Routing

Hi Experts, 

Can someone please help me understand if this is an expected behavior?

I am trying to set wide-metrics-only in the following topology:

R1----R2---R3

Being R1 on L2 and R3 on L1, of course R2 will on L1/L2.

When I set the  wide-metrics-only on R1 for L2, on R2 for L1 and L2 and R3 on L1, I am able to see routes leaking from L1 to L2 but no the opposite, when I looked at R3 routing table I am still receivig the default route 0.0.0.0/0 as an internal preference 15.

Why am I not receiving specifi prefixes from R1 via?

Please see the configuration, 

R1:

export EXPORT;
level 2 wide-metrics-only;
interface lt-0/0/0.0 {
level 1 disable;
}

from {
protocol direct;
route-filter 10.0.1.1/24 exact; ------ This prefix is the one I want to be received on R3
}
then accept;

R2:

[edit logical-systems R2]
emoraga@Cacique# show protocols isis
level 2 wide-metrics-only;
level 1 wide-metrics-only;
interface lt-0/0/0.1 {
level 1 disable;
}
interface lt-0/0/0.2 {
level 2 disable;
}

R3:

[edit logical-systems R3]
emoraga@Cacique# show protocols 

isis {
export EXPORT;
level 1 wide-metrics-only;
interface lt-0/0/0.3 {
level 2 disable;
}
}

[edit logical-systems R3]
emoraga@Cacique# show policy-options
policy-statement EXPORT {
from {
protocol direct;
route-filter 10.0.0.1/24 exact; ---- This is the prefix I am sending to R1.
}
then accept;
}

Which I do see on R1:

emoraga@Cacique# run show route 10.0.0.0 logical-system R1

inet.0: 7 destinations, 7 routes (7 active, 0 holddown, 0 hidden)
+ = Active Route, - = Last Active, * = Both

10.0.0.0/24 *[IS-IS/18] 00:06:15, metric 30
> to 172.16.0.1 via lt-0/0/0.0

but I am not seeing this on R3, what I see is the default route:

emoraga@Cacique# run show route 10.0.1.0 logical-system R3

inet.0: 6 destinations, 6 routes (6 active, 0 holddown, 0 hidden)
+ = Active Route, - = Last Active, * = Both

0.0.0.0/0 *[IS-IS/15] 01:05:49, metric 10
> to 172.16.0.2 via lt-0/0/0.3

Is this expected?

why is the leaking only happening from L1 to L2?

I would have expected to see both, the default and the specific routes.

Any clue?

Hello,

---

@emoraga wrote:

Hi Experts, 

 

Can someone please help me understand if this is an expected behavior?

 

I am trying to set wide-metrics-only in the following topology:

 

R1----R2---R3

 

Being R1 on L2 and R3 on L1, of course R2 will on L1/L2.

 

When I set the  wide-metrics-only on R1 for L2, on R2 for L1 and L2 and R3 on L1, I am able to see routes leaking from L1 to L2 but no the opposite, when I looked at R3 routing table I am still receivig the default route 0.0.0.0/0 as an internal preference 15.

 

Why am I not receiving specifi prefixes from R1 via?

<skip>

 

Any clue?

 

---

Of course.

Wide-metrics or no wide-metrics, ISIS L2 routes are not leaked to L1 by default.

You have to apply a policy on R2 to leak the L2 routes into L1.

This is working as designed, is a well-known JUNOS behaviour and is richly documented.

HTH

Thanks

Alex 

Hi Alex, 

Thanks for replying.

After posting the query I did find that it is working as designed.

Is there any specific reason why it works that way? does it have any logic or just due to desing? 

Thanks.

Hello,

---

@emoraga wrote:

Hi Alex, 

 

 

Thanks for replying.

 

After posting the query I did find that it is working as designed.

 

Is there any specific reason why it works that way? does it have any logic or just due to desing? 

 

Thanks.

---

Yes

RFC 1195 https://www.ietf.org/rfc/rfc1195.txt

- Level 1 intermediate systems -- these nodes route based on the ID
     portion of the ISO address. They route within an area. They
     recognize, based on the destination address in a packet, whether
     the destination is within the area. If so, they route towards
     the destination. If not, they route to the nearest level 2 router.
<skip>

A level 1 router routes as
   follows:

   - If a specified destination address matches an [IP address, subnet
     mask, metric] reachable within the area, the packet is routed via
     level 1 routing.

   - If a specified destination address does not match any [IP address,
     subnet mask, metric] combination listed as reachable within the
     area, the packet is routed towards the nearest level 2 router.

RFC 2966 http://www.ietf.org/rfc/rfc2966.txt

an L1L2 router should be manually configured with
   a set of prefixes that summarizes the IP prefixes reachable in that
   L1 area.  These summaries are injected into L2.  RFC 1195 specifies
   no further interactions between L1 and L2 for IPv4 prefixes.

I strongly suggest You read this excellent book by W.Goralski http://www.amazon.com/Juniper-Cisco-Routing-Protocols-Multivendor/dp/0471215929 

It will save You countless hours staring at the screen trying to figure out why the stuff happens and/or countless hours of arguing over RFC-(non)compliant nitty-gritty protocol details in forum posts 

HTH

Thanks
Alex

Hi,

as a summary, for IPv4 prefixes:

- by default no IPv4 prefixes are leaked from L2 to L1

- by default L1 routers install a last resort gateway to one L2 router (due to the presence of the Attached bit in the LSP sent by the the L2 router)

- by default only internal IPv4 prefixes are leaked from L1 to L2

All these defaults can of course be altered with routing policies.

Also when you use wide-metrics-only on your L1, you loose the ability to distinguish between internal and external IPv4 prefixes (because you're not sending the TLV #130) therefore all routes are considered internal and therefore all IPv4 routes are leaked to L2.

And it is better to mention that this behavior is slightly different with IPv6 prefixes.

--
If this post solves your problem, please mark this post as "Accepted Solution".
Kudos are appreciated

Hi there,

I don't understand. why route 10.0.0.0 has preference 18? I think it must have preference 165 because route  10.0.0.0 is route external , learn from protocol direct.