Dell - Internal Use - Confidential 

I think I've encountered this exact problem or a similar problem as well on the 
6500 on 12.2(33)SXJ5. Nothing was making sense to me as the preferred BGP 
next-hop wasn't being used until I refreshed the routes from the upstream peer. 
Then things became consistent. It was showing the best next hop as one thing, 
but forwarded traffic to another peer entirely. No weird PBR or anything was in 
play. Actually, it looked more like an inconsistency between the RIB and FIB 
from what I remember as I thought the RIB showed what BGP was showing, but the 
actual forwarding path was different until BGP routes were refreshed... so 
maybe not the same issue but it sounds familiar. My memory of the problem could 
also be slightly off.

-Vinny

-----Original Message-----
From: cisco-nsp [mailto:[email protected]] On Behalf Of Tim 
Huffman
Sent: Tuesday, October 22, 2013 3:26 PM
To: [email protected]
Subject: Re: [c-nsp] 6500 IOS recommendation?

Now THAT is good to know. Thanks!

--
Tim

-----Original Message-----
From: cisco-nsp [mailto:[email protected]] On Behalf Of Chris 
Welti
Sent: Tuesday, October 22, 2013 7:13 AM
To: Phil Mayers; Jon Lewis; [email protected]
Subject: Re: [c-nsp] 6500 IOS recommendation?

A little word of advice for those that use BGP: Don't use 15.1(1)SY and 
15.1(1)SY1.
They have a nasty little BGP bug that can create black holes or loops for 
random prefixes due to stale RIB entries of withdrawals that are not processed 
correctly (CSCuh43027).
Some withdrawn prefixes are being removed from the BGP table, but not from the 
RIB!
It's very hard to debug and even notice as it usually only affects a couple 
thousand prefixes, but it's annoying.
A lot of 15.1, 15.2. and 15.3 trains have the same issues, so be careful also 
on the 7600 "S" train.
It is fixed in 15.1(2)SY and 15.2(4)S4 only.
So if you want to use 15.1 on Sup720/Sup2T, use 15.1(2)SY.

Regards,
Chris

Am 10/21/13 8:44 PM, schrieb Phil Mayers:
> Jon Lewis <[email protected]> wrote:
> 
>>
>> Are people actually upgrading to 15.1SSY, or just running late 
>> 12.2(33)SXI or SXJ until these boxes run out of resources?
>>
> 
> We're on sy1 everywhere, seems fine. A couple of minor niggles, no worse than 
> any average sx release.
> 

_______________________________________________
cisco-nsp mailing list  [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

_______________________________________________
cisco-nsp mailing list  [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
_______________________________________________
cisco-nsp mailing list  [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

Reply via email to