[c-nsp] Maximum spannig tree instances

[Tim Durack]

On Tue, Jul 14, 2009 at 11:43 AM, Jon Lewis <jlewis at lewis.org> wrote:

> On Tue, 14 Jul 2009, Geoffrey Pendery wrote:
>
>  "I've wondered about that...if we were to move to MST, we're going to
>> have to assign every VLAN to an MST instance, which could get messy."
>>
>> That's exactly why I was warning about the 16/64 instance limit.  This
>> was my mindset when moving from PVST to MST, and I suspect there are
>> many others out there thinking this way.  But if you have more than 64
>> VLANs, you can't do that.  You'll have to look at their topology and
>>
>
> That's not what I meant.  I just meant we'd have to decide which instance
> (of likely just a few of them) to assign every VLAN to...as every VLAN has
> to be assigned to some instance.  I should setup a lab of switches again and
> play around with MST.  IIRC, the docs I've read about MST on cisco.comgenerally split up the VLANs between MST instances 2 and 3.
>
>
We left everything in MST0, and pull a few VLANs into MST2 for
load-balancing reasons. Core-1 is root for MST0, Core-2 is root for MST2.
Works for a simple topology, where every switch has redundant links back to
a couple of core switches. Not sure it would be so great for the kind of
topologies being discussed here.

However, as soon as I want to add another VLAN to MST2, I have touch *every*
switch in the MST region. And during the process MST is inconsistent -
either I adjust the two core switches first, and every edge switch flips
over to MST0, or I do every edge switch first, core last. Either way it's a
lot of STP fun.

I'm going to guess the standards body that came up with MST doesn't do too
much network configuration work...

Tim:>