Spanning tree question, I have layer 3 switches connected via fiber in geographically different locations. These sites are connected via switchport mode access and the VLAN associated with the SVI that is being used for the layer 3 traffic. I have a lot of OSPF adjacency changes on a daily basis ( the timers are set to 1 second ) and noticed at times when changes were being completed at one site it appeared to take out parts of the network at another geographically separated location and I started working from ground up trying to figure out the issue.
I found that I am sending BPDU's between sites and seeing TCN's being sent out from the root bridge out to the rest of the network for switches listening on that vlan. For instance if site Site A is connected to Site B with a point to point IP on VLAN 900. The two sites are configured with site A g4500x t1/0/1 switchport mode access vlan 900 and site B 4500x T/10/1 switchport mode access vlan 900. This seems easy enough, both sides decides who is root bridge for that particular VLAN and all is good.
What happens when you have other switches connected internally to both site A and site B and they are connected via trunk ports and VTP is running and VLAN 900 is not pruned from the internal network. Would a TCN generated internally at SITE A on vlan 900 cause spanning tree to recalculate VLAN 900 between Site A and Site B ( or vice versa direction doesnt matter ) causing traffic to black hole for the spanning tree recalculate time on that particular VLAN?
I am trying to figure out if this is possible.
1.) Since we do not prune VLAN 900 and for some reason something connected to a trunk port internal to the LAN causes a TCN on VLAN 900 will that cause Spanning-Tree to recalculate on VLAN 900?
2.) If so while spanning-tree recalculates could that cause traffic between SIte A and Site B to blackhole and effectively stop communication between the two sites while spanning tree recalculates.
I appreciate any information!
