Monday, October 8, 2018

Multi-vender STP question, and other weird observations

I have a client who currently uses several Adtran switches that were sold to them from a previous vendor. We just added two Cisco switches into a brand new IDF, and I discovered that the Adtrans run a form of single instance RSTP, and by default, the new Cisco's running rapid-pvst were only participating with the Adtrans's STP in the native VLAN.

I set the Cisco's to MST and they seem to be working ok, but in the beginning while testing it on the first Cisco, it seems like the r-pvst BPDU's from the second Cisco were being looped around through a second, redundant interface, and the Adtrans were flooding them, and when they hit the first Cisco's trunk port, it sent it into err-disabled state. Switching both Ciscos to MST at the same time predictably prevented this from happening again.

I changed the STP priorities on various switches to ensure they were all participating nicely, and it seems like they are, but the Cisco lists the Adtran peers as an MST Boundary. Even though it seems to be working fine, I've always read that MST doesn't work past a region boundary. The Adtrans don't seem to have a way to edit any of the instance info, so I don't think I really have any other options until we phase these Adtrans out. Any opinions on this?

On a similar note, this same site also has three Cisco WAPs, all three connected to different Adtran switches, but now they're showing up in the new remote IDF as CDP neighbors to the aforementioned new Cisco swtiches. It seems that, like the r-psvt BPDU's, the Adtrans doesn't really seem to know what do to, and is forwarding them. I have never seen this before. All the Adtrans are up to date.



No comments:

Post a Comment