12-31-2012 07:27 PM - edited 03-07-2019 10:50 AM
Hi everybody
This is what my book says:
Ospf elects DR/bdr on NBMA int. Dynamic neighbor discovery is not permitted therefore we need to use " neighbor A.B.C.D " command.
I found something diffferent:
R1-----f0/0----sw------f0/0----R2
Above I changed the ospf network type from broadcast to NBMA on both routers:
R1
f0/0
200.200.200.1/24
ip ospf network non-broadcast
router ospf 1
network 200.200.200.0 0.0.0.255 area 0
R2;
f0/0
200.200.200.2/24
ip ospf network non-broadcast
router ospf 1
network 200.200.200.0 0.0.0.255 area 0
===================================================
In absence of neighbor command, no hellos are sent by both routers.
Next I configure " neighbor 200.200.200.2" on R1 only.
Strange things happens. As expected, R1 starts sending hellos to R2. R2 does not have any neighbor command but yet starts sending hellos once it verifies all the hellos' parameters received match its own. i.e all the timers , hello, dead interval etc.
I did another lab to make sure it is indeed the timers
R1f0/0--------sw------f0/0 R2
I only change f0/0 on R1 to default ospf network type i.e broadcast.
R1 starts sending hello to R2.
R2 generates a log message " mismatch hello parameters" ( default for NBMA hello 30sec, dead timer 120)
R1 's hello carries hello 10 sec, dead timer 40 sec.
But when I change the timers on R2 as:
ip ospf hello 10
ip ospf dead-interval 40.
R2 starts sending hello to R1 and neighbor relationship forms even though R2 is not configured with any " neighbor 200.200.200.1 " command.
It appears to me a router will form a neighbor relationship over a nbma int as long as the hello parameters received in hello match the hello parameters of its NBMA int without the need of " neighbor A.B.C.D" command .
I appreciate your input
Happy New Year !!!!
Solved! Go to Solution.
12-31-2012 08:05 PM
Hi,
Yes. As you might be aware, OSPF on receiving hello will just check hello, dead intervals, area, subnet (skip in some case), auth (if enabled). There is no field as such in hello packet to carry wht network type the interface is enabled with.
If you have different network type on same link but above parameters matches, neighborship may come up. But OSPF database will have differnt link type in Router LSA and so the topology will be incomplete which in turn will reflect as incomplete RIB.
HTH,
Nagendra
12-31-2012 08:05 PM
Hi,
Yes. As you might be aware, OSPF on receiving hello will just check hello, dead intervals, area, subnet (skip in some case), auth (if enabled). There is no field as such in hello packet to carry wht network type the interface is enabled with.
If you have different network type on same link but above parameters matches, neighborship may come up. But OSPF database will have differnt link type in Router LSA and so the topology will be incomplete which in turn will reflect as incomplete RIB.
HTH,
Nagendra
12-31-2012 11:49 PM
Thanks Nagendra
This is exactly I observed in my lab. Incomplete ospf database when there is inconssitency in router lsa.
Happy New year to you. This is my first post for 2013!!!
01-02-2013 02:31 AM
Dear nagendra,
u r right
If you have different network type on same link but above parameters matches, neighborship may come up. But OSPF database will have differnt link type in Router LSA and so the topology will be incomplete ospf database..
Dear sarahr202
neighbourship may b up bt lsa type 5 will nt match bcoz lsa type 5 external LSA - these LSAs contain information imported into OSPF from other routing processes. They are flooded to all areas unchanged(except stub areas).
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide