Where x is the number of the router's letter (D=4; F=6; G=7; H=8). Also this is true of the token-ring interface addresses (F's tok0 is 10.4.4.6, etc...)

Router H's IP table looks like this:

Gateway of last resort is not set

===============

222.1.6.0/32 is subnetted, 1 subnets

O 222.1.6.1 [110/7] via 10.4.4.6, 00:00:13, TokenRing0

215.0.0.0/24 is variably subnetted, 3 subnets, 2 masks

i L2 215.0.0.14/32 [115/20] via 10.4.4.7, TokenRing0

i L2 215.0.0.13/32 [115/30] via 10.4.4.7, TokenRing0

O 215.0.0.12/30 [110/70] via 10.4.4.7, 00:00:13, TokenRing0

10.0.0.0/24 is subnetted, 1 subnets

C 10.4.4.0 is directly connected, TokenRing0

C 222.1.8.0/24 is directly connected, Loopback0

Router-H#

===============

Yes there's also OSPF in there routing some of these networks but there's no redistribution configured, and since every NET begins with 49.0001, the same area ID, why would these be seen as L2 routes?

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 3.3 (3 ratings)
Loading.
Giuseppe Larosa Fri, 12/12/2008 - 02:24

Hello Suey,

I would start by investigating the serial link

do

sh clns int serX/y on RG and RD

what type of adiacency do they have formed ?

second thought:

if 215.0.0.12/30 is the ip subnet of the serial link that we see an OSPF route the two /32 entries 215.0.0.13/32 and 215.0.0.14/32 are the ip addresses of the serial interfaces:

i L2 215.0.0.14/32 [115/20] this is on RG

i L2 215.0.0.13/32 [115/30] via 10.4.4.7, TokenRing0 this is on RD (highest metric)

among the several TLVs of IS-IS there is also one TLV 132 ip interface ip address that should be present in : hellos, LSP, L1 and L2.

serial interfaces have only one type of hellos (not distinct l1 and l2).

see RFC 1195 for IS-IS routing of IP.

TLV 132 is the ip address of the transmitting interface.

if on the serial link only a L2 adjacency is built what you see is correct.

check if you have any isis circuit-type level-2 on RD or RG

Hope to help

Giuseppe

CriscoSystems Fri, 12/12/2008 - 14:16

They each show

Number of active level-1 adjacencies: 1

!

Number of active level-2 adjacencies: 1

But still, all NET's contain the exact same area ID. I suppose the fact that I left the default routing level (1 and 2) allowed or caused all routers to think that 49.0001 is the backbone area?

Still though, I had thought any routing within one area is considered level 1 routing, even if it's the backbone area. Is that not true?

I looked up TLV132 in RFC 1195, but I'm afraid I don't quite see the relevance of it.

The 10.4.4.0/24 subnet is in D's table as level 1.

Giuseppe Larosa Sat, 12/13/2008 - 04:09

Hello Stuey,

>> I looked up TLV132 in RFC 1195, but I'm afraid I don't quite see the relevance of it

the misbehaving routes are actually the ip addresses of the routers on the serial link they are /32. Aren't they ?

From this I thought of TLV 132 as the possible origin of these /32 routes then I added that on a serial link there is a single unified hello instead of two distinct hellos.

I think the problem is only of these two "prefixes" because the real ip subnet is /30.

But some prefixes are masked by lower AD of OSPF.

As a confirmation of this comes your last note:

The 10.4.4.0/24 subnet is in D's table as level 1

So if IP prefixes downstream router D and advertised only by IS-IS are correctly seen as i L1 in routing table you are fine.

These two /32 can be seen as a minor issue

Hope to help

Giuseppe

CriscoSystems Mon, 12/15/2008 - 11:55

Hmmm, you know this brings us back to something I've wondered for a long time, which is why the heck does a 32-bit "prefix" even show up in any ol' routing table in the first place?

Isn't it just supposed to list routes, rather than hosts?

Giuseppe Larosa Mon, 12/15/2008 - 21:41

Hello Stuey,

I agree for this reason I thought of an effect of that TLV 132.

If you were using another routing protocol those /32 entries wouldn't be there I suppose

Hope to help

Giuseppe

Actions

This Discussion