RSTP & Trunk Ports

sixarm

First of all, let me describe my topology...I have three 2950's connected in a 'triangle' format [SW1 (root bridge), SW2 & SW3] with a Cisco 2611XM hanging off SW3 as a router-on-a-stick. My test PC was connected via SW2. I have three extra VLAN's configured, each with a corresponding sub-interface on the 2600.

I have configured all three switches to use RSTP and the root bridge is set the same across all VLAN's. I have enabled spanning-tree fastport on the interface that my PC connects to and also the trunked interface to the 2600.

Now, this is the test that I was doing in order to see the difference between STP & RSTP. My PC was configured in one of the VLAN's and I was running a constant ping across to the sub-interface on the 2600. I then disabled the interface on SW1 to SW3, forcing STP to open the link from SW2 to SW3. I expected to see a seamless fail over, however there wasn't really a difference from when I previously tested using basic STP.

In order to troubleshoot the issue I returned the interface on SW2 that my PC connects to back to the native VLAN, removed all the sub-interfaces on the 2600 and configured the interface on SW3 connecting to the 2600 to access mode instead of trunk.

On retesting there was no pings dropped when I shut the interface on SW1 and then only 1 dropped when I opened it again. Now this is what I'd expect from RSTP.

Am I missing some config here that would cause this?

Thanks in advance.

sixarm

No thoughts?

/usr

Are you saying that RSTP was acting the same as STP when you were routing between VLANs, but after you removed the VLANs and subinterfaces, everything worked as expected?

I'm a bit confused by your scenario.

sixarm

/usr wrote: »

Are you saying that RSTP was acting the same as STP when you were routing between VLANs, but after you removed the VLANs and subinterfaces, everything worked as expected?

I'm a bit confused by your scenario.

Yep, that's exactly what happened.

ptlinva

It's pretty cool to see so many people working on the same things that you are..

I tried your scenario but ran into problems because one of my switches was a CatOS switch - 2948 -and I'm having issues troubleshooting with it.

I'll should have my third 2950 delivered today and a 4th switch (3550) here by Thursday. I purchased these do test this exact scenario.

You'll here back from me shortly with the results...

Good Luck SixArm!

sixarm

I'd be very interested to hear how you get on.

I'm tempted to wipe all my switch configs and start again to see if this makes any difference.

/usr

I will try this when I get home tonight or tomorrow. I've got a 2620XM setup with subinterfaces and VLANS, like you described, so I won't have to do much work to replicate your scenario.

sixarm

Did either of you manage to test out this scenario?

My PC won't boot at the moment, so I've not been able to have a play with my Cisco's for the past few days. Bloody Windows.

ikarugamatt

If you can post up your port switch config that may help to troubleshoot it. It sounds like you've got spanning-tree portfast enabled and spanning-tree mode rapid-pvst (I think that's the right one) as well but config will always help things further.