Halt app on Home Assitant complement

Hi.
I have a Conbee II usb into a Raspberry Pi 3 running Home Assistant OS. Its plugged through an extension usb cable as recommended
Deconz complement and integration installed and working all Ok (i have this setup since 2019).
Suddenly, all the devices linked to the Conbee sttoped working. I’ve not updated nor touched anything. I have disabled automatic updates for the complements and also for the OS itself since many months ago.
Looking the Deconz complement addon i can see its stopped everytime this happens. Checking the app logs i can see there’s a problem that says “Halt app”, then sometimes the complement restarts itself (as enabled in the settings) but sometimes it just stops and is no longer starting until i do it manually. In both cases thats so annoying, because i loose the funcionality while reatarting or ill loose it until i interact.

I’m not able to see whats causing this, my knowledge its not enough, but until now I’ve tried:

Uninstall and reinstall the complement, and the HA integration.
Update the complement to the last version.
Reset the Conbee.
Change USB port.
Change Zigbee channel.

None of that is solving the problem.

Here’s a piece of the log:
deCONZ
09:05:33:172 sd-card cid: 035344534836344780ffffffff0161bb
09:05:34:008 DB sqlite version 3.42.0
09:05:34:014 DB PRAGMA page_count: 64
09:05:34:014 DB PRAGMA page_size: 4096
09:05:34:014 DB PRAGMA freelist_count: 0
09:05:34:014 DB file size 262144 bytes, free pages 0
09:05:34:014 DB PRAGMA user_version: 10
09:05:34:014 DB cleanup
09:05:34:024 DB create temporary views
09:05:37:289 DB skip loading sensor Puerta principal 43, handled by DDF Open/close sensor (MCCGQ01LM)
09:05:37:293 DB skip loading sensor Sensor dormitorio 39, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:294 DB skip loading sensor Puerta baño 28, handled by DDF Open/close sensor (MCCGQ11LM)
09:05:37:295 DB skip loading sensor Sensor pasillo 5, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:296 DB skip loading sensor Sensor vestidor 36, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:297 DB skip loading sensor lumi.sensor_motion.aq2 29, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:297 DB skip loading sensor Sensor salon A 31, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:298 DB skip loading sensor Sensor baño 7, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:299 DB skip loading sensor Sensor habitacion F 24, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:299 DB skip loading sensor Sensor dormitorio 38, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:300 DB skip loading sensor Sensor vestidor 37, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:301 DB skip loading sensor Sensor habitacion F 25, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:301 DB skip loading sensor Sensor pasillo 6, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:311 DB skip loading sensor Sensor baño 8, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:311 DB skip loading sensor Sensor salon A 32, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:312 DB skip loading sensor Sensor cocina 30, handled by DDF Motion sensor (RTCGQ11LM)
09:05:37:347 [INFO] - Timezone read is ‘Etc/UTC’
09:05:37:348 [INFO] - Environment variable TZ found: Europe/Madrid…
09:05:37:449 Started websocket server on 0.0.0.0, port: 8081
09:05:37:485 [INFO] - Found file containing button maps. Parsing data…
09:05:37:549 [INFO] - Button maps loaded.
09:05:37:579 found node plugin: /usr/share/deCONZ/plugins/libde_rest_plugin.so - REST API Plugin
09:05:37:755 found node plugin: /usr/share/deCONZ/plugins/libstd_otau_plugin.so - STD OTAU Plugin
09:05:38:405 LightNode 1: Configuration tool 1 added
09:05:39:175 COM: /dev/serial/by-id/usb-dresden_elektronik_ingenieurtechnik_GmbH_ConBee_II_DE2427219-if00 / serialno: DE2427219, ConBee II
09:05:39:769 dlg action: Read binding table
09:05:39:838 DEV found DDF for 0x00158D0006330AE1, path: /usr/share/deCONZ/devices/xiaomi/xiaomi_mccgq01lm_openclose_sensor.json, result: 1
09:05:39:847 DEV found DDF for 0x00158D000839B896, path: /usr/share/deCONZ/devices/xiaomi/xiaomi_rtcgqq11lm_presence_sensor.json, result: 1
09:05:39:849 DEV found DDF for 0x00158D0007C57FC1, path: /usr/share/deCONZ/devices/xiaomi/xiaomi_mccgq11lm_openclose_sensor.json, result: 1
09:05:39:854 DEV found DDF for 0x00158D00057B3EF6, path: /usr/share/deCONZ/devices/xiaomi/xiaomi_rtcgqq11lm_presence_sensor.json, result: 1
09:05:39:861 DEV [09:05:42] WARNING: Existing API key invalid (HTTP 404), will try to create a new one…
[09:05:42] WARNING: No API key found – trying to create a new one…
[09:05:53] INFO: API key successfully created after 1 seconds.
[09:06:33] INFO: Successfully sent discovery information to Home Assistant.
[07:07:25] INFO: Service restart after closing
[09:07:31] INFO: Starting VNC server (local - yes)…
[07:07:34] WARNING: Halt app
s6-rc: info: service legacy-services: stopping
2026/04/03 09:07:34 [notice] 147#147: signal 15 (SIGTERM) received from 144, exiting
2026/04/03 09:07:34 [notice] 455#455: exiting
2026/04/03 09:07:34 [notice] 455#455: exit

