Watchdog: Home Assistant restarts at irregular intervals

Hello all,

I am at a bit of a loss. HA is actually running wonderfully. But since yesterday HA restarts out of nowhere and without operation. From the logs I can unfortunately not take much and a real solution I have not yet found here in the forum.

Version: core-2022.2.0
supervisor-2022.01.1

The logs show the following. First message unfortunately only as screenshot:

After HA just rebooted, only disconnecting power and reconnecting helped. Only visible error in logs after reboot:

22-02-03 14:50:10 ERROR (MainThread) [supervisor.homeassistant.api] Error on call http://172.30.32.1:8123/api/config: 
22-02-03 14:50:14 ERROR (MainThread) [supervisor.homeassistant.api] Error on call http://172.30.32.1:8123/api/config: 
22-02-03 14:50:19 ERROR (MainThread) [supervisor.homeassistant.api] Error on call http://172.30.32.1:8123/api/config: 
22-02-03 14:50:24 ERROR (MainThread) [supervisor.homeassistant.api] Error on call http://172.30.32.1:8123/api/config: 

These problems also appeared before the update to the latest version today
Unfortunately, I can’t find any clues as to what triggers these irregular restarts. Does anyone have any ideas?

Just after about 2 hours again a reboot out of nowhere. I am at a loss.

22-02-03 19:43:05 WARNING (MainThread) [supervisor.homeassistant.websocket] Connection is closed
22-02-03 19:43:42 WARNING (MainThread) [supervisor.misc.tasks] Watchdog miss API response from Home Assistant
22-02-03 19:43:57 WARNING (MainThread) [supervisor.misc.tasks] Watchdog found a problem with Home Assistant Docker!

Hey,
Mine is not restarting but it’s just getting unaccessible.
I put a WiFi plug to restart it… but today, WiFi plug also failed!
I will have to wait for the weekend to restart it :-((

I think I have determined the trigger. It must have been related to the Eufy Security Addon. This was the last change before the irregular restarts occurred. After deactivating all components associated with the addon, there have been no more restarts.

Owner of integration and add-on is here, i am not having this issue and no one reported this. Not sure how it is related or the root cause.

Unfortunately, I can not say 100% whether it is related to the integration (in whatever way). I had test last week Friday again activated the integration - Saturday was HomeAssistant out of nowhere again not accessible and only a disconnect from the power helped. So I deactivated the addon, since last Saturday no more problems. Logs unfortunately not meaningful or no error documented, so I can not identify the error / trigger exactly.
Very sad, because the addon is really great and I would like to use it very much

You can give a try install add on but no integration. We can find out if it is causing issues when integration is working or not.

So, home assistant has just said goodbye again after exactly one week and it only helped to briefly interrupt the power supply. So it can’t have been the Eufy integration / addon, as I still have it deactivated. The same error as above was in the logs. I am at a loss as to what the trigger is.

I have a similar random restart every 1-2 days, I only figure it out with telegram notification.
Nothing in the logs except in the supervised one. this is only observed after 202202xx versions and supervisor-2022.01.1. .

Any idea how to save the log of home assistant before restart, logs are vanished after home assistant container restart

 WARNING (MainThread) [supervisor.homeassistant.websocket] Connection is closed
 WARNING (MainThread) [supervisor.misc.tasks] Watchdog found a problem with Home Assistant Docker!

Actually I found the reason for my restart in home-assistant.log.1 which supposed to be old log before restart

it is something related to the streaming

CRITICAL (stream_worker) [libav.generic] Assertion next_dts <= 0x7fffffff failed at libavformat/movenc.c:1071
1 Like

Hello - did you ever fix the problem? I have the same issue with my HA installation. And there is no help to find anywhere

Check the end of home-assisstant.log.fault for errors

If it is due to the streaming component as me, my issue was due to the new latency streaming component, which was switched by default since that time. And caused issue with my single rtsp camera streaming lookback

What i did , i have to switch off it in the config and no random restart since then.

stream:
ll_hls: false

Thanks for you answar - so what you are saying is that in your configuration.yaml is putting this line in:
steam:
ll_hls: false

Coming back here as the owner of eufy security custom integration and related add on. I have been having similar issues on my setup and i am trying to find out root cause. I had disabled the integration and add-on, all is working fine.

I will run the add-on without the integration and update back here.

I will run both of them and disable hls and update back here.

Update: used add-on and integration for 12 hours with disabