Skip navigation
Cisco Learning Home > CCIE Routing and Switching Study Group > Discussions
This Question is Answered 1 Helpful Answer available (2 pts)
4252 Views 49 Replies Latest reply: Apr 6, 2012 1:15 PM by gregsykes RSS Go to original post 1 2 3 4 Previous Next

Currently Being Moderated
  • aaceituno 242 posts since
    Feb 18, 2009

    I will try tomorrow, right now i dont have time to deal with it.




  • aniata 1 posts since
    Jun 30, 2010



    while redistributing OSPF in the BGP, use external keyword. It will work; try and let us know.


    router bgp 1

    redistribute ospf 1 match external

  • Hey Greg,


    I have a question about R1 on your attached packet tracer. I want to know if the 10.x.x.x network  was supposed to be in  bgp or ospf domain; and/or running both bgp &OSPF. I'm asking this because the 10.x.x.x network command is on both ospf and bgp. Please, clarify



  • Not sure if you guys were able to ping R1 or R0 from R2 & R3. I wasn't able to ping.

    There was no route from R2 & R3 to the network (Bgp domain).


    = NO need to redistribute bgp to ospf nor vise versa since you have already a static route on R1.

    = Remove all duplicate network statement in R1

    =under OSPF, issue the command


    router ospf 100


    network area 0

    network area 0

    default-information originate


    router bgp 100

    bgp log-neighbor-changes

    no synchronization

    neighbor remote-as 200

    neighbor next-hop-self  ( OPTIONAL)

    network mask (OPTIONAL)


    ip classless

    ip route


    I have tested this already. OK


    This O*E2 [110/1] via, 00:00:48, Serial0/0 now appeared in R2 & R3 routing table.


    I can attach a working network here if need be.

  • all devices cannot ping with the command you pasted. Your network is both bgp and OSPF, why? again, this is EBGP, why are you redistributing ospf instead of using static route ponting to you ISP (R0) . Try your ping again and submit true result here. check R2 and R3 routing table, they don't know how to get to 172.18.x.x network. Try a ping.

  • Greg,


    All of these depends largely on what you intend to achieve: your aim for the lab, your goal for the configurations especially between R0 and R1. Your config could be right if it achieve desired goal or attained expected result. That is why I am asking my questions, just to understand your intentions for your initial configs not to impose mine. As your group member, I want to understand goal so I can better tailor my responses. "why do you have 10.x.x.x network in both routing protocol then stil redistribute OSPF in bgp" and both R0 and R1 has default route pointing to each other. Maybe, there may not be any loop issue, but I still don't get it. Maybe we can attach some workstations and servers then test.

  • no worries. we'll meet in the group chat

  • Vijay Swaminathan 491 posts since
    Aug 29, 2008

    Hi Glory,


    Just curious.. I'm still in the process of learning BGP and could not resist myself in posting on this thread.


    I was going through the config posted in post # 23, and went through the bgp configuration.


    In that, the network 172.18.1.x is advertised on the bgp process.. do we really need them as it is a directly connected interface on R2 and since the other OSPF speaking routers (r3 / r4) has got a default route to R2 (as part of default -originate command) they would obviously have connectivity to the link (R1's interface).. just trying to understnad  what could be the significance?


    I have the following configs on R1 and R2 (bgp alone) and this works fine.


    On R1:



    R1#sh run | sec bg

    router bgp 200

    no synchronization

    bgp log-neighbor-changes

    neighbor remote-as 100

    no auto-summary



    On R2:



    router bgp 100

    no synchronization

    bgp log-neighbor-changes

    redistribute ospf 1

    neighbor remote-as 200

    neighbor next-hop-self

    no auto-summary



    and from R3 / R4 the following ping works:





    Type escape sequence to abort.

    Sending 5, 100-byte ICMP Echos to, timeout is 2 seconds:


    Success rate is 100 percent (5/5), round-trip min/avg/max = 4/28/64 ms






    Type escape sequence to abort.

    Sending 5, 100-byte ICMP Echos to, timeout is 2 seconds:


    Success rate is 100 percent (5/5), round-trip min/avg/max = 1/32/108 ms



    - Vijay

  • Hi Vijay,


    No, we don't need to advertise it to outside network. It was meant to be ISP IBGP advertisement, it was my playing around with bgp features. It is not advertise back to OSPF though. My point was to emphasize the use of configured default route instead of redistributing ospf or bgp between R0 and R1. The network statement in bgp I included has nothing to do with original Lab. It personal  additional Lab attached to R0 that is not on that Packet tracer.


    So, please ignore it. That is why I didn't redistribute. Great catch Vijay!

  • aaceituno 242 posts since
    Feb 18, 2009

       Excuse me excuse me, but looking at the Greg configurations looks like have no idea about bgp ...... 


       First, BGP is the internet protocol, so often used for ISP connections to internet.

       Never never never redistribute an IGP into BGP, unless you have an output filter or aggregate that propagate only your public pools into a ISP network.

       If you redistribute an igp this is very possible you will be banned and your bgp connection disabled from the ISP. There is a maximum prefix feature, if you propagate more prefix than allowed then ISP could disable you bgp connection, could drop the excess routes, and of course will only allow your legitimate public routes.


       You could redistribute an igp into bgp only under MPLS/VPN configurations, not for isp or internet connections. And BGP is not a IGP protocol, dont use this if you dont really need this.


      Since you use a redistribute ospf, your networks command are useless.


      Also, remember, for every network command there should be an exact match route for that network command. You configured several networks with a /32 mask. This mask is not like the access lists mask, this is not a "match this bit", this is the real network mask of the route. For example, if you want to propagate the route, will have to configure network command netmask, if you configure netmask you instruct the router to look for the route, these are not the same route.


       If you want to redistribute a default route into ospf, whenever the bgp connection is active, its better to use the conditional route advertisement. You have to configure a route-map, that will match on a bgp learned prefix. This way, if you learn this prefix or prefixes, you will propagate the default route.


       There is no need to use default routes with next hop your bgp peer, because if you use a default route then you dont need a bgp session to learn prefixes, you are going to send all outgoing traffic through that connection regardless of what you learn.

        If you learn bgp prefixes is because you are going to use this someway. If you dont need lo learn nothing because you are going to use a default route, you should apply an incoming route filter that deny everyting. This way your router will save memory, and cpu. Better if your peer apply this filter and dont send prefix to your router.


       If you need to redistribute ospf into bgp, then you will need the match internal and/or external. If you redistribute bgp into ospf, you will redistribute only EBGP prefixes. from other ***. Will not redistribute internal prefix unless you configure the bgp redistribute internal.



        There is lot of things you should learn about bgp yet.










More Like This

  • Retrieving data ...

Bookmarked By (0)