From: David Ham (ccieau@hotmail.com)
Date: Tue Nov 19 2002 - 06:03:14 GMT-3
Hi all.
I have weird BGP symptom to ask.
Scenario:
AS64666
---------------------------------------------|
| ----------------------- |
| | MSFC1--- MSFC2 | AS 64605 |
| --/----\---/------/--- |
| / \ / / |
| / / \ / |
| / / \ / |
| ------------------------ |
| | MSFC3----- MSFC4 |AS64610 |
| ------------------------ |
--------------------------------------------|
As you can see MSFC1,MSFC2, MSFC3, MSFC4 are fully meshed.
MSFC1 and MSFC2 belong to BGP AS 64605.
MSFC3 and MSFC4 belong to BGP AS 64610
These AS are sub as and belong to confederation BGP AS 64666.
BGP connection is OK. Configured NETWORK statement for all the subnets
between the links so that BGP routes can
be seen in any MSFCs. Also used "maximum-path" to do the load balancing.
The problem occurs after I reload the MSFC3 and MSFC4. All the BGP routes
showing at MSFC2 are trying to use path
from MSFC1 via IBGP not by the directly connected path to MSFC3 or MSFC4.
I have used "weight" to control path but no success. Only way I can use
normal directly connected path to MSFC3 and
MSFC4 is to do "Clear IP bgp *"
MSFC1 is using directly connected path to MSFC3 and MSFC4.
Am I missing something wrong or Is this IOS bug !! I am using
c6msfc-jsv-mz-12.1(11b)E.
I will really appreciate if you can spare some time to sort this problem.
Thanks in advance.
David Ham
Regards,
David Ham
This archive was generated by hypermail 2.1.4 : Tue Dec 03 2002 - 07:23:06 GMT-3