[c-nsp] BGP Cpu

[Gary Stanley]

At 11:23 AM 6/21/2007, hjan\@libero\.it wrote:
>Hi, someone have idea on how a "clear ip bgp * soft in" and "clear 
>ip bgp soft out" can smooth out CPU use ?
>Before the clear:
>CPU utilization for five seconds: 99%/0%; one minute: 69%; five minutes: 66%
>  PID Runtime(ms)   Invoked      uSecs   5Sec   1Min   5Min TTY Process
>  174     97860441440825344          6 47.56% 53.47% 52.70%   0 BGP Router
>After the clear:
>CPU utilization for five seconds: 18%/0%; one minute: 37%; five minutes: 37%
>  PID Runtime(ms)   Invoked      uSecs   5Sec   1Min   5Min TTY Process
>  174    101175361440872287          7 16.15% 22.41% 23.23%   0 BGP Router

Are you taking full routes from 1 peer? If you only have 1 peer you 
could just take a default route, no point in taking transit routes 
from just 1 upstream.



G. Stanley
IP Engineering (AS36352)
Velocity Servers, INC
gary at velocity-servers.net
http://leaf.dragonflybsd.org/~gary