[j-nsp] BGP error messages

[Julian Eccli]

>=20
> However, our bgp is resetting pretty frequently. One of the=20
> recent cause of
> resetting is due to a TCP error on both ends.

Please open a case with JTAC since they can help resolve your ongoing
problem.

> In addition,=20
> what does the
> statement "connection collision prefers 192.168.1.2+1926=20
> (proto)" means?

This means both peers tried to open a TCP session with each other at the
same time.  The peer logging the above message was the one with the
lower BGP identifier so closed it's local session.  RFC1771, section
6.8, gives a good description on what this means exactly.

>=20
> Error messages from Juniper (JUNOS 5.0R5.1):
>=20
> Oct 14 07:36:01 juniper2 rpd[15846]: bgp_traffic_timeout:=20
> NOTIFICATION sent to
> 192.154.24.74 (External AS 9989): code 4 (Hold Timer Expired=20
> Error), Reason:
> holdtime expired for 192.168.1.2 (External AS xxxx)

Hold timer expired since did not receive any keepalives/updates within
hold time period.

> Oct 14 07:36:01 juniper2 rpd[15846]: bgp_event: peer=20
> 192.168.1.2 (External AS
> xxxx) old state Established event HoldTime new state Idle

Logging new state change from established to idle.

> Oct 14 07:37:07 juniper2 rpd[15846]: bgp_pp_recv: NOTIFICATION sent to
> 192.168.1.2 (External AS xxxx): code 6 (Cease), Reason:=20
> dropping 192.168.1.2
> (External AS xxxx), connection collision prefers=20
> 192.168.1.2+1926 (proto)

Already defined above.

> Oct 14 07:37:07 juniper2 rpd[15846]: bgp_recv: peer=20
> 192.168.1.2 (External AS
> xxxx): received unexpected EOF

Received unexpected End Of Frame (not file).  This means other side
closed connection without sending a cease message first.

> Oct 14 07:37:17 juniper2 rpd[15846]: bgp_event: peer=20
> 192.168.1.2 (External AS
> xxxx) old state OpenConfirm event RecvKeepAlive new state Established

Log message on state change to established.

>From what I can see from both log files it appears both sides tried to
open a connection at the same time after hold time expired.  When the
connection collision occurred 92.168.1.1 closed it's local connection as
it should have (as seen in it's log files as well as 192.168.1.2 cease
message log entry).  No big deal from the exact events above except why
the hold timer expired in the first place.  You need to find out why you
are not seeing keepalive's/updates before the hold time expires, which
may not always be a BGP issue (could be a lower layer).

If this is an ongoing problem, always open a case with JTAC.  They would
need the whole log file from each peer, not just BGP messages.


-Julian