Most of our LANs in Europe are using for spanning-tree today the PVST. One LAN with about 2000 used switch-ports is using Rapid PVST (and we have no problems there); one other site with about 2000 used ports is using MST (MST with using only one instance; instance 0)and we have a lot of spanning-tree issues. - After some investigations I understood, that the MST installation should be using other instances like "0" and a number of instances. But, after reading in Design Guidelines I do no longer understand what reasons exist for using MST instead of Rapid PVST (we have about 50 VLANs in each of them). So my question: Are there any reasons to use in a LAN with about 2000 switch ports MST instead of Rapid PVST?
what really counts is the number of different Vlans involved.
With 50 Vlans you can use PVST without any concerns.
If you decide to use MST you need to provision some instances in addition to the IST (istance 0) if your VLans can be divided in some groups that will share the same topology (the same root bridge).
Configuration of region name, region revision number and what vlan ranges are associated to each STP instance have to be configured manually on every switch and must match.
So I agree that with 50 VLans and 2000 user ports PVST is much more manageable.
In a MAN context you would use MST for its scalability.
If you would have 130 Vlans not every Cisco switch can run 130 STP instances and here starts the problems with PVSTP.
Some old switches can have only 64 STP instances running at the same time.
These older switches cannot run MST.
Hope to help
thanks very much for your response. Your reply is helpful. Especially important was to get a feedback in regards of the additional instances, which should be created (when sharing the same topology etc.)
I want to add some more info. On the backbone are 2 Cisco 6509, which can handle (I suppose) also bigger number of Vlans. On the access-site we have mainly 2950 switches there, but the maximum of Vlans they have to handle is about 8 (data-vlan, different voice-vlans, wireless-vlan, guest-vlan, mangement-vlan). A maximum of 2 x 2950 are in a chain and connected to the both 6509s. - But at the moment there is no "pruning" and no "allowed" command configured.
I described that to you, to maybe get your feedback again, if this described situation is confirming, that PVST (or Rapid-PVST+) is much more manageable than MST here or if this additional info is changing your comment.
Your feedback would be very much appreciated.
Regards ****** Karl
core 6509 switches are not a problem they can support 4096 STP instances ,one for every possible vlan, or at least 1024 STP instances in older versions.
C2950 should have a limit of 128 STP instances I think and can be dependent of IOS version.
Be aware that even if you have no access ports on a vlan if the vlan is permitted on the uplink trunk the switch will try to start a STP instance for it.
VTP Pruning affects only the propagation of user data with broadcast or multicast or unknown unicast destination address, but it does not prevent the creation of an STP instance.
To be sure to avoid overloading access layer switches specify in the allowed vlan list only the vlans that are needed.
If using VTP there can be other limits also for this protocol a switch that receives VTP updates for more vlans that can support (in VTP) will revert to VTP transparent mode (I've seen this).
I think you can stay with PVSTP. Consider to define the allowed vlan list on trunk ports.
Hope to help
The release note of the software release you're using will give you some indication on the number of instance port that a cat6k typically supports. MST goes way beyond RPVST, so if you have lots of vlans or lost or trunk or both, you might want MST. The other advantage for MST is that it is a standard protocol and will thus interact in an easier way than PVST with third party devices.
I am wondering, if there is a link or cisco documentation that says 4096 instances in Cisco 6509s? I have been searching for this for a while and could not locate one.
Best Practices for Catalyst 4500/4000, 5500/5000, and 6500/6000 Series Switches Running CatOS Configuration and Management
From CatOS 5.1, the Catalyst switches can map 802.1Q VLAN numbers greater than 1000 to ISL VLAN numbers. In CatOS 6.x, Catalyst 6500/6000 switches support 4096 VLANs in accordance with the IEEE 802.1Q standard. These VLANs are organized into these three ranges, only some of which are propagated to other switches in the network
Sorry. I didn't understand what your point is.
I don't know if you you understood my original question correctly. What I am trying to ask is: Is there a limit on number of spanning tree instances? If so, what's it. 6500 series switches support 4094 vlans, but do they support 4096 spannig-tree instances?
The Catalyst 6500 running IOS 12.1E supports up to 16 MST instances (numbered 0 through 15). When running MST, you map your individual VLANs (up to 4096 of them on Catalyst 6500's) to MST instances.
This is completely described in, and according to: http://www.cisco.com/en/US/docs/switches/lan/catalyst6500/ios/12.1E/native/configuration/guide/spantree.html#wp1064097
HTH. please rate helpful posts.
HI, I think this question is now answered.
Both MST and PVRST+ are Cisco derived/inspired, they both supppot Portfast, BPDU guard, BPDU filter, root guard and loop guard.
MST requires Medium/High processor bandwidth whereas PVRST+ requires Higher.
*MST supports (some references cite 16 others 64,read) 64 instances mapping to 4096 VLANs on a 6500 switch.
CiscoPress 'Implemeting Cisco Swithed Network (SWITCH) page 121