not receiving OSPF external routes

[Aoi]

Hi
Could someone please help me solve a problem with an OSPF router that is receiving external routes(O E2) on different interfafces?

There are 4 routers in the lab.
R1 has two interfaces each connected to a different router (R2 and R3).
R2 and R3 are both connected to R4 each using a differnet interface.

  |----R2----|
R1                R4
  |----R3----|

Here, R1 is running only OSPF, and R2 and R3 are running OSPF and RIP. They all belong to the area 0. R4 is only running RIP and advertising its prefixes.
R2 and R3 are both redistributing OSPF to RIP, and RIP to OSPF.

My question is, would R1 be able to receive two routes with the same destination of R4 when all AD and metrics are equal? In other words, would it be possible to load balance on R1? 
The project I am working on just doesn't seem to work like that. It only keeps a single route (R1--R2--R4 or R1--R3--R4).

Sorry but i won't be able to share the configs for this.
Thank you.

[balaji.bandi]

in addition to @paul driver 

I was thinking when you redistribute each other as per your diagram (there will be a loop right ?)

[MHM Cisco World]

sure you can if 
R1-R2 path cost is equal to R1-R3 path cost then the R1 will add the two equal path to RIB 
OSPF support ECMP.

[Giuseppe Larosa]

Hello @Aoi ,

given your network scenario and your configuration:

>> Here, R1 is running only OSPF, and R2 and R3 are running OSPF and RIP. They all belong to the area 0. R4 is only running RIP and advertising its prefixes.
R2 and R3 are both redistributing OSPF to RIP, and RIP to OSPF.

You have a race condition : the first one between R2 and R3 that injects an OSPF external route in OSPF for each of the RIP learned prefixes will make R1 and the other ASBR to install an OSPF O E1 or OE2 that will be preferred over the RIP routes for its lower AD 110 < 120.

So in your scenario only one copy of the external routes will be seen in the OSPF domain.

The second ASBR cannot redistribute into OSPF RIP routes if they are not anymore in the IP routing table.

This is what you see in your tests

Hope to help

Giuseppe

[MHM Cisco World]

there are two issue here in your network, 
1- multi point redistribute, you redistribute RIP into OSPF and then OSPF into RIP, 
this since both routing protocol have different AD make one Router have two copy one from OSPF with AD=110 and other from RIP AD=120 , it will select OSPF even so it is not optimal path. 

2- I change AD and make it same but still R1 dont have multi path ??
Yes I run lab I see this case, I go back I found the issue, the issue is in R4, R4 advertise route (let assume it is LO connect to R4) to R2 and R3, 
the issue is if you do 
show ip ospf database external <<- in R2 and R3 you will see the router advertise this LO is one, either link R1-R2 or R1-R3 
this make 
R2 or R3 not advertise the route into R1 instead I will use path through R1 for LO connect to R4. 

so sorry that can not done.

[Aoi]

@paul driver @balaji.bandi @Giuseppe Larosa @MHM Cisco World 
Thank you all so much for your answers!
Please let me get my head around each and will reply back.

[Aoi]

Could i rephrase my question in a different way?
When I do "show ip ospf database external" on R1, I only see R2 as the "Advertising router" at this moment. But why am i not seeing R3 here as well? I was expecting to see two routers R2 and R3 because I have everything equeally set up.

@Giuseppe Larosa 
>The second ASBR cannot redistribute into OSPF RIP routes if they are not anymore in the IP routing table.
This i am not quite sure what you mean. Both ASBR R2 and R3 have RIP routes from R4 redistributed to the OSPF RIB as "redistributed". Why won't R2 and R3 both just send the redistributed routes to R1. Sorry if i am not seeing the problem here.

[MHM Cisco World]

Why R1 see one path through R2 or R3, I clarify this point in my previous comment.

[paul driver]

Hello

---

@Aoi wrote:

This i am not quite sure what you mean. Both ASBR R2 and R3 have RIP routes from R4 redistributed to the OSPF RIB as "redistributed". Why won't R2 and R3 both just send the redistributed routes to R1. Sorry if i am not seeing the problem here.

---

Prior to redistribution R2/R3 will have rip routes both pointing to R4 for its advertised routes

When the rip gets redistributed into ospf, whatever ospf rtr (r2/r3) performs the action first will then advertised R4s  now ospf routes to the other ospf rtrs and as per network administrative distance metrics ospf  AD (110) is preferred over rip AD (120)

so let say R2 wins the race of redistribution , it will advertise R4 routes to R1 who will advertise onwards to R3

R3 will see two paths for R4 routes, 
- ospf via R1 ad 110
 - rip via r4 ad 120

as a result only a single best path will be entered in the route table of R1/3
R2 will see R4 rip route directly via rip
R1 will see R4 routes via R2
R3 will see R4 routes via R1

now if you change the AD seed metric of RIP to be "better" than OSPF seed metric , then both R2/R3 should have direct rip routes to R4 and both should advertise them towards R1 , As such R1 should see two best paths (ECMP) and enter both redistributed R4 routes into its own route table.

[paul driver]

Hello
on the r2/r3 , try changing RIP seed metric to be more preferred over the ospf default seed metric of 110,

Example: R2/R3
conf t
router rip
distance 109

[Joseph W. Doherty]

You can easily run into issues when doing mutual redistribution (which I believe was mentioned skimming the earlier replies).

You can also easily run into issues between routing protocols that have a whole different ways of "looking" at routes and their metrics.  For one, if you're using RIP, it's a classful routing protocol while OSPF is classless.

If you doing this is Packet Tracer, it would be helpful to also post the PT file.