Re: [nsp] troubleshooting stuck in active routes

From: Andrey Kostin (ankost@lightcom.ru)
Date: Fri Oct 12 2001 - 08:04:49 EDT


This two bugs are very similar with your case, but 12.0S isn't mentioned
here:

Bug Id : CSCdu79625
Headline IP-EIGRP: Routing loop and incorrect topology table entries
Product c6msfc Model
Component eigrp-ip Duplicate of
Severity 2 Status R
Version Found 12.1(7)E Fixed-in Version 12.1(10.1) 12.2(4.1)
12.2(4.1)S 12.1(08a)E03 12.2(4.1)PI 12.2(5.2)T 12.1(9.5)E
Release Notes

Release-note:
-------------

Under rare conditions, routers running EIGRP may point to each other for
certain
IP prefixes, causing a routing loop. This behavior was caused by the fix
for
CSCdr91621.

Bug Id : CSCdr91621
Headline EIGRP:SIA caused by routers waiting on each others reply
Product all Model
Component eigrp Duplicate of
Severity 3 Status R
Version Found 12.0 12.1 Fixed-in Version 12.1(7) 12.2(1)T 12.2(1)
12.2(1)PI01 12.1(6)E03 12.1(7)E 12.1(7)EC 12.2(1)S 12.2(4)PB
Release Notes

In very unusual circumstances, EIGRP may get a stuck-in-active route with
routers waiting on each other. There is no workaround.

---------------
Andrey Kostin, ankost@lightcom.ru +7-095-956-4951
Lightcom Corp. http://www.lightcom.ru

----- Original Message -----
From: "Gert Doering" <gert@greenie.muc.de>
To: "CLAEREBOUDT Elke" <ECLAEREB@mail.mobistar.be>
Cc: <cisco-nsp@puck.nether.net>
Sent: 5 октября 2001 г. 1:34
Subject: Re: [nsp] troubleshooting stuck in active routes

Hi,

On Thu, Oct 04, 2001 at 11:25:10AM +0200, CLAEREBOUDT Elke wrote:
> When those routes dissapear, we sometimes have the SIA's and the clearing
of
> neighbourships. It isn't always the same router which isn't responding to
> the queries, there's really no line that is congested, no trouble with cpu
> or memory.
> There seems to be a loop in the queries.

This is something that I have seen as well. We have a topology with a
number of loops purposely built into it. Sometimes when routes
disappear completely ("line goes down, no route anymore"), they get SIA
and EIGRP flaps to recover.

I have not fully understood why this is happening, as it is also neither
a CPU nor memory problem on the boxes affected (7200s, 7500), and it
doesn't happen "all the time" but only "very rarely" (maybe once per a
few hundred withdrawals). Last time I actually managed to get a "show
ip ei top active" from the routers involved, and they pointed to each
other in a circle... from the docs, this should never happen (if a
router queries one that is already active, it should send a negative
response immediately).

12.0S, various subreleases, no obvious "open bugs" found.

gert

--
Gert Doering
Mobile communications ... right now writing from *Prague*
... mobile phone: +49 177 2160221 ... or mail me:  gert@greenie.muc.de



This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:19 EDT