From: Howard C. Berkowitz (hcb@gettcomm.com)
Date: Fri Sep 13 2002 - 20:48:18 GMT-3
>Hi Group,
>
>I'm confused over a new command i have come across for BGP. Picture this if
>you will,
>
>R5------------------------R1----------------------R2
>(BGP) (BGP/ RIP) (RIP)
>
>Essentially, R1 is running BGP and RIP. R1 also has IBGP connections to
>other routers. RIP and BGP are mutually redistributed on R1.
>
>R2 learns about BGP routes which are only attached directly to R1 (Loopback
>0, Serial 1), however, all the other BGP routes which R1 has learnt from its
>iBGP peering are not present in R2's routing table.
Try putting an R3 router and R4 router to the right of R2. R3 is in
your current AS, and R4 is in a different AS. Run eBGP between R3 and
R4. iBGP peer the loopbacks of R1 and R3.
Without trying it, I suspect R5 would see the routes learned from R4.
Just as a start, iBGP has an AD of 200 and RIP has one of 120.
Most real world situations try very hard to redistribute the minimum
of BGP into the IGP. The IGP is mostly communicating reachability of
BGP next hops to the BGP speakers. So, if iBGP redistribution has to
be explicitly enabled, that's consistent with best current practice
for ISPs.
>
>The only way I got this resolved was by typing in the command "bgp
>redistribute-internal" under R1's BGP process.
>
>I was under the impression that when redistributing BGP into any IGP, the
>"redistribute bgp x" was sufficient. But what the following link states is
>that if you wish to redistribute IBGP routes, the above mentioned command
>must also be used.
>
>http://www.cisco.com/univercd/cc/td/doc/product/software/ios121/121cgcr/ip_r
>/iprprt2/1rdbgp.htm#xtocid17
>
>Can anyone shed some light on this it terms of the functionality of this
>command. Essentially, what I am trying to say is that prior to today, i was
>under the impression that "redistribute bgp x" was the ONLY command needed
>?!
>
>Thanks in advance,
>
>Raj.
This archive was generated by hypermail 2.1.4 : Mon Oct 07 2002 - 07:43:51 GMT-3