RPVST+ at Distribution and PVST at Access Switch

[anthonyluna]

Hi. I have two Cat6506 (redundant collapsed-core config) configured w/ RPVST+ and connected to this are a series of Cat2950s configured with PVST. My question is, is this a good practice configuration of STP? (running both RPVST+ and PVST on the same topology) Will there be any problems with convergence? We recently are adding some vlans on the network and we experienced an outage for most of the vlans. The Cat6500s registered duplicate HSRP addresses. Any recommendation to fine-tune this network (considering that the Cat2950 does not support RPVST+) would be helpful. thanks.

[r.sneekes]

Adding a vlan should not lead to outage. Unless there is some mis configuration.

PVST+ en RSVT+ can interact. Convergence on links between pvst+ and rpvst+ domains use the pvst timers for convergence up to 50 sec outage.

BTW cat2950 do support RPVST+.

So using RPVST+ on the all the switches would be the prefered way to go.

http://www.cisco.com/en/US/products/hw/switches/ps628/products_configuration_guide_chapter09186a00801cde7d.html#1150156

You may consider implementing MSTP depending on how much vlan u will be needing. Since there in much less cpu overhead compared to RPSVT+ and PVST+ well vlans increase.

[anthonyluna]

thanks for the inputs. yup, the Cat2950 supports rpvst+. so you're saying that although i'm running rpvst+ on my root bridge (Cat6509s), still the convergence would take the 50 sec pvst+ standard (since the STP timers are dependent on the root bridge config) making the rpvst+ setting on my cat6509s irrelevant since it would default to pvst+ anyway? if this is the case, it seems that running PVST+ on access and RPVST+ on the distribution (on the same STP instances)is somewhat not a best practice.

[gpulos]

you can run rpvst+ and pvst+ simultaneously.

rpvst+ has been built to interoperate with pvst+.

an rpvst+ configured switch will revert to pvst+ in a pvst+ rootBridge scenario.

cisco recommends a slightly different configuration in this case:

1) run two STP instances, one for rpvst+ and one for pvst+.

2) the rpvst+ instance must have an rpvst+ switch as the rootBridge; the pvst+ instances must have an pvst+ switch as the rootBridge.

3) keep the pvst+ switches at the edge of the network; likewise, keep the rpvst+ switches at the core of the network.

please see this link for more info on STP modes:

http://www.cisco.com/en/US/products/hw/switches/ps5213/products_configuration_guide_chapter09186a0080216672.html#wp1150156

please see the following link for much more STP info:

http://www.cisco.com/en/US/tech/tk389/tk621/tsd_technology_support_protocol_home.html

[bauti1428]

This is exactly what I'm looking for. Do you mean that if we have 2 6509 setup with RPVST+ and the edge devices only running regular STP the 6509 will converged faster than the edge devices? Also our end devices in this location are mostly 3500XL which are not supported.