2

enter image description here

in this scenario R4 redistribute 8.0.0.0/8 network from an external source into EIGRP domain, R2 and R3 are doing mutual redistribution from OSPF into EIGRP and vice versa, so we have mutual redistribution on multiple devices, when R2 and R3 learn 8.0.0.0/8 network, they put it in their Routing table as EIGRP external route, the question is regarding mutual redistribution on both R2 and R3 they both must advertise that to R1 as OSPF external route, so R1 should have 2 routes toward 8.0.0.0/8 network in its Routing table, and also based on this event, R2 must learn 8.0.0.0/8 as E2 and also R3 must have that route as E2 too,which will result in having a loop but this is not happening, so what's the reason could be possibly?

Improve this question
1
  • Did any answer help you? if so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could provide and accept your own answer.
    – Ron Maupin ♦
    Commented Aug 12, 2017 at 5:26

2 Answers 2

Reset to default
2

Since you don't provide any configuration I can't give you any specific answer.

In general when you perform two-way redisitrbution between two network protocols at multiple points you tag the routes so that when it reaches another redistribution point it will be filtered and hence not redistributed back into the original routing domain.

R1 should have two routes for 8.0.0.0/8 and possibly both in the routing table if it's the same cost. R2 and R3 will only have one route each with next hop of R4 in their routing tables since EIGRP has Administrative Distance of 90 which is more believable than 110 of OSPF.

Edit: Two-way redistribution at multiple points is always a bit tricky to understad but I'm gonna try to explain what I think is happening.

R3 retrieves route with ad 170 from R4, redistributes into OSPF with ad 110. R1 receives a route to 8.0.0.0/8 with ad 110 and passes it to R2. R2 has now got 2 routes: ad 110 via ospf and ad 170 via eigrp, the ospf route will be inserted into routing table and then passed along to R4 vid ad 170 since it's external.

R4 retrieves the ad 170 advertisement from R2 but since it's has got a shorter path to 8.0.0.0/8 it will continue to advertise it's own path instead of the one coming from R2.

R2 retrieves route with ad 170 from R4 and route with ad 110 from R1 for network 8.0.0.0/8 and hence will insert route from R1 into routing table instead of the one from R4 and will not redistribute the eigrp route since it has already got a "better" route coming from R1.

Ofcourse it could be the other way around depending on who is first.

Improve this answer
5
  • since you did not read the question correctly, I believe you did not notice that routes are injected into EIGRP from an external source so on R2 and R3 those have AD of 170 and not 90!!! plus all the costs and metrics and configs here are equal but R1 do not install 2 routes toward 8.0.0.0 network and just got one!
    – mila sarosh
    Commented Aug 21, 2015 at 12:13
  • @milasarosh Sorry, I apparently missed that part. I've updated my answer now with what I believe is happening. You should still update your question with your config.
    – Jimmy
    Commented Aug 21, 2015 at 12:43
  • R2 and R3 both are doing mutual redistribution, which basically means both are injecting 8.0.0.0/8 into OSPF domain with AD110, so R1 should tell R2 that it has a route to 8.0.0.0/8 via OSPF from R3, and also R1 should tell R3 that it has a route to 8.0.0.0/8 via OSPF from R2! this supposed to create loop, but R1 installs only 1 route to 8.0.0.0 next-hop R2, and R3 installs 1 route to 8.0.0.0 next-hop R2, so no loop happens, and its wired that it does not cause loop in OSPF domain, here focus is just only on what happens inside OSPF domain, I did not provide config to keep it simple and clear
    – mila sarosh
    Commented Aug 21, 2015 at 12:53
  • @milasarosh as I stated in my answer I believe the one who is quickest will send it's route to R1 and the other one will receive that route and hence not send its own.
    – Jimmy
    Commented Aug 21, 2015 at 12:56
  • the answer is split-horizon rule built in to link-state protocols, and the way it treats external routes as distance vector, I found my answer here, learningnetwork.cisco.com/thread/72466, but thank you anyway.
    – mila sarosh
    Commented Aug 21, 2015 at 13:08
0

I found the answer to my question here Cisco Learning Network but do not know how to copy or use that stuff here in this forum!

Improve this answer
1

Not the answer you're looking for? Browse other questions tagged or ask your own question.