Hi there,

may i ask a question about a rapid spanning tree protocol (RSTP).

Just the short background to understand: I got an an Omada on-premise environment with a headquarter and there are about 20 switches (TL-SG3428XMP) in controller mode (Omada SDN). I'm about to built an Spanning-Tree Configuration like this (f.e. outlined a way simple to the bottom) and it works. SW1 & SW2 are the core switches. SW1 is the root bridge (ID 4096) and SW2 is for backup (ID 8192). The Access-Switch SW11 ist default. My goal is to reboot the core switches one after another for maintaining reasons, without loss of connection to the WAN-Router.

So, if i disconnect and reconnect one uplink manually or disconnect the powersupply, then SW2 takes over and this is fine and like i would expect it, maybe just 1 IP-packet gets lost. This will be absolutely okay and my wish. But for some reason, when SW1 reboots normally (f.e. installing firmware updates) and it comes back, or even if i manually change the ID of the root-bridge, the connection is interrupted every time by about 30 seconds. This is a big problem, because f.e. running video conferences gets affectes and disturbed.

So in my humble opinion i understood the way RSTP works, i expected, that SW1 would be first listening for a while an is in "LEARNING" state and when everything is fine, it takes the connection back to the root bridge SW1 without interrupting the connection. And by rebooting the switch in the normal way, the connection has to kepp alive.

So, does anyone know, why does the connection gets interrupted about 30 seconds, whe the root-bridge switch reboots and after the root bridge is back to work?

Thanks in advance!

Best regards