Distributed Trunking + MSTP

[Dr1zzt]

Hello

We have created a pretty simple set-up with procurve 3500 switches using distributed trunking to connect to 1 or 2 access switches and to the core. There is a simplistic topology image attached for more information.

We have repeated the topology (in the attached image) in a 5 floor building giving each floor its DT-Stack (and 1 or 2 access switches) with the stack connected to the Core switch in the basement (also with DT-trunking).

We have enabled MSTP in all switches (same config name and revision number, no instances), we have given priority 0 to the core and priority 15 to all other switches (totally 15 switches in DT and access connections).

Problem is, every access switch (that is SW3 in topology image) has also become root in the MSTP. The Core switch is also root in MSTP.  Since we now have 5 root switches in total (same region) this is not normal of course. The network seems to work fine, but there is still optimization to be done.

Maybe we need to disable MSTP in all access switches when they are in DT trunks? Is this the best practice? In all access switches we have mostly IP phones and workstations, so there is no probability for a loop.

I will give you more information as requested.

Thank you all in advance.

[Dr1zzt]

OK, I figured it out why I have so many root bridges.

All dt-trunks have bpdu-filtering enabled by default so they dont exchange BPDUs with the dt-stack. Each switch thinks itself as the root.

Problem is, I don't want to disable this filtering, since its the way DT is working. Only solution I can think of, is disabling STP on every root switch (except core ofc).

Any suggestions?