Re: STP Root port selection - missing step?

 On Wed, 4 Apr 2012 23:24:00 -0500, Tom Kacprzynski wrote:
> Hello
> I'm trying to figure this out something regarding root port
> selection. The
> root port selection per most documents are:
>
> 1. Lowest cumulative root path cost (sum of received cost + incoming
> port's
> cost). This step selects the best path.
> 2. Lowest upstream Bridge ID (designated bridge). This selects a
> unique
> designated switch.
> 3. Lowest upstream port ID (designated bridge's port id). This
> selects a
> unique switch port, which should become the root port.
>
> My questions is what happens if you reach step 3 and everything is
> still
> equal?
>
> How is this possible? If you have a simple topology like
> SW1----SW2===SW3,
> where
>
> - SW1 is the root with only one connection to SW2.
> - SW2 has spanning tree disabled.
> - SW3 has two connection to SW2 (Fa0/20 and Fa0/21)
>
>
> When I did this in a lab, SW3 would see SW1 as the designated and
> root
> bridge. SW3 would always pick the lowest port, in my example fa0/20
> instead
> of fa0/21. So is there a step 4 to the root port selection process,
> where
> it will pick the lowest port id of the local switch?
>
>
> Thanks
>
> Tom

 Hi Tom,

 Yes the 4th step is to choose the lowest port. Since the port ID could
 also be a tie there must be something certain to break the tie. I am
 surprised that you would find documentation not describing this step
 since it is essential to STP. There are some good documents on STP @
 Ciscos website and I would also recommend reading anything Petr Lapukhov
 has written on STP.

 So if you want to manipulate path to root you can either change cost,
 change upstream port priority. In the lab they might tell you to "You
 can't configure switch x" which would point you to the method they are
 asking for.

 Good luck.

 Best regards,

 Daniel

Blogs and organic groups at http://www.ccie.net
Received on Thu Apr 05 2012 - 06:48:41 ART