Jsondb: things configuration lost over and over

I noticed that my rules did not work anymore, that my InfluxDB didn’t get populated properly anymore, my heating controller didn’t work as expected and so on… the reason for that was all the same:

All of my things were deleted!
The file /var/lib/openhab2/jsondb/org.eclipse.smarthome.core.thing.Thing.json contained 0 bytes.

After restoring the things from the backup folder (openHAB was shut down), everything worked fine again. A couple of hours later, the things were lost again. After a couple of restores, I made a backup of the sd card and transferred it to another sd card. I rebooted the system using /sbin/shutdown -F -r now in order to repair file system defects on startup. But again, a couple of hours later, all my things were gone.

How can I mitigate this behavior? What could be the reason for this?

Best regards,

  • Platform information:
    • Hardware: Raspberry Pi 3 Model B Plus Rev 1.3
    • OS: openHABian / Linux 4.19.66-v7+
    • Java Runtime Environment: OpenJDK Runtime Environment (Zulu8.40.0.178-CA-linux_aarch32hf) (build 1.8.0_222-b178)
    • openHAB version: 2.5.3-1

Are you using the amazonechocontrol binding and experoence frequent out-of-memory-errors? Similar issues have been reported.

Yes, I’m using amazonechocontrol. Out-of-memory was a good hint, I found this in the logs:

2020-06-14 16:34:07.018 [WARN ] [me.core.internal.events.EventHandler] - Dispatching event to subscriber 'org.eclipse.smarthome.core.internal.items.ItemUpdater@c0f1f1' takes more than 5000ms.
2020-06-14 16:39:25.713 [WARN ] [mmon.WrappedScheduledExecutorService] - Scheduled runnable ended with an exception: 
java.lang.OutOfMemoryError: Java heap space

Would you recommend following your advice from Web Socket Error on AmazonEchoControl binding every 65 seconds although I’m using OH 2.5.3-1?

Yes, there are no differences except for addons in 2.5.x

Ok, thank you! I’ve installed the updated version as described!

Memory usage is still growing, but perhaps a bit slower. From what I observed so far, the memory leaks seem to originate from reloading of my rules, too.

I’ll keep an eye on it. Thank you very much!