cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
468
Views
0
Helpful
3
Replies

Redistribution in NSSA Area - Against to Redistribution Rule

OBD
Level 1
Level 1

Hello Dear Community,

 I hope you are doing well!

I have a very weird situation which I couldn't explain it. As you know , if you want to redistribute routes into different protocol , it has to be in routing table from the particular protocol which you want to inject into other. 

OBD_0-1708716491898.png

In my case , I have 2 router in a NSSA area-1 in OSPF Process-1 . One of the routers also have Loopback0 which belongs to Router OSPF-2 Process with Area-1. To summary , one router is running 2 different OSPF process. I redistributed OSPF-2 into OSPF-1 which is NSSA , and I saw this Loopback0 which is directly connected to my router , redistributed as Type-7. But in the routing table Loopback0 seems as a directly connected, so router doesnt know it from Router OSPF-2. 

 So how can this Router advertised it into NSSA even Loopback0 has not been learnt by OSPF-2.  In the normal circumstances , it doesnt advertise it unless learn it from Router OSPF-2. But in NSSA , this strict rule is not applied somehow.

Could you please enlighten me ? 

OBD_1-1708716836808.png

OBD_2-1708716861677.png

OBD_3-1708716897922.png

OBD_4-1708716938169.png

 

 

 



 

2 Accepted Solutions

Accepted Solutions

The LO since it ditect connect it have lowest ad and router will show it this in RIB' the router select lowest AD in RIB.

So Q If it appear in router as direct that meaning it not inject into ospf ? No 

The LO can appesr in ospf database if we redistrubte or config it under ospf process.

MHM

View solution in original post

Giuseppe Larosa
Hall of Fame
Hall of Fame

Hello @OBD ,

>> But in the routing table Loopback0 seems as a directly connected, so router doesnt know it from Router OSPF-2. 

For connected routes the rule that you have mentioned is relaxed because connected routes have smaller AD then IGP and also because it is safe to redistribute them in another routing protocol for the fact they are local routes that can be checked.

In OSPF network

command allows to inject internal routes and it is better then using redistribute connected subnets that creates unnecessary external routes.

Another important point to be able to redistribute static routes in OSPF is that static route's next-hop must be learned via an internal route and not via an external route. The forwarding address has to be learned via an internal route otherwise the LSA type 5 is not generated.

This is a consistency check to avoid to create an external route LSA type 5 that would depend by another LSA type 5 the one describing the next hop .

Hope to help

Giuseppe

 

View solution in original post

3 Replies 3

The LO since it ditect connect it have lowest ad and router will show it this in RIB' the router select lowest AD in RIB.

So Q If it appear in router as direct that meaning it not inject into ospf ? No 

The LO can appesr in ospf database if we redistrubte or config it under ospf process.

MHM

Thank you so much for your answer.

I got it now. So for the directly connected interfaces which are enabled at IGP level can be advertised even they are not in the routing table from this particular IGP process. 

 Apparently . it is very enough to redistribute connected interfaces  while they are enabled at IGP level.

Thank you

Giuseppe Larosa
Hall of Fame
Hall of Fame

Hello @OBD ,

>> But in the routing table Loopback0 seems as a directly connected, so router doesnt know it from Router OSPF-2. 

For connected routes the rule that you have mentioned is relaxed because connected routes have smaller AD then IGP and also because it is safe to redistribute them in another routing protocol for the fact they are local routes that can be checked.

In OSPF network

command allows to inject internal routes and it is better then using redistribute connected subnets that creates unnecessary external routes.

Another important point to be able to redistribute static routes in OSPF is that static route's next-hop must be learned via an internal route and not via an external route. The forwarding address has to be learned via an internal route otherwise the LSA type 5 is not generated.

This is a consistency check to avoid to create an external route LSA type 5 that would depend by another LSA type 5 the one describing the next hop .

Hope to help

Giuseppe

 

Review Cisco Networking for a $25 gift card