@lukass2000 

Gi1/0/5 or Gi1/0/24 is port 5 and port 24 in your switch.

  @lukass2000 you seem to have a loop going on between port 5 and port 24 or you have RSTP Rapid Spanning Tree Protocol kicking in since it may that both of those ports go to the same location or there is a loop somewhere else where your packets are going out on 5 and coming back on 24 or vice versa.  Typically in a spanning tree arrangement one port should be blocked and the other should be active so that the packets cant come back and create a round robin loop.  That will take down any network when you get a packet storm and the network is flooded with noise. 

Would help to see your topology but...as I said, typically if you do have 2 links going to the same switch or you have a 3 way arrangement with all switches talking to each other...someone has to break the loop so that you dont wind up with flooding or packet storm. RSTP knows how to manage that.

If you use a hub/spoke arrangement where everyone only talks to one device (all switches going to one router but no switch has a direct connection to another switch) you wont see that situation occur.

Thank you for your information :)
 

@Fleegle61  Here is a very simple structure of the Omada SDN.
Of course, the other ports are also partially occupied, but that's the Omada system and the problem ports :)



Again and again I get the message of a loop from the SG3428 on port 5. A "normal" Netgear SG105 switch is connected to this port.

Now and then I get a loop reported from the uplink ports on the SG2218 and T1600G-T28.
But it is much worse, that the "blocked" message keeps coming at these ports.

It always takes some time until the network is up and running again, and of course also Wifi :(

But there is really no reason why the OmadaSDN blocks the uplink ports?

One switch is connected to the other.
That should work without any problems, right?
Thanks :)

  @lukass2000 Thanks for the picture.  That tells alot!

Now, I don't see any reason for a loop at first glance or why a port would go blocked.  What I would ask; are you meshing the AP's???  That could be perhaps where the loop is coming from since that would create another path between the switches.  

If meshing, then the 1600 has 2 ways to get to the 2218, via port 24 direct and port 22 via AP that could or would be meshed to the AP on the 2218 on port 13. 

For port 5, your diagram shows that the 105 is connected to it.  On the 105, is there anything running WiFi or are they all just direct connected?  If your PS4, TV or PC has WiFi and its on and connected to one of the AP's, you have created a loop.  

Normally a TV, PC or PS4 should only use one of the other but I know that in some laptops and other PC's you can have both on, both get their own IP address assigned via DHCP and, they may at some point decide to load balance and try to send some packets in one direction and other packets in another direction.  Port 5 sends something to one of those 3 devices and it expects the reply to come back but not the original packet it sent coming in from another port which could be from the AP side.  So if port 5 sends and it sees its own packet coming back from lets say port 20 or 22, that is considered a loop and it will block.  

Hope that kinda explains it some more.  

@Fleegle61 Thanks :)

Sometimes my laptop is also connected to the Netgear switch via a LAN cable.
The Laptop would also have WiFi, maybe that's causing the problem?

I will watch that

The three access points are within the respective Wifi range.
I think, they could then reach each other via mesh.

Can I turn this off somehow?
So without affecting the WiFi itself?

With this problem it is very bad that the blocked port is immediately blocked completely.
The entire network behind it always fails in the short term 

  @lukass2000 Hallo Lukas, du kannst die Funktion "Loopback detection" unter den Diensten (Services) vom Switch abschalten. Hier ein Screenshot vom Omada:

  @lukass2000 

I've had something similar on unifi network, wired access points have started and meshed and created a loop in the network,

if you don't use mesh you can switch this off in the site config.

I have also seen on some forums that this can also happen with Sonos, they have their own mesh and create a loop in some cases.

the fact that there is a PC that has connected wifi and cable at the same time should not create a loop in the network, had that been the case there would have been a lot of problems in many networks.

  @lukass2000 It's actually not a BAD thing to have multiple paths to the same desitnation.  That gives you diverse routing so that if one link fails, the other will work just fine.  RSTP is there just for that reason.  To keep the looping from happening and to block one of the 2 paths so that it is contained.  

Your case with port 5 has me guessing that something on your 105 switch has 2 ways to travel.  

Make sure you have RSTP turned on with any managed switch that you have.  Unmanaged switches do not have that capability.  

You can also try turning off the mesh to see if that changes anything.  

Lastly, check those devices on the 105 switch to make sure they dont have any wifi turned on if they are hardwired to the 105 switch.  

@Fleegle61 
@btx 

Okay, have seen, all my Switches have activated "Loopback Detection", and also RSTP dectivated.

Should i deactivate "Loopback detection" on the SG2218, or is better i activate RSTP on the SG2218?
Thanks :)
 

  @lukass2000 

For now I would recommend that you keep RSTP active on all managed switches and watch the alarms to see if any other port complains of a loop or blocking.  

It doesnt hurt to keep that protocol enabled.