Very strange behaviour of Deconz

Hail to the quality product.

grafik

It’s the device, not deconz.

It’s Deconz, not the device.
If I pair the device with:

  1. SILVERCREST BRIDGE (Tuya) = OK
  2. ZB BRIDGE (Sonoff) = OK
  3. HUE BRIDGE = OK

If I pair the device with Conbee = NOT OK

So, please quit to save Deconz.

Oh, I wasn’t aware that I’m in need to save anything here.

Almost the whole thread here is about the question (rightfully if I may add) why the LIDL/Tuya devices pop a new endpoint all out of a sudden. The answer has already been given: because they just send inappropriate data. That’s fact and proven, end of story.

And I confess: it’s for that I do not like any Tuya devices and do not touch them anymore. Very personal choice I have to live with.

Blockquote
Almost the whole thread here is about the question (rightfully if I may add) why the LIDL/Tuya devices pop a new endpoint all out of a sudden

Only smart plugs.
Lights don 't suffer this problem; Multiple sockets don’t suffer this problem.
And it’s not “sudden”. It happens if I interrogate the device by “read”

Tuya? With Deconz I have “cluster problems” only.
Sonoff ZBmini are worse. Version 1.0 are kicked out from the mesh after 6 hours (but you can rejpoin them); 2.0 are kicked out sometimes but you have to switch off and then on the electrical cabinet to have them online again. For this reason I’m changing course to eMylo (Tuya).
Deconz doesn’t understand anymore when an Ikea bulb is phisically switched off. Once upon ago, it reported the bulb offline after few minutes. Now it needs hours. Firmware bug? Again?
When I reported many devices in my mesh reset alone (switching on light and switching off subwoofer and amplifier) what was the diagnosisin this forum/github?
Firmware bug. Unbelievable because the issue involved many brands (Osram, Vimar, Ikea and Tuya).
After 2 or 3 Conbee firmwares the issue has been solved.
I think developers should pay attention to “regressions” when a new release is published.
In the last months I saw regressions for Tuya smartplugs, Ikea lights (just written above) and Aqara end devices (maybe solved).
Listen to the customers… somethimes! :slight_smile:

Almost the whole thread here is about the question (rightfully if I may add) why the LIDL/Tuya devices pop a new endpoint all out of a sudden

Only smart plugs.
Lights don 't suffer this problem; Multiple sockets don’t suffer this problem.
And it’s not “sudden”. It happens if I interrogate the device by “read”

I’m sorry, but I was again just responding to what you refer here in that thread, which is smart plugs, being Tuya devices. Might be true that other device classes or even other plugs with potentially different hardware and firmware do not have that issue, but the plugs in your possession you reference here and the one I had 2 years back undeniably have the same issue. And sending irrational data on a simple read request is… let’s call it disturbing.

I cannot comment on any Sonoff devices, as I don’t own any. Same goes for any Ikea bulps, but I do just have a handful of other lights which are all recognized in a couple of minutes as being offline.

As for the device reset, I just had that doubtful pleasure myself recently. It was a specially programmed 4-gang switch, so that wasn’t funny at all. Although I had the sniffer running, I had no active debug log, so no clue where it came from.

Speaking just for myself here: I’m almost always curious if users report unusual behavior and that already led to quite a number of ironed out (nasty) bugs. Would be interesting to also have access to deconz core and the firmware, there might be more dwelling in the depths.