Its someone able to put some light into this trouble ?

Sorry, but it looks like the app and integration need to be reinstalled (remember to back up your Phoscon data). See here.

I’ve done that, posted in the OP.
Works for a few hours, then suddenly go back to the same issue.

First, you should not change the Zigbee channel because some devices do not automatically adjust the channel and may need to be reconnected.
To find the actual problem, you should change the log levels to document the error.

I’ve changed channel just to test and try. All my devices are working correctly no mather what channel i use, but thats not the problem.
I can’t see any option to change the log level on the deconz complement.
Or you’re meaning on the HA system ?

See the app documentation.

I suppose its dbg_error but…can you please point wich one of this i must enable ?

dbg_info
dbg_aps
dbg_zcl
dbg_zdp
dbg_ddf
dbg_dev
dbg_ota
dbg_error
dbg_http

Thanks for your time and patience.

I would start with dbg_error and dbg_http. If a restart then occurs, the messages before that are of interest.

Hello, wich one deconz version are you using ?
As the bug is rapid, try to install deconz on another machine and move the conbee (and migrate the database).
You can configure HA to use this second installation, but this will help to see if the problem is from the deconz version, your database, or the host.

Enabled, waiting for the next halt issue.

Right now, using the last available deconz version. But was happening also on earlier version.
I dont have any other resources on to install HA to try your way.

@Smanar Excellent point.
You don’t need HA; deCONZ can be installed and run on many systems (e.g., Windows, Linux, and macOS). So if it happens again, you should also rule out the database. The Phoscon homepage and Google can help you with that.

Sadly, i dont have any other machines than my phone and a very old tablet.
Also, after enabled dbg_error and dbg_http i can’t see any different or related messages on the log, just the same/similar posted on the OP.
@Smanar wich database are you meaning, the deconz one or the HA ? If its Deconz, it is not recreated when i uninstall and delete the complement and the integration ?

I mean the deconz zigbee network setting, you can create and load backup using Phoscon.

Only the device database can be removed, if you are able to see your devices in deconz, the answer is no.

Yeah, I m asking that because last version can have this kind of issue.

Try with loading deconz with command line, to have only deconz log and not HA ones.

Because " signal 15 (SIGTERM) received from 144, exiting" ca, be lot of thing, can be just another app that kill deconz.

It’s strange, since I’m happy with the current HA version. You should try making a Phoscon backup and restoring it with a brand new version from the app store. It’s worth a try.

Brand new version of HA SO, or Deconz app…?
Because if its Deconz I’ve already do that many times… with no positive results, the whole thing works for maybe 24 hours or so, and then back again the dame behavior. I saw on the logs that even its halted when is simply restarting the complement.

I’d like to add i really appreciate all your help and patience. Also, im sorry for somet hings i do not understand or be able to do.

I don’t have issue too on my side.
But there is at least V2.32.5: deconz.service: Main process exited, code=killed, status=11/SEGV

Have you tried with older deconz versions ?

Try to run deconz with command line, to have direct logs, no the ones from HA.

Issue start happened on a older version of deconz, i was non updating anything because all was runnign fine for my needs.
I’ve tried at least two HA backups, wich are including earlier versions of deconz, but the issue persists, so i can’t understand where the problem begins.
I dont know how to run deconz from command line and make it work on HA, my abilities are limited.

For HA problems, the ha supervisor repair function, for example from the Advanced SSH & Web Terminal app in the App Store, can help. Alternatively, you can try disabling the watchdog function on the app’s info page.

Tried both things a week ago or so. No luck.