Hmmm.. that's definitely strange. Have a pair of 7600's with 720-3BXL's in them doing similar (roughly 250 public peers per box, a few transit connections per box, dozen downstream customers per box rough) and even though the boxes are loaded down pretty good we don't see what you describe. Our CPU sits around 40% most of the time on 5 min average.
We're running 12.2(33)SRA7 if that helps at all... Paul -----Original Message----- From: cisco-nsp-boun...@puck.nether.net [mailto:cisco-nsp-boun...@puck.nether.net] On Behalf Of Andy B. Sent: Thursday, March 11, 2010 12:41 PM To: Arie Vayner (avayner) Cc: cisco-nsp@puck.nether.net Subject: Re: [c-nsp] Long list of route-maps On Thu, Mar 11, 2010 at 6:36 PM, Arie Vayner (avayner) <avay...@cisco.com> wrote: > BTW, what kind of feeds are you taking from the 150 peers? I hope these > are not full route feeds... Make sure this is not the case... > This is public peering. Feeds usually go from 1 prefix to 10.000. Most are below 1000. The transit session is coming from a direct crosscable to the provider's demarcs, but the router is still the same, just a different interface/vlan You are probably right that I could concentrate ingress route-maps to a single route-map, since this is mostly static. But I do have peers where I want to change localpref because I prefer them at a different location, so there again I loose granularity. _______________________________________________ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/ _______________________________________________ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/