We have 2 iBGP peers which are flapping. We verified that all configuration looks normal(including AS,MTU(9000 on vlan, 9216 on physical int., address including lo0, IGP routing, extended ping with different MTU size). We have multiple iBGP peers, all other are working, these two are the only pair flapping. We compared the configuration with other iBGP peers, they are all identical. The following is some debug ip bgp keepalive logs and sh ip bgp neigher logs. The once caught our eye is the tcp closed windows.

But wired thing is, ospf has always been up with these two peers, it's the BGP that is flapping. I also attached one of the interface configuration, we are trying QoS, just in case somebody can point out some of our misconfigurations.

Any help is greatly appreciated.

symtoms log:

.Jul 9 13:17:09 EDT: %BGP-5-ADJCHANGE: neighbor 192.168.253.14 Down BGP Notification received

.Jul 9 13:17:39 EDT: %BGP-5-ADJCHANGE: neighbor 192.168.253.14 Up

.Jul 9 13:20:39 EDT: %BGP-5-ADJCHANGE: neighbor 192.168.253.14 Down BGP Notification sent

.Jul 9 13:20:39 EDT: %BGP-3-NOTIFICATION: sent to neighbor 192.168.253.14 4/0 (hold time expired) 0 bytes

.Jul 9 13:21:09 EDT: %BGP-5-ADJCHANGE: neighbor 192.168.253.14 Up

test-c1#sh logg | i 192.168.253.6

Jul 9 11:23:43 EDT: BGP: 192.168.253.6 sending KEEPALIVE (io)

jul 9 11:24:43 EDT: BGP: 192.168.253.6 sending KEEPALIVE (io)

Jul 9 11:25:43 EDT: %BGP-5-ADJCHANGE: neighbor 192.168.253.6 Down BGP Notification sent

Jul 9 11:25:43 EDT: %BGP-3-NOTIFICATION: sent to neighbor 192.168.253.6 4/0 (hold time expired) 0 bytes

Jul 9 11:26:14 EDT: BGP: 192.168.253.6 sending KEEPALIVE (rcv_open)

Jul 9 11:26:14 EDT: BGP: 192.168.253.6 received KEEPALIVE, length (excl. header) 0

Jul 9 11:26:14 EDT: %BGP-5-ADJCHANGE: neighbor 192.168.253.6 Up

Jul 9 11:26:14 EDT: BGP: 192.168.253.6 sending KEEPALIVE (updgrp)

Jul 9 11:26:14 EDT: BGP: 192.168.253.6 sending KEEPALIVE (io)

Jul 9 11:27:15 EDT: BGP: 192.168.253.6 sending KEEPALIVE (io)

Jul 9 11:28:15 EDT: BGP: 192.168.253.6 sending KEEPALIVE (io)

test-br#sh logg | i 192.168.253.14

.Jul 9 11:24:44 EDT: BGP: 192.168.253.14 received KEEPALIVE, length (excl. header) 0

.Jul 9 11:25:44 EDT: %BGP-5-ADJCHANGE: neighbor 192.168.253.14 Down BGP Notification sent

.Jul 9 11:25:44 EDT: %BGP-3-NOTIFICATION: sent to neighbor 192.168.253.14 4/0 (hold time expired) 0 bytes

.Jul 9 11:26:14 EDT: BGP: 192.168.253.14 sending KEEPALIVE (rcv_open)

.Jul 9 11:26:14 EDT: BGP: 192.168.253.14 received KEEPALIVE, length (excl. header) 0

.Jul 9 11:26:14 EDT: %BGP-5-ADJCHANGE: neighbor 192.168.253.14 Up

.Jul 9 11:26:14 EDT: BGP: 192.168.253.14 received KEEPALIVE, length (excl. header) 0

.Jul 9 11:26:14 EDT: BGP: 192.168.253.14 received KEEPALIVE, length (excl. header) 0

.Jul 9 11:26:45 EDT: BGP: 192.168.253.14 sending KEEPALIVE (io)

.Jul 9 11:27:15 EDT: BGP: 192.168.253.14 received KEEPALIVE, length (excl. header) 0

.Jul 9 11:28:15 EDT: BGP: 192.168.253.14 received KEEPALIVE, length (excl. header) 0

test-br#

test-br#sh ip bgp neigh 192.168.253.14

bGP neighbor is 192.168.253.14, remote AS 65553, internal link

Description: iBGP to core

Member of peer-group internal for session parameters

BGP version 4, remote router ID 192.168.253.214

BGP state = Established, up for 00:00:34

Last read 00:00:34, last write 00:00:34, hold time is 180, keepalive interval is 60 seconds

Keepalives are temporarily in throttle due to closed TCP window

Neighbor capabilities:

Route refresh: advertised and received(new)