Don't receive Homematic states and figures from 2 am onwards

Platform information: OH3 - RPi 4 - 4GB
Issue of the topic:

I’m using OH2 in parallel and rebuilding my actual environment at OH3 step by step.

However, each night my OH3 do not communicate which my Homematic any longer. At 2am all connections seems to stop. I viewed the logs but there are no information at all. Only ntp and Modbus are working without error. There are no information whats happent with Homematic communication.
See also attached screenshot.

The homematic communication on OH2 is not affected and still working after 2am.

What I did:
Restart OH3 several times, uninstal Homematic binding, resinstal homematic binding, delete Things and redefine Things. Nothing helps so far.

Any idea?
Thank you in advance

Hallo Andreas,

what kind of Homematic devices are you using? I have seen problems if you are using especially HmIP devices and more than one OH instance is connected to the CCU at the same time.

During my testing, I even sometimes got a message on the CCU the HmIP service crashed. You could also ssh into the CCU and check whether you can find some more information in the logs on the CCU (

What is the status of the things in the OH 3 installation?

In the last weeks I have made some improvements in the binding that will officially available with OH 3.1. But if you like you can download a test version from here GitHub - MHerbst/openhab-addons-test


Hi @MHerbst, i am facing similar problems with my HmIP things. How can i install your newest binding version in my openhab 3.0.1 release build?

If you follow the link, you will also see a file that explains what do to.

Hi Martin,
thank you ver much for the information.
I’ve problems with the weather station homematic:HmIP-SWO-PL and homematic:HmIP-STHO.
Homematic Switches (none IP) are working fine.

What I recognize: I’ve a confab in place - some historical thing:
00 2 * * * sudo service openhab2 restart
which restarts open hab service on RPi with OH2 at 2am.
I changed yesterday the restart to 3 am. And the communication on OH3 to homematic:HmIP-SWO-PL stoped now at 3am. OH2 and OH3 are running on different RPi’s.

However, I’ll now download your new binding and let you know if this fix the problem.

Best regards and thanks again

Bildschirmfoto 2021-04-07 um 20.33.08

Don’t get the new binding installed.
What I did - as you described:

Download org.openhab.binding.homematic-3.1.0-SNAPSHOT.jar to local laptop.

  1. Uninstall the Homematic binding
  2. Stop openHAB
  3. Clean the cache
  4. Copy the test jar into the addons folder (copy it into
  5. Start openHAB (on slow devices like Raspberry you have to start openHAB a second time because of the emptied cache) - did it but no homematic binding is now available at all

Also run Karaf

feature:install openhab-transport-upnp
feature:install openhab-transport-serial

Sorry, didn‘t look into the link. I‘ll try the next days. Thx so far!

@MHerbst i noticed this behaviour today as well. I can’t get your testversion installed neither on latest snapshot nor on 3.1.0 M3.

@muelli Do you see any messages in the log file?

Can you also check in the Karaf console the output of bundle:list | grep Homematic

Maybe there is some sort of dependency problem with M3, but in the latest Snapshot it should work.

@andi1306 It should work on the latest snapshot. I will test it again. Are there any messages in openhab.log?

Can you also check in the Karaf console the output of bundle:list | grep Homematic

That’s interesting. It seems that the restart of OH2 confused the CCU. As I said before there are problems if you connect more than one OH instance at the same time to a CCU. The old HM devices are normally don’t have any problems, but with the HmIP devices there can be some real strange problems.

You can try the following in the OH3 installation after it does no longer receive updates:

  • open the Karaf console
  • execute bundle:list | grep Homematic and note the number in the first column
  • execute bundle:restart <number>

Does this help? As a bypass you could create a cron rule on the OH3 machine too and perform a restart at the same time.

ok, i will download the latest snapshot over the weekend and run a test again.

3.1.0 M3 has been stable over the last days ( with the usual binding, not your tweaked version for the addon folder )

How would roughly the “then” part of a cron rule ( or even check if an IP channel does not update for some time ) look like to restart the homematic bundle? In case i run into issues with M3 i could do this as a workaround.

I will check tomorror whether the test version really works with the latest snapshot.

As far as I know the bundle:restart is only possible in the Karaf console (I will check this). My idea with the cron rule was to use the same technique that @andi1306 is already using for his H2 installation and restart both OH installations at the same time.

ok, got it. Then i potentially work on a rule to alert me with a Telegram in case one of my IP devices does not report updates or stays at the same value vs. 1 or 2 hours ago… so that i can just restart the homematic bundle in the karaf manually. i dont want to restart the whole OH3 instance if not neccessary.

I wait for your feedback on the snapshot, this one did not work with me:

   3.1.0-SNAPSHOT - Build #2310

OK and thank you very much for the information.
I think I’ll go a more straight forward way. Next weekend will be the weather condition very badly. I’ll migrate whole OH2 → OH3 and shut down OH2 afterwards.
This will hopefully fix the issue due to the fact that then is only one connection to CCU3.
I’ll send an update on Sunday. Please cross your fingers :hugs:
Best regards and thanks again for the information so far

That doesn’t look right.

OK, things seem to work for me now with the “normal” binding from the release build 3.0.1.

But there is one weird thing I do not understand: I linked 2 timestamp items to my ActualTemperature thing. One of them using UI the other using an .items file. The item defined in the .items file receives NULL while the item defined by UI receives correct timestamps. Both of them have a profile and a stateDescription.

.items file:

DateTime tempWallFreddieTS "Letzte Messung [%1$td.%1$tm.%1$tY %1$tH:%1$tM]" <temperature>   ["Point"]  {channel="homematic:HmIP-WTH-B:3014F711A0001F5A49940499:00265D89897A70:1#ACTUAL_TEMPERATURE"[profile="timestamp-update"], stateDescription=""[pattern="%1$td.%1$tm.%1$tY, %1$tH:%1$tM Uhr"]}

I must be doing something wrong in my .items file. Can somebody help please?

I have uploaded a new a new test version. But it won’t work with 3.1.0.M3 because there have been some updates in openHAB core after the milestone was released. I have tested the binding successfully with the latest snapshot (2319).

Migrated yesterday nearly everything to OH3 and switched off OH2.5. The OH3 environment run through without loosing the connection over night :slightly_smiling_face: Perfect, well done so far.

However, I have problems with some basics at all.
Former I used a rule to build an average of illumination. I would like transfer this rule to ECMA. Former OH2.5:
var helligkeit_avg = (Wetterstation_ILLUMINATION.averageSince(now.minusMinutes(10)) as Number).intValue
is not running through.
How to handle it in ECMA with item.Registry.getItem(‘Wetterstation_ILLUMINATION’).getState() and rrd4j requests?
I don’t get the right information in the internet.
Thank you very much in advance and best regards

See here: Blocky: Calculating the average of light intensity last minutes

Replying to myself in case anybody might be interested: The problem was the name of my item defined in the items-file. There is one called “tempWallFreddie” for the temperature and the failed one named “tempWallFreddieTS” for the timestamp of update.

If I rename the last one to “tempWallFreddie_TS” evrything works fine. Obviously items have to have names that differ in a minimum of 3 characters. ???