OH is losing connection to KNX IP Interface

Add-ons Bindings
1

Hey everybody,

I am using OH and KNX for more than 5 years now and never had any problems with my connection. But since 48 hours my OH loses the connection to my KNX IP Interface sporadically. It is a MDT SCN-IP000.02 Gateway.
I even set up a completely fresh OH on my Pi 5 with only the knx bindig installed. Still, it does disconnect a lot. I doubt it might be a OH problem, but i hope one of you can help me.

Here comes the log:

2026-05-20 08:56:45.635 [TRACE] [nx.internal.client.AbstractKNXClient] - Received a Group Write telegram from '1.1.2' to '5/0/5' with value '[1]'
2026-05-20 08:56:45.636 [TRACE] [nx.internal.client.AbstractKNXClient] - Address '5/0/5' is not configured in openHAB
2026-05-20 08:56:46.590 [TRACE] [nx.internal.client.AbstractKNXClient] - Received a Group Write telegram from '1.1.3' to '3/1/33' with value '[0, 0, 2, 17]'
2026-05-20 08:56:46.590 [TRACE] [nx.internal.client.AbstractKNXClient] - Address '3/1/33' is not configured in openHAB
2026-05-20 08:56:51.778 [TRACE] [nx.internal.client.AbstractKNXClient] - Received a Group Write telegram from '1.1.3' to '3/1/32' with value '[0]'
2026-05-20 08:56:51.778 [TRACE] [nx.internal.client.AbstractKNXClient] - Address '3/1/32' is not configured in openHAB
2026-05-20 08:56:52.392 [TRACE] [nx.internal.client.AbstractKNXClient] - Received a Group Write telegram from '1.1.4' to '3/1/11' with value '[67, -38, 0, 0]'
2026-05-20 08:56:52.393 [TRACE] [nx.internal.client.AbstractKNXClient] - Address '3/1/11' is not configured in openHAB
2026-05-20 08:56:53.220 [TRACE] [nx.internal.client.AbstractKNXClient] - Received a Group Write telegram from '1.1.3' to '3/1/33' with value '[0, 0, 2, 18]'
2026-05-20 08:56:53.221 [TRACE] [nx.internal.client.AbstractKNXClient] - Address '3/1/33' is not configured in openHAB
2026-05-20 08:56:55.188 [DEBUG] [nx.internal.client.AbstractKNXClient] - KNX link has been lost (reason: server request on object tunneling link (closed) 192.168.178.47:3671 TP1 medium, device 0.0.0, hopcount 6)
2026-05-20 08:56:55.189 [DEBUG] [nx.internal.client.AbstractKNXClient] - Bridge knx:ip:0083-557f0310 scheduling connect in 60s
==> /var/log/openhab/events.log <==
2026-05-20 08:56:55.190 [INFO ] [ab.event.ThingStatusInfoChangedEvent] - Thing 'knx:ip:0083-557f0310' changed from ONLINE to OFFLINE (COMMUNICATION_ERROR): server request
2

For me this sounds like a network issues. I have a simliar setup and it works without any problems.

And as you are not mention your openhab version, I expect you are running he latest stable openhab version.

3

Hey Holger,

jea, right. I am using latest stable version. But I can recreate the problem on older versions, too.

How can I test my network for problems? I changed all the relevant switches but the problem still exists.

Ty

4

There was an issue with auto discovery switching the thing from tunnel to router that caused the bridge to go offline. It happened for me with the same MDT interface model as yours.

This was resolved, now it discovers two different things, one for router and one for tunnel.

Maybe try to use the discovered thing of the type you need as a bridge. And simply ignore the second one. I assume you have not done so yet?

For me this solved the flapping of the interface type.

5

Hello again,

You were absolutely right. I bought a new router (Unifi Router) and ditched my 16 year old Fritzbox. Now everything is running smoothly again.

Thank you for your hint.

6

For such important pieces of infrastructure, I am usually trying to not depend on routing and DHCP at all, setting static IPs for important devices.

7

Nah, I was using static IP for all relevant devices. There must be something wrong with my old router. Everything is fine with the KNX Gateway.

8

Alright, guys,

i finally found a real soloution to this problem. My router was not broken at all. I am just writing this, so it might help others in the future.

The problem was not my router at all. The disconnects were caused by far too many multicast protocolls in my network. I found this out by using wireshark. Mainly these where caused by two echo dots (an update maybe?) and some Apple Devices. They where sending mdns packets like crazy. This caused the knx IP interface to stop communicating at all, until the storm was over.

My solution was, to put alexa devices into an own VLAN and isolate them there. So they only can access the internet now. Furthermore i put the KNX IP Interface into an own VLAN and blocked all incoming/outgoing traffic with firewalls. Exept my OH Pi and PC ofc.

Works like a charm. Didnt have any disconnects for 3 days now.

Greez