Home page logo

nanog logo nanog mailing list archives

Re: slowing down every 60 seconds due to BGP scaner
From: Alex Rubenstein <alex () nac net>
Date: Sat, 15 Sep 2001 17:12:50 -0400 (Eastern Daylight Time)

12.1 and 12.2 both have some very significant BGP CPU-race conditions.
We've run in to it many times. It's hardly documented in cisco-land, but
one TAC person I spoke to knew about this.

For us, very recent 12.1E versions have this situation fixed. (12.1.8aE2)

We ran into this in all 12.1 mainline, certain 12.2, and early 12.1E had
it. I think 12.1T had it as well (I am anti-12.1T for other reasons).

CEF seems very good in later 12.1E (but, strangely, I've not had many CEF
problems except on a 6509).

On Sat, 15 Sep 2001, Robert Boyle wrote:

BOOTFLASH: 7200 Software (C7200-BOOT-M), Version 11.1(19)CC1, EARLY

Consider updating your bootflash to be somewhere near currrent
(something from 12.2T would be good, identical to the current version

Good advice.

Except, that older 7200's don't have enough flash to do it. Official cisco
opinion is the latest 12.0S.

:) We are using  12.2(1) in production. It is stable except... CEF is
wierd, PPP Multilink doesn't work, and there is a small memory leak
with BGP. We NEED to run it for PPPoE support over ATM. It is
realatively well behaved considering it is bleeding edge.

PPPoE is very well supported and works perfectly in 12.1T (which I hate,
as spoken to above, but since we don't run BGP on DSL aggregation routers,
it's moot). One 7206 here with a lowly NPE-200 is handling well in excess
of hundreds of PPPoE sessions over DS3-ATM.

Heck, we've even a 4700M running 12.1T with many PPPoEoL2TP sessions
running fine.

-- Alex Rubenstein, AR97, K2AHR, alex () nac net, latency, Al Reuben --
--    Net Access Corporation, 800-NET-ME-36, http://www.nac.net   --

  By Date           By Thread  

Current thread:
[ Nmap | Sec Tools | Mailing Lists | Site News | About/Contact | Advertising | Privacy ]