Hi Jerônimo,

There is no reason a computer could send 11K packets of ARP, except
there is a switching loop there.

Say, do you use in your network those small and cheap switches, said,
DLink DES-1008, Encore ENL-901NWay, etc.?

When these devices are installed directly on user's rooms, it's easy
that the users change the way the cables are mounted, and create
switching loops. I passed this problem myself...

PS/Off-topic: Sou do Rio de Janeiro, se quiser posso te ajudar a
identificar o ponto onde está acontecendo esse loop.

2008/4/14, Jerônimo Bezerra <[EMAIL PROTECTED]>:
> Hello All,
>
>  i'm sorry for comma, my intention was tell 11 000 pps :) Follow my scenario:
>
>  80 VLANs and each of then with 100 until 600 computers;
>  my ntop's NIC is tagged to 3 vlans ( 14, 145, 137 );
>  some unmanaged switchs, some hubs, e some managed switchs on each vlan;
>
>  In one vlan ( 145 ) one computer was sending 11 000 pps of ARP
>  broadcast, and my ntop was telling me just 300 pps. That's my question:
>  why 300 pps?
>  My core router was 99% of CPU.
>
>  Jeronimo
>
>  Graeme Fowler escreveu:
>
> > On Mon, 2008-04-14 at 11:06 -0500, Gary Gatten wrote:
>  >
>  >> 11 or 100 pps is nothing - not even close to anything to worry about.  A 
> 10Mb Ethernet "network" does over 19K pps.  Most broadcast storm control 
> features default to several thousand pps, so really - 11 or a 100 is a tiny 
> fraction of a percent or available bandwidth.
>  >>
>  >
>  > I think Jeronimo's email ost a bit in translation - it was 11kpps,
>  > phrased as "11.000 pps". Not every written language uses a comma as a
>  > decimal separator for positive powers of ten :)
>  >
>  >
>  >> Switching Loops don't cause broadcast storms.  If there is a loop it 
> won't be found looking for excessive broadcasts.
>  >>
>  >
>  > Loops in ethernet networks cause all manner of lunacy, because they
>  > amplify anything that isn't unicast. After some time (depending on
>  > hardware), they amplify unicast too as the L2 devices involved age out
>  > or conflict out their MAC tables; once most switches see MAC addresses
>  > on several ports they can get a little confused!
>  >
>  > Jeronimo - you gave no indication of your network topology, and only a
>  > vague description of what happened so it's tricky to tell you why you
>  > didn't see the problem with ntop.
>  >
>  > Graeme
>  >
>  > _______________________________________________
>  > Ntop mailing list
>  > [email protected]
>  > http://listgateway.unipi.it/mailman/listinfo/ntop
>  >
>
>  _______________________________________________
>  Ntop mailing list
>  [email protected]
>  http://listgateway.unipi.it/mailman/listinfo/ntop
>
_______________________________________________
Ntop mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop

Reply via email to