My network consists about a 5x SG350 and 1x CBS350. They are all connected in star with leaves topology. (1->2 1->3 1-4 3->5 5->6 etc). Some are connected with LAG with 2 ports, most are just single-cable no-LAG connected. There are loops in the network.
From time to time we have a power failure somewhere as those switches have single power. Not an issue for us until a year ago.
It seems that after upgrading the SG350's that since one of the later codelevels communication can be blocked after a reboot of one of the switches. A down/up on the downlink port from the central switch always fixes the issue. But it's a very weird issue and leads to serious downtime now with each power failure.
The logs display normal link-up's and traffic-forward messages. Also CDP display normal neighbor-switch information.
I have the feeling that smartport plays a role in this, as I struggled from the same time as this issue started with Native-VLAN-mismatch errors, that appeared to be caused by auto-voiice vlan changing the isl ports from my intended VLAN configuration to a default vlan configuration.
To exclude smartport from being the cause I tried disabling smartport on one of the switches. This was a bad idea.... The switch in question was not reachable and up/down of the isl did not bring it back online. I had to fore reboot (go back to previous config) to get the switch online again. So take care when you at switch level disable smartport... Could cause you downtime!
example of the configuration on the isl port (example direct connected ports without LAG):
!
interface GigabitEthernet28
flowcontrol on
spanning-tree link-type point-to-point
switchport mode trunk
So from my perspective switch configs look good at both sides.
Any ideas? I read this post which seems to be same issue, however the solution there was to downgrade firmware. This does not really help me as it also occurs between SG350 and my new CBS350.
