Issue with Aquiring DHCP

Andy Ciordia andy.ciordia at pgdc.com
Thu Jun 10 14:06:29 UTC 2004


Scott Mewett wrote:
> Brians on the right track with regard to port-fast.
> What you want to enable there is on a per port basis.
> It's not a general spanning tree config.
> I mention this because your earlier mail only showed some general
> spanning tree settings.

> Kickstart then resets the interface, which make the interface leave the
> bridge, and thus when it reconnects the switch port goes back to
> blocking mode and goes through the cycle again.

You know I hadn't thought of that and I'll have to go back and look but 
you're probably right.  If its flushing that port its going to take a 
chunk of seconds before it accepts that link again.

> Also notice the warning it gave. So you don't want to plug in a switch
> to that port or it could cause problems. To protect against that
> versions of CatOS 5.4.1 have a feature called BPDU Guard. If you enable
> this then the port will get disabled if it detects the device being
> connected is running spanning tree. You can control how long it will
> disable for as well.

Right now the machine thats being tested is hooked into the Cat5k but 
the array I'm will be pushing the work to has 100unmanaged switch 
connect into the 5k..  I'm curious if there will be a problem there.. 
Since the Cat5k isn't opening a new port it might just breeze on 
through.  That'd be refreshing ;)

I need to develop my other KS's for the other roles in that rack and see 
what happens.

And thanks go to Philip Rowlands as well. Nice to see that patch got 
wrapped already into U2, I don't have an install tree for that atm but I 
might consider it. :)

-a





More information about the Kickstart-list mailing list