Hi exprienced deCONZ users,
I did use the Conbee Stick 2 with Deconz at my Synology NAS DS716+ since ages and it was always running quiet good. Since I replugged all my hardware on the weekend it is not working anymore. Some Zigbee devices have endless delays other don’t work at all.
I don’t find the problem.
Does someone can help me finding the error?
My setup:
I did plug the conbee 2 Stick with a 1m USB extension to the first usb port of my Synology NAS DS716+
I run in docker the deconzcommunity/deconz:2.24.2 container (as before)
In the phoscon app the sensors get back to (not grey) after a very long time and I need to use them first.
If use a light switch it can take more then 20seconds that the light turns on. very often it does not work at all.
I did replug the hardware and rebootet the docker container again. same problem.
the docker container constantly takes like 20% CPU time. This was before usally only 1% or so.
Thanks for your time and help!
I did install the 2.25.3 now .
It works just a little better but I’m still having massive delays. Also the CPU load remains higher then 20%.
Any other advice?
Mine was workung perfect with 2.23.02 all of the versions beyond made the system unresponsive and lagging. for me it got better with the conbee II with version 2.26.0 hope it helps.
I’m running v2.26.1-beta now. CPU from this container is now down to 16%.
The lights react again, but still with a greater delay than usual. Devices that I had just used seem to work immediately. Devices that I haven’t used for a while have a delay of 1-5 seconds.
I’ll keep monitoring this.
Does any of this in log may cause the CPU load?
libpng warning: iCCP: known incorrect sRGB profile
GW sd-card image version file does not exist: /opt/deCONZ/gw-version
a lot of 'DB legacy loading sensor' for example:
DB legacy loading sensor Consumption 30 30, later handled by DDF wireless relay with neutral
a lot of 'skip legacy loading' for example:
skip legacy loading 00:17:88:01:06:03:51:95-0b / LWA001
a lot of 'DB skip loading sensor' for example:
DB skip loading sensor WS Zisterne 5% 60L 163, handled by DDF Aqara Water Leak Sensor SJCGQ11LM
Device firmware version 0x26780700 ConBee II
a lot of 'TY_DATA_RESPONSE' for example:
TY_DATA_RESPONSE: seq 207, dpid: 0x04, type: 0x02, length: 4, val: 96
My setup is the following:
Synology NAS DS716+ with docker.
In docker I ran the container deconzcommunity/deconz:2.26.1 with
DECONZ_VNC_MODE: 1
DECONZ_VNC_DISPLAY: 0
DECONZ_VNC_DISABLE_PASSWORD: 0
ad so on.
I did login from two different computers with different VNC clients:
RealVNC VNC Viewer 6.21.406
RealVNC RealVNC Viewer 7.10.0
Both times everything is turned on in the vnc config:
I’m able to see the log. The shortcut for select all works. I copy it.
But I can’t paste it to any editor in my 1. Windows 10 or 2. Windows 11 system.
IDK for other but on my side I don’t see something special on logs :
No missing bind in loop.
No tuya device spamming
In general no spamming request
No error
No “busy” error
You have a constant 20% cpu load ?
for TY_DATA_RESPONSE They are from the device itself, and deconz manage only 1 request, and the same device only do it every 10 mn.
There is an issue about that for some TRV, but you are far away the spamming result of them
@Smanar Thank you so much for looking into that. Sadly you didn’t find anything I could change and my system is still under >23% CPU load and absolutly not running smooth. I have very long delays since I did replug all the hardware. I did restart everything many times now, but it does not help.
You can’t install deconz on an other machine to test if it’s from deconz itself ?
You can clone your zigbee network using the phoscon backup/restore feature.
Make a test with a more recent version ? ATM there is a lot of update on performance on deconz version.
@Smanar@Mimiix: I would not be able to circle the devices because I have the delays on all items. If I send a light on via openhab it can take 10-20 seconds till the Hue bulb turns on. If I do it via a Aqara switch the time is around 20-50 seconds…
As I did not find a solution with 2.24.2, 2.25.3 and 2.26.1 I now installed 2.23.02 and everything is running perfectly again. It just takes milliseconds to turn on the lights and the CPU of the docker container is now only 3%.
To double check I then did start the container again with deconz 2.26.1. The Container has >15% CPU load seems to go till over 20% in the next hour. And the delays are back. So something changed in 2.24.2, 2.25.3 or 2.26.1 which makes the system not stable anymore.
I will have to stick with 2.23.02 till a solution is found.
I did share my log already and don’t now what else I could do to assist.
Can anyone inform the developers of this issue?
@Developer and @Mimiix As I still had the same issue I did try my system with the new 2.27.0. I again had from the beginning a high CPU load and big delays.
So I again switched back to: 2.23.02.
So sadly the problem is still unfixed.
Is this still the case with current version v2.27.6 / v2.28.0-beta?
We had some devices fixed which showed more traffic / CPU usage due heavy polling.
I don’t notice delays for controlling lights in my setups after (re)starting deCONZ. But if other devices are heavily polled this could be an indicator why controlling messages take longer. If this is still the case with current versions we need to figure out which devices are affected and adjust the DDFs of them.
I did restart everything many times now, but it does not help.
