[c-nsp] Re-thinking (remembering) how a switch operates

[Matt Buford]

Also, keep in mind that every time any port in the VLAN changes state, the 
CAM table for that VLAN is cleared (well, technically it is all set to 
expire in 6 seconds, which is almost the same) on every switch.  On a 
reasonably large VLAN with someone rebooting a server here and there, you 
can be wiping your CAM regularly.  For example, if I flap a link somewhere 
on my network, I see a significant increase in traffic on all links for that 
VLAN that then slowly dies down as the CAM table is populated again.

Enabling portfast on the host ports avoids this.  A link flap on a portfast 
port will not cause a spanning-tree TCN, and not clear the CAM.

----- Original Message ----- 
From: "David J. Hughes" <bambi at hughes.com.au>
To: "Lincoln Dale" <ltd at cisco.com>
Cc: "'cisco-nsp'" <cisco-nsp at puck.nether.net>
Sent: Thursday, April 28, 2005 1:03 AM
Subject: Re: [c-nsp] Re-thinking (remembering) how a switch operates


>
>> i'm wondering if its a bad cable such that you actually have a
>> unidirectional link that can receive but not transmit...
>
> Nahh, just a side effect of people running longer arp caches than cam
> table timers.  We reconfigure the mac timeout to match the arp timeout
> to ensure this doesn't happen.
>
>
> David
> ...
>
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
>