Rapid-PVST

[mjarkovsky]

Dear all,

I am trying to implement Rapid-PVST in our company network. Our network consists of 20 switches. We have two HSRP L3 switches. I was thinking what would be the best choice for root switch in STP topology. I thought that choosing active HSRP L3 switch would be good choice. However, as I read Rapid-PVST convergence time is significantly worse when root switch is dead. It means that if active L3 switch would be dead, stp convergence would occur (and "electing" of new root switch) and the disruption in the traffic would occur until stp convergence is done and standby router goes active. Is it better choice to choose some "neutral" switch in such topology? What would be better practise?

Thanks for you help in advance.

Michal

[jj27]

Since you have two switches running HSRP I assume they are the 'core' of the network.

Use this command on the primary switch:  spanning-tree vlan 1-4094 root primary

Use this command on the secondary switch: spanning-tree vlan 1-4094 root secondary

Obviously replace vlan 1-4094 with your VLANs.

[mjarkovsky]

I know how to change root switch.. My concern is what is better approach to speed up the convergence.

From what I have read Rapid-STP convergence is significanlty longer when root bridge fails. If the root bridge would be also active HSRP it means that in worst case scenario it can take up to 20-30 seconds for network to converge. Means - no routing..

Maybe choosing the standby HSRP as root bridge would be good idea?

[Mitchell Dyer]

Can you site your sources on poor RPVST convergence times? The only information I found regarding the convergence of RPVST after a root bridge failure was regarding mis-configurations of RPVST.

This might help with understanding why there shouldn't be poor convergence times:
http://www.cisco.com/en/US/tech/tk389/tk621/technologies_white_paper09186a0080094cfa.shtml

My concern with setting the standby HSRP switch as the root is that typically in a full mesh environment with either a dedicated core or collapsed core is that this will cause all traffic to traverse the link between the standby and active HSRP switches, as links from other switches to the active should be blocking. This may produce a bottleneck to the active HSRP device. This may or may not be true for your particular topology.

Sent from Cisco Technical Support iPhone App

[mjarkovsky]

I have read the following article about RSTP:

http://blog.ine.com/2009/09/07/rstp-and-fast-convergence/#more-1908

and more detailed comparision is here:

http://blog.ine.com/wp-content/uploads/2010/04/understanding-stp-rstp-convergence.pdf

in that PDF file, page 32 author mentioned 30 seconds convergence times in worst cases.

topology that we have looks similar to this:

[Mitchell Dyer]

I still don't think that setting the root bridge to something other than the active HSRP switch will provide any benefit, regardless of which switch is the root, it's failure will result in non optimal conditions regardless of where its located in the network.

HTH

Sent from Cisco Technical Support iPhone App