Homematic binding only works partially

I have 2 homegear hosts connected to openHAB on latest Milestone and since quite some time I’m having an issue: When I restart openHAB my Dimmers are working fine but my Max Thermostats aren’t. When I issue a bundle:restart command for the Homematic Binding then the thermostats start working but sometimes the lights stop working then (I can no longer send commands and status updates are not received). Has anyone else seen something similar? Does anybody know what might be going on here?

I found some entries in the logs that could be related :

2019-11-03 07:13:38.013 [WARN ] [ternal.handler.HomematicThingHandler] - Channel not found for datapoint 'XXXXXXXXXX:0#RSSI_DEVICE'
2019-11-03 07:13:38.015 [WARN ] [ternal.handler.HomematicThingHandler] - Channel not found for datapoint 'XXXXXXXXXX:0#RSSI'
2019-11-03 07:13:38.017 [WARN ] [ternal.handler.HomematicThingHandler] - Channel not found for datapoint 'XXXXXXXXXX:0#SIGNAL_STRENGTH'
2019-11-03 07:13:38.021 [WARN ] [ternal.handler.HomematicThingHandler] - Channel not found for datapoint 'XXXXXXXXXX:0#LOWBAT'
2019-11-03 07:13:38.026 [WARN ] [ternal.handler.HomematicThingHandler] - Channel not found for datapoint 'XXXXXXXXXX:1#ACTUAL_TEMPERATURE'
2019-11-03 07:13:38.028 [WARN ] [ternal.handler.HomematicThingHandler] - Channel not found for datapoint 'XXXXXXXXXX:1#CONTROL_MODE'
2019-11-03 07:13:38.030 [WARN ] [ternal.handler.HomematicThingHandler] - Channel not found for datapoint 'XXXXXXXXXX:1#LOCKED'
2019-11-03 07:13:38.032 [WARN ] [ternal.handler.HomematicThingHandler] - Channel not found for datapoint 'XXXXXXXXXX:1#SET_TEMPERATURE'
2019-11-03 07:13:38.035 [WARN ] [ternal.handler.HomematicThingHandler] - Channel not found for datapoint 'XXXXXXXXXX:1#VALVE_STATE'

The log messages are not relevant, they only say that you don´t have the Gateway-Extras configured.

What does the Homegear log say?
You should find them in /var/log/homegear

kind regards

They indicate that there was a restart of homegear (doing unloading and loading of everything) but that’s about it, nothing else is in there. Should I raise the debug level?

I have checked the OpenHAB trace logs of org.openhab.binding.homematic but they don’t indicate anything, just the ping and pongs are shown aswell as commands for the items that work properly. I do not have autoupdate=false set, so I do not understand why the item is “predicted to become MANU-MODE” even though it received BOOST-MODE (it was in MANU-MODE before).

23:19:09.046 [INFO ] [smarthome.event.ItemCommandEvent     ] - Item 'Thermostat_Mode' received command BOOST-MODE
23:19:09.059 [INFO ] [arthome.event.ItemStatePredictedEvent] - Thermostat_Mode predicted to become MANU-MODE

This usually happens when you set the item to something not vaild for that item.

If i recall correctly, you cannot send a mode command to the mode item, you need to trigger the correct switch, because the mode is readonly.

Something like #BOOST_MODE.sendCommand(ON) and not #CONTROL_MODE.sendCommand(BOOST_MODE)

No, that’s wrong. This has nothing to do with Gateway-Extras. This device-thing is wrongly initialised.

The log messages indicate the binding received state messages for datapoints to non-existing channels. The thing object does not contain these channels, although these should exist. Have a look here: Homematic Binding: "Channel not found for datapoint" errors for definitely existing channels

My problems with that were solved by delaying the rule engine on a start: Cleaning up the startup process / renaming rules (windows possible)

I made an interesting discovery: The item doesn’t have any channels for some reason. When I request manually channelGroups is empty, also in PaperUI it is empty when I check at one of the things causing issues. So for some reason the binding isn’t even creating the channels, and that is most likely also causing the prediction (and entire system) to show the behaviour described above.

Now the interesting question is: Why would the binding not create the channels? I guess that is part of the “Channel not found for datapoint”-thread, also look at who made the last post in that :wink:

I am already delaying the rule loading, unfortunately that doesn’t help with this issue (it helps with some others, it’s stupid that this is necessary but it is what it is at the moment).

That’s exactly what i wrote one post above… Although it’s the thing which is missing the channels.

I did not examine your posting history. If you already are aware of that thread, why did you create another one regarding the same topic?

You did not wrote that, so i didn’t knew anything about that.

It’s a race condition/timing issue without a general solution.

I didn’t write that/confirm that before, so nobody was aware that this is indeed the issue I am experiencing.

I was not aware that this is the same thing until I noticed it yesterday. Before I was thinking that something in the command handling was wrong or the connection was interrupted for some reason.