OSPF questions, batch 2

Topology info: RID = device number; ospf area 0; all routers have P2P and loop 0 (RID) in ospf.

image

 

Questions:

1) given the above snippets, why does R3 OSPF database show metric 10 to reach 4.4.4.4, whereas RIB show metric 2?

Having set OSPF cost on R4, link R4-R3 (gi1.34) to 10, 

 

  • as expected I see R4 reaching R3 loopback through R5
  • on R3 however, link R3-R4 is still preffered. Whats strange though is that OSPF database reflects the correct metric, but not RIB

 

on R3: show ip ospf database router 4.4.4.4

...


    Link connected to: a Stub Network

     (Link ID) Network/subnet number: 4.4.4.4

     (Link Data) Network Mask: 255.255.255.255

      Number of MTID metrics: 0

       TOS 0 Metrics: 1


    Link connected to: a Transit Network

     (Link ID) Designated Router address: 100.45.0.5

     (Link Data) Router Interface address: 100.45.0.4

      Number of MTID metrics: 0

       TOS 0 Metrics: 1


    Link connected to: a Transit Network

     (Link ID) Designated Router address: 100.34.0.4

     (Link Data) Router Interface address: 100.34.0.4

      Number of MTID metrics: 0

       TOS 0 Metrics: 10




R3#sh ip route 4.4.4.4

Routing entry for 4.4.4.4/32

  Known via "ospf 1", distance 110, metric 2, type intra area

  Last update from 100.34.0.4 on GigabitEthernet1.34, 00:03:08 ago

  Routing Descriptor Blocks:

  * 100.34.0.4, from 4.4.4.4, 00:03:08 ago, via GigabitEthernet1.34

      Route metric is 2, traffic share count is 1




2) on R4, setting max-metric router-lsa produces results as expected, however same issue as before - R3 ospf databse shows correct metric (65535), rib shows metric 2


OSPF,RIB have been refreshed so they show the newest metric.



3) quote: If cost is manually set on an interface, it effects inbound updates only

Can't say I understand this logic. The way I understand,  when router sets cost manually, its reflected in LSA L1, thus all neighbors should see the updated cost on the merit of LSA L1.



4) LSA L2: DR receives all link states from DROthers through LSA L1. Does it then use LSA L1 or L2 to propagate these link states to DROthers?

I initially thought its through L2, but LSA L2 contains only adjacency RIDs. Could you confirm it indeed just resends the L1 LSAs it receives.



5) quote: inter-area routes – summarized by ABR. If multiple ABR’s are within an area, only 1 of them announces the prefixes to the area

Likely the person meant NSSA L7-to-L5 translator, as it doesn't make sense for ABR to be doing that. Its false, right?

Comments

  • Hi daqua,

    the fact that you are changing locally on R4 the interface cost does not imply a possible recomputation cost on R3 database or RIB(at least in this case)

    The reason why you are not seeing metric changing on R3 is that OSPF, in order to compute the shortest path towards a destination, must first compute the shortest path to the node identifier.

    So each OSPF node identified by the link-node ID number set itself as the root of the bestpath calculation towards the nodes belonging to the same area. Once this occurs the interface cost metric are considered and evaluated in a cumulative fashion.

    In your example you are viewing the database from the perspective of R4, that's why you see a correct metric assigned of 10.


    on R3: show ip ospf database router 4.4.4.4

    ...


        Link connected to: a Stub Network

         (Link ID) Network/subnet number: 4.4.4.4

         (Link Data) Network Mask: 255.255.255.255

          Number of MTID metrics: 0

           TOS 0 Metrics: 1


        Link connected to: a Transit Network

         (Link ID) Designated Router address: 100.45.0.5

         (Link Data) Router Interface address: 100.45.0.4

          Number of MTID metrics: 0

           TOS 0 Metrics: 1


        Link connected to: a Transit Network

         (Link ID) Designated Router address: 100.34.0.4

         (Link Data) Router Interface address: 100.34.0.4

          Number of MTID metrics: 0

           TOS 0 Metrics: 10

     

     

    But then you check the RIB from the perspective of R3.


    R3#sh ip route 4.4.4.4

    Routing entry for 4.4.4.4/32

      Known via "ospf 1", distance 110, metric 2, type intra area

      Last update from 100.34.0.4 on GigabitEthernet1.34, 00:03:08 ago

      Routing Descriptor Blocks:

      * 100.34.0.4, from 4.4.4.4, 00:03:08 ago, via GigabitEthernet1.34

          Route metric is 2, traffic share count is 1

    So in this case you see the metric of 1 (Gig link between R3-R4) plus another metric of 1 (stub network on R4 for the loopback interface)

    result metric = 2

     

     

    try instead to visualize the ospf database or RIB accordingly so try to issue:

    show ip ospf database router 3.3.3.3 you will see that the picture will match the one of the RIB.

     

  • [email protected],

    following the correct logic, what would then be the approach to metric manipulation how I wanted - so that none of the routers take the R3-R4 path to reach R4?

    Also, any docs regarding this? Somehow cost manipulation is portrayed as being straightforward, but then nothing works.

  • following the correct logic, what would then be the approach to metric manipulation how I wanted - so that none of the routers take the R3-R4 path to reach R4?

    Assuming that you are leaving the default ospf cost on every interface of your network which, if you are using Gigabit ethernet interfaces, should be equal to 1, you need just to increase the cost on the R3 Gigabit interface that face R4 in order to make it less preferrable than the other paths to reach R4.

    In this case the value needed should be one more than the worse case cost value. Or alternatively you could just increase it to the maximum.

    Since the routers are connected with P2P interfaces (as you stated) in a daisy chain fashion the worse case cost should be from the perspective of R3 that should ignore its directly connected link in order to reach R4 loopback and follow the path towards R2.

Sign In or Register to comment.