From: Godswill Oletu (oletu@inbox.lv)
Date: Wed Jun 21 2006 - 11:59:08 ART
Fredrick,
It looks more like a password problem. If the BB router is configured with a
password, you will not get any notification, however the BB router will be
displaying sometime like:
*Jun 21 14:38:03.519: %TCP-6-BADAUTH: No MD5 digest from 172.1.17.6(179) to
172.1.17.7(61687) (RST)
But since, you do not have access to the BB router, there will be no way of
seeing that error message.
The other school of thought, as someone else suggested, will be a wrong AS.
However, if you have the wrong AS configured, you will know because, your
router will display an error message similar to this:
*Jun 21 14:36:18.851: %BGP-3-NOTIFICATION: sent to neighbor 172.1.17.6 2/2
(peer in wrong AS) 2 bytes 00FA FFFF FFFF FFFF FFFF FFFF FFFF FFFF FFFF 002D
0104 00FA 00B4 AC01 1106 1002 0601 0400 0100 0102 0280 0002 0202 00
The leading 00FA in the error message above is an indication that, the eBGP
peer is expecting a peering on AS 250 (0x00FA).
Since, your post did not state that, you got that type of error message, it
might not be a wrong AS problem.
'debug ip tcp transactions'
Is a good tool that will help in isolating the problem and point you one way
or the other:
R3640-R2#debug ip tcp transaction
R3640-R2#
*Apr 18 07:48:32.580: TCP: sending SYN, seq 1869623725, ack 0
*Apr 18 07:48:32.580: TCP0: Connection to 172.1.17.7:179, advertising MSS
1460
*Apr 18 07:48:32.580: TCP0: state was CLOSED -> SYNSENT [22587 ->
172.1.17.7(179)]
*Apr 18 07:48:34.580: 172.1.17.6:22587 <---> 172.1.17.7:179 congestion
window changes
*Apr 18 07:48:34.580: cwnd from 1460 to 1460, ssthresh from 65535 to 2920
*Apr 18 07:48:34.580: TCP0: timeout #1 - timeout is 4000 ms, seq 1869623725
*Apr 18 07:48:34.580: TCP: (22587) -> 172.1.17.7(179)
*Apr 18 07:48:38.300: MD5 received, but NOT expected from 172.1.17.7:61745
to 172.1.17.6:179
*Apr 18 07:48:38.304: TCB654938A0 created
*Apr 18 07:48:38.304: TCP0: state was LISTEN -> SYNRCVD [179 ->
172.1.17.7(61745)]
*Apr 18 07:48:38.304: TCP: tcb 654938A0 connection to 172.1.17.7:61745, peer
MSS 1440, MSS is 516
*Apr 18 07:48:38.304: TCP: sending SYN, seq 1080374276, ack 3010331734
*Apr 18 07:48:38.304: TCP0: Connection to 172.1.17.7:61745, advertising MSS
1440
*Apr 18 07:48:38.580: TCP0: timeout #2 - timeout is 8000 ms, seq 1869623725
*Apr 18 07:48:38.580: TCP: (22587) -> 172.1.17.7(179)
*Apr 18 07:48:40.300: MD5 received, but NOT expected from 172.1.17.7:61745
to 172.1.17.6:179
*Apr 18 07:48:40.300: TCP0: bad seg from 172.1.17.7 -- bad sequence number:
port 179 seq 3010331733 ack 0 rcvnxt
<result truncated>
'debug ip bgp'
Is another good tool, but it will not give your alot of actionable
information like the 'debug ip tcp transaction'
HTH
Godswill Oletu
----- Original Message -----
From: <dfredrick@gmail.com>
To: <ccielab@groupstudy.com>
Sent: Wednesday, June 21, 2006 9:57 AM
Subject: eBGP Neighbors
> Hello,
>
> I just finish an online lab. I had trouble with getting an eBGP neighbor
> to come up. It was a BB router, so I didn't have access to see the config.
> Any who... I was recieving RIP updates for it. It was directly connected,
> on the same vlan, and I could ping it.
>
> I tried the following and still didn't work...
>
> - used the ebgp 255 (just incase it was further than I thought)
>
> - used the update loopback0 (just incase it was peering to my loopback)
>
> - and stripped it down to just basics on the router...
> (meaning, didn't have any confederations or anything... just the AS
> it was going to peer with)
>
> Then I turned on all debugging for BGP...
>
> The only message I could get was "connection refused from remote" or
> something like that...
>
> Then I turned on debug ip packet detail... to see if I was recieving any
> packets via TCP port 179... and I wasn't... all I saw from the backbone
> router was rip updates via 224.0.0.9.
>
> Then I the only last thing I could think that could prevent updates... was
> an ACL... I looked at the whole path... no ACLs anywhere...
>
>
> So what else could prevent eBGP neighbors from forming?
>
> Thanks,
>
> Daniel Fredrick
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Sat Jul 01 2006 - 07:57:33 ART