[j-nsp] OSPF Confusion

[Crist Clark]

This may be something simple, but I've been staring at it a while and am
confused now. I have a rather simple network,

    1    1

R1 --------- R2

 \           /20

  \10       /

   \       /

    \     /

     \   /

    10\ /20

      R3


The numbers are the OSPF metrics for each interface. The R1-R2 link is 10
Gbps. The interface metric is manually set to 1 on R1 and R2. The other two
links are both 1 Gbps media, but the R1-R3 is limited to 500 Mbps and the
R2-R3 to 100 Mbps by their respective carriers. (Well, not really. This is
a lab set up meant to simulate the real network.) I've used the 10 and 20
metrics on the interfaces as shown to tell OSPF something about that.


What I want to happen is all traffic from R3 to go to both R1 and R2 via
the R1-R3 link unless that link is down. With the costs configured as they
are, I would think it would do that, but it's not working for R3.


This is all a NSSA. R3 is distributing a default route into the area. Both
R1 and R2 are importing static routes into the area. The routing on R1 and
R2 works how I want. R2 sees R1 as the best next hop for the default and
all of R1's statics. R1 sees R3 as the default next hop and sees R2 as best
for all of R2's statics.


The problem is that R3 sees R2 as the best next hop for all of the statics
on R2. I don't understand why. The cost of the path from R3 to R2 is lowest
via R1, 11 vs. 20, right?


R3 is a EX4500/4200 chassis running 11.1R3.5. In trying to troubleshoot
this, I'm also a bit baffled by the router LSA that R3 is sending out to
the area. Here's the IP info for the links,


R1-R2:  160.33.151.85-160.33.151.86/30

R1-R3:  10.56.1.14-10.56.1.1/28

R2-R3:  10.56.1.22-10.56.1.17/29


But when I look at the LSA for itself in R3's database,


    OSPF database, Area 0.0.0.1

 Type       ID               Adv Rtr           Seq      Age  Opt  Cksum
 Len

Router  *10.56.1.1        10.56.1.1        0x80000030  2005  0x20 0x9786  48

  bits 0x2, link count 2

  id 10.56.1.1, data 10.56.1.1, Type Transit (2)

    Topology count: 0, Default metric: 10

  id 10.56.1.17, data 10.56.1.17, Type Transit (2)

    Topology count: 0, Default metric: 20

  Topology default (ID 0)

    Type: Transit, Node ID: 10.56.1.17

      Metric: 20, Bidirectional

    Type: Transit, Node ID: 10.56.1.1

      Metric: 10, Bidirectional

  Gen timer 00:09:48

  Aging timer 00:26:35

  Installed 00:33:25 ago, expires in 00:26:35, sent 00:33:23 ago

  Last changed 00:36:20 ago, Change count: 22, Ours

It looks like its listing itself in as its neighbors? Wha'?

The other devices' router LSAs look like I expect. BTW, the other two
routers are Palo Alto Networks firewalls.

Like I said, I'm probably missing something easy. Haven't done much OSPF in
JUNOS or tried much traffic shaping before.