openhab-addons: [homematic] Heating groups broken after upgrade from 3.1 to 3.2

Expected Behavior

Heating groups (HM-CC-VG1) devices should just work, especially for SET_TEMPERATURE. For this it’s required that the things for these virtual devices are online.

Current Behavior

Currently all HM-CC-VG-1 (INT0000*) things are offline with OH 3.2.0-1.

In events.log I see messages like:

Thing 'homematic:HM-CC-VG-1:ccu:INT0000008' changed from UNINITIALIZED (BRIDGE_UNINITIALIZED) to INITIALIZING
Thing 'homematic:HM-CC-VG-1:ccu:INT0000008' changed from INITIALIZING to OFFLINE (CONFIGURATION_ERROR): Device with address 'INT0000008' not found on gateway 'ccu'

In openhab.log on startup I see warnings (linebreaks manually added) like:

[WARN ] [ommunicator.AbstractHomematicGateway]
Can't load device with address 'INT0000008' from gateway 'ccu':
java.util.concurrent.ExecutionException: java.io.EOFException: 
HttpConnectionOverHTTP@236f05f4::SocketChannelEndPoint@573bc467
{l=/192.168.42.2:41458,r=homematic.sail.spinnaker.de/192.168.42.28:9292,ISHUT,
 fill=-,flush=-,to=0/0}
{io=0/0,kio=0,kro=1}->HttpConnectionOverHTTP@236f05f4(l:/192.168.42.2:41458 <-> 
r:homematic.sail.spinnaker.de/192.168.42.28:9292,closed=false)=>
HttpChannelOverHTTP@17c70d83(exchange=HttpExchange@334d439e
{req=HttpRequest[POST /groups HTTP/1.1]@6856a0a6[TERMINATED/null] 
res=HttpResponse[null 0 null]@61c74937[PENDING/null]})
[send=HttpSenderOverHTTP@4e43eb99(req=QUEUED,snd=COMPLETED,failure=null)
[HttpGenerator@174b107c{s=START}],recv=HttpReceiverOverHTTP@4ac5c1a3
(rsp=IDLE,failure=null)[HttpParser{s=CLOSED,0 of -1}]]

Possible Solution

The problem occurred on upgrade from 3.1.1 to 3.2.0. As a workaround downgrading OpenHAB (including the homematic binding) to 3.1.1 solves the problem for me.

Steps to Reproduce (for Bugs)

I run OH 3.2.0-1 with text file configuration. The homematic bridge and things are configured as

Bridge homematic:bridge:ccu [ gatewayAddress="homematic.sail.spinnaker.de", callbackHost="192.168.42.2" , callbackRegTimeout=180, timeout=30 ]
{
    Thing HM-RCV-50            BidCoS-RF     "CCU2"
    Thing GATEWAY-EXTRAS-CCU   GWE00000000   "GATEWAY-EXTRAS"
    Thing HM-CC-VG-1           INT0000008    "HeizungRolandVirt"
...
}

I defined items as

Switch Gr_Roland_Mode_Comfort   {channel="homematic:HM-CC-VG-1:ccu:INT0000008:1#COMFORT_MODE"}
Switch Gr_Roland_Mode_Lower     {channel="homematic:HM-CC-VG-1:ccu:INT0000008:1#LOWERING_MODE"}
Switch Gr_Roland_Mode_Boost     {channel="homematic:HM-CC-VG-1:ccu:INT0000008:1#BOOST_MODE"}
Switch Gr_Roland_Mode_Auto      {channel="homematic:HM-CC-VG-1:ccu:INT0000008:1#AUTO_MODE"}
Number Gr_Roland_SollTemp  "Soll-Temperatur Roland [%.1f °C]" <temperature> (All, OG_Roland, RRD, gT_Roland) {channel="homematic:HM-CC-VG-1:ccu:INT0000008:1#SET_TEMPERATURE"}

With this I observe the above mentioned errors and in the Main UI the heating group things show up as “OFFLINE”.

Context

If you use heating groups (combining HM-TC-IT-WM-W-EU and HM-CC-RT-DN), you have to send all updates to the group (HM-CC-VG1), since the group overwrites changes, that are sent directly to HM-TC-IT-WM-W-EU or HM-CC-RT-DN. This means, that this issue breaks all changes.

I reported the problem at https://community.openhab.org/t/homematic-hm-cc-vg-1-heating-groups-broken-in-oh-3-2/131296 before and there seems to be at least one other user affected by the same problem. I generated a debug log, which is available at https://www.spinnaker.de/tmp/homematic.log

Your Environment

• Version used: OpenHAB 3.2.0-1 as amd64 Debian package
• Java from Azul Zulu 11.0.13-1
• Debian 11.2 on a normal PC.
• Homematic with a CCU2 2.57.5 with 47 devices/things (incl. 7 HM-CC-VG-1).
• Operating System and version (desktop or mobile, Windows 10, Raspbian Buster, …):