What you are seeing here is R3 liking the IBGP learned R4 route prior to seeing the R2 learned EBGP version. When it sees the EBGP version, it prefers it and must let R2 know of this change. It could do this by sending an update with the same NLRI and different attributes which would be an implicit withdrawal, or it could explicitly send a withdrawal for the NLRI. Because it learned the better route from R2 and Cisco has a split horizon like rule wrt to re advertising prefixes, the best bet is to send the explicit withdrawal which is what you see it do near the end. It
At 02:31 PM 8/11/2002 +0000, Pierre-Alex Guanel wrote: >More supsense .... > >R3 sent an update but the message was that 172.16.0.0/16 is unreacheable. >That should explain why R2 did not accept the update! > > >R3# >1d11h: %BGP-5-ADJCHANGE: neighbor 10.10.2.1 Down User reset >1d11h: %BGP-5-ADJCHANGE: neighbor 10.10.3.1 Down User reset >1d12h: %BGP-5-ADJCHANGE: neighbor 10.10.2.2 Down Peer closed the session >1d11h: %BGP-5-ADJCHANGE: neighbor 10.10.3.1 Up >1d11h: BGP(0): 10.10.3.1 computing updates, afi 0, neighbor version 0, table >version >1, starting at 0.0.0.0 >1d11h: BGP(0): 10.10.3.1 update run completed, afi 0, ran for 0ms, neighbor >version 0 >, start version 1, throttled to 1 >1d11h: BGP(0): 10.10.3.1 rcvd UPDATE w/ attr: nexthop 10.10.4.2, origin i, >localpref >50, path 100 101 >1d11h: BGP(0): 10.10.3.1 rcvd 172.16.0.0/16 >1d11h: BGP(0): Revise route installing 172.16.0.0/16 -> 10.10.4.2 to main IP >table >1d11h: BGP(0): 10.10.3.1 computing updates, afi 0, neighbor version 1, table >version >2, starting at 0.0.0.0 >1d11h: BGP(0): 10.10.3.1 update run completed, afi 0, ran for 0ms, neighbor >version 1 >, start version 2, throttled to 2 >1d11h: %BGP-5-ADJCHANGE: neighbor 10.10.2.1 Up >1d12h: %BGP-5-ADJCHANGE: neighbor 10.10.2.2 Up >1d11h: BGP(0): 10.10.2.1 computing updates, afi 0, neighbor version 0, table >version >2, starting at 0.0.0.0 >1d11h: BGP(0): 10.10.2.1 send UPDATE (format) 172.16.0.0/16, next 10.10.2.2, >metric 0 >, path 100 101 >1d11h: BGP(0): 10.10.2.1 1 updates enqueued (average=48, maximum=48) >1d11h: BGP(0): 10.10.2.1 update run completed, afi 0, ran for 12ms, neighbor >version >0, start version 2, throttled to 2 >1d11h: BGP(0): 10.10.2.1 rcvd UPDATE w/ attr: nexthop 10.10.2.1, origin i, >path 102 1 >01 >1d11h: BGP(0): 10.10.2.1 rcvd 172.16.0.0/16 >1d11h: BGP(0): Revise route installing 172.16.0.0/16 -> 10.10.2.1 to main IP >table >1d11h: BGP(0): 10.10.3.1 computing updates, afi 0, neighbor version 2, table >version >3, starting at 0.0.0.0 >1d11h: BGP(0): 10.10.3.1 NEXT_HOP part 1 net 172.16.0.0/16, next 10.10.2.1 >1d11h: BGP(0): 10.10.3.1 send UPDATE (format) 172.16.0.0/16, next 10.10.2.1, >metric 0 >, path 102 101 >1d11h: BGP(0): 10.10.3.1 1 updates enqueued (average=53, maximum=53) >1d11h: BGP(0): 10.10.3.1 update run completed, afi 0, ran for 16ms, neighbor >version >2, start version 3, throttled to 3 >1d11h: BGP(0): 10.10.2.1 computing updates, afi 0, neighbor version 2, table >version >3, starting at 0.0.0.0 >1d11h: BGP(0): 10.10.2.1 send unreachable 172.16.0.0/16 >1d11h: BGP(0): 10.10.2.1 send UPDATE 172.16.0.0/16 -- unreachable >1d11h: BGP(0): 10.10.2.1 1 updates enqueued (average=26, maximum=26) >1d11h: BGP(0): 10.10.2.1 update run completed, afi 0, ran for 8ms, neighbor >version 2 >, start version 3, throttled to 3 >[Resuming connection 1 to r4 ... ] > >On my next post, I will attempt to explain why the network was unreacheable! >Any one wants to guess? Gheese, am I playing alone? :) Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=51197&t=51169 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
