Re: Route reflector -Cause

Brian ,

Thank for the quick answer . The platform is very old 12000 PRP 1 .

Lets say both RRs have the same policy towards the client Z . Does this
mean client Zs CPU will calculate only once the Update ?

Also can this different policies which is ONLY applied to one peer in the
iBGP routers -Z , affect the other ibgp peers
 because high CPU usage is also visible on the other routers who have only
one policy from both RRs .

On Sat, Mar 8, 2014 at 10:52 PM, Brian McGahan <bmcgahan_at_ine.com> wrote:

> What are the platforms and memory? A full BGP feed can surely cause high
> CPU. Check the "show proc cpu sorted" during times of high CPU and it'll
> tell you if the BGP process is the one pegging it.
>
> Also if the two peers have different policies they fall into separate BGP
> update groups, which means that the CPU has to calculate the update
> separately for the two peers as opposed to just once.
>
> Brian McGahan, 4 x CCIE #8593 (R&S/SP/SC/DC), CCDE #2013::13
> bmcgahan_at_INE.com
>
> Internetwork Expert, Inc.
> http://www.INE.com
>
>
> -----Original Message-----
> From: nobody_at_groupstudy.com [mailto:nobody_at_groupstudy.com] On Behalf Of
> Mark Szwed
> Sent: Saturday, March 08, 2014 2:48 PM
> To: Cisco certification
> Subject: Route reflector -Cause
>
> Hi guys ,
>
> I have a situation in which the production routers in my company are
> showing huge cpu usage . After carefully troubleshooting and analyzing all
> configuration , I come to the conclusion that in the Core routers , on one
> of the iBGP peers (lets call Z) has huge update messages that keep on
> increasing .Keep in mind that Z has got eBGP peering as well .
>
> Z is peering with 2 RRs and on RR1 is a route map which sets a community
> to the outgoing routes to Z and the other RR2 sends all routes without
> community .
>
> router Z is receiving less routes from RR1 due to the communty settings
> ,can this kind of RR peering be the cause of high CPU load .
>
> My thought was RR2 sends all routes to Z , but Z receives less routes
> from RR1 , and this triggers constant update message from RR1 .
>
> Anyone to help explain if it is possible to peer with an 2 RR with
> different policies towards the client ?
>
>
> Thank you!
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html

Blogs and organic groups at http://www.ccie.net
Received on Sat Mar 08 2014 - 23:32:38 ART