Sorry I don’t know. I don’t work with nor use the addon currently so I zero knowledge about it
ok, no problem. thx for comming back to me.
Does this anybody else know?
PS.: There are indeed situations where the addon doesnt restart…
thx
Watchdog of the add-on seems to do nothing indeed.
This is odd.
Today I noticed that all my zigbee sensors were unavailable.
Checked the logs but there was nothing, at least not what I saw.
The watchdog is enabled but in this case it didn’t work.
So yeah. The watchdog doesn’t seem to work.
thanks for your feedback, so i am not alone, that’s good to know.
Who is the maintainer of this addon?
Got the same last night at 02:30 am. deCONZ add-on was stopped and had not restarted. Seems the watchdog is not working (properly).
Same here again. For me it was 02:10.
Its officially supported addon, Im note sure there are explicitly stated owners of addon like there is with integrations.
I also have an issue where the Deconz addon isn’t starting. Either after reboots, or crashes. Also my logs are filled with thousands/millions of this warning:
Unexpected thermostat mode None
2022-08-05 22:04:34.560 WARNING (MainThread) [pydeconz.models.sensor.thermostat] Unexpected thermostat mode None
2022-08-05 22:04:34.563 WARNING (MainThread) [pydeconz.models.sensor.thermostat] Unexpected thermostat mode None
2022-08-05 22:04:34.565 WARNING (MainThread) [pydeconz.models.sensor.thermostat] Unexpected thermostat mode None
2022-08-05 22:04:34.567 WARNING
Please let me know if there is anything I can provide to help debugging
Yes, I got exactly the same problem. My wife had to get up at 3 a.m. and not a single lamp was working. I got this message in my log (addon stopped, no auto restart, watchdog was on):
2022-08-13 02:47:29 INFO (MainThread) [pydeconz.websocket] Reconnecting to deCONZ (core-deconz)
2022-08-13 02:47:30 INFO (MainThread) [pydeconz.websocket] Reconnecting to deCONZ (core-deconz) failed, scheduling retry at an interval of 15 seconds
But I was able to fix that, I simply clicked on start addon.
Yeah, that the addon stops for everyone at a certain specific time is already reported like 2 years ago, but still no fix or cause… And the issue went stale, seems no owner for this addon to look at?
Do you have a link for me?
I have said this before in this thread.
I run Deconz on a separate machine (Intel NUC i5) with wired network and fixed IP address and I install it on a Debian directly with the official Dresden Elektronik debs. And nothing resets at night ever.
For me stability is to reduce the length of the dependency chain.
Look at the list of things that can go wrong which will make Dezonz not work at all
- Deconz daemon can crash. If Deconz daemon is running started by systemd as it is with standard deconz.service it will restart on failure.
- The host OS can crash - common problem. But a Debian OS is known to be conservative, boring, out of date with the latest applications, and therefore rock solid stable.
- The USB driver can crash
- The Conbee can break or crash
- Home Assistant core can crash
- The Deconz integration can have a bug and disconnect. I must say I am impressed with the stability of the integration itself.
Now we move Deconz to a docker container and on top of the list above we now also get…
- Docker itself can crash
- The deconz container can crash and this is not just one failure mode. The container runs a subset of an OS (but not an entire OS like a virtulized machine) and any of these elements can fail. Like DNS client, USB bridge, the shared libraries, etc. If the entire container crashes the supervisor watchdog should restart it. But if the container still runs but with a crashed subcomponent, then it becomes hard for a generic watchdog to realize it. Almost all the things that can go wrong in the host OS can also go wrong inside the container.
- As far as I can tell the IP address of the container is assigned by a local DNS as part of the HA environment. That can go wrong so the IP suddenly changes and you get a lost connection
- USB detection and function now has two layers. Host OS and Docker container. We have seen race condition issues with the device discovery when starting the container.
- The Dresden Elektronik team incl most of the open source participants do not test Deconz as a HA addon. They repeatedly tell people on their forum that they do not support the various installation methods made by 3rd party
- The addon for Deconz is not maintained by a dedicated developer on a daily basis. My impression is that they are built by Frenck who maintains quite an impressive number of addons while also being involved in HA release management. My guess is that a new Deconz release is tested for a few minutes. If you install from a deb and you install it on the same host OS as the developers you have 1000s of hours of testing. If you install deconz on a host OS that is more exotic then you have even less test coverage than using the addon because very few runs this way. This is not a critique of Frenck. He does an amazing job but he is only a human and he needs to sleep and eat and be with his family and do so many things for the project. It cannot be any other way.
An addon based installation will always have more ponts of possible failures than a bare metal installation on the same host OS as the original developers. And that goes for all addons unless the addon is a unique Home Assistant feature.
But - for a newbie it us much much easier to install an addon. And you do not need an extra computer. Addon has a great value. But it comes with less stability. A small extra raspberry pi cost nothing. It consumes like 5 W. If you pair the important switches with light groups it even works at a basic level when HA is dead. I an not looking back.
Here is the first one created
thx guys for your feedback. In my case it was 03:01 in the night.
Does anybody know how HA is fetching core updates and when?
I know it sound’s strange but 2022.8.4 was released.
I think it has todo with deconz itself crashing, some versions are more stable than others.
But that’s a really strange issue!
PS.: is there an automation to trigger when an addon (deconz) stops? For notification purposes.
Oh that’s unfortunate
I hoped it did anyway.
Sorry but I don’t know a way to set the timezone on a raspberry pi any other way.
Maybe google knows
Br,
Raymond
How can I do this? I tried to bind my Philips dimming switch to my one light group. Then stopped the deconz addon → doesnt work anymore.
Via deconz (vnc) and the bind dropbox I can’t get it to work, too.
Hi everyone
When setting up Deconz see the internal given ip adress of my Conbee 2 stick (i guess)
When using deconz to log into the GUI inside HA, i see the same address
But after a reboot or two, the integration fails to connect (Settings\Devices and Services)
And when i check the IP address it differs from the address given to the device upon fort setup. Often only by plus or minus 1 on the last digit.
To remedy this i remove the integration, and re-add the integration again. Annoying.
-Can i fix this?!
Hey, I hope this is the right place to ask. I recently had my Conbee II soft-brick my HA install. It seems like this has happened to other people, but I decided to plug it in to my PC, and to do a firmware update, all good. HA is now booting fine, however all my devices are missing. I have never done a backup in Phoscon, so I do not have a .dat file to import. I do however have a backup of my HA install from before it started acting up. Is it possible to extract the devices from that or am I SOL?
The address you’re using is a docker address and its not static, you can change the address to be the machine address instead and it should stop breaking on reboots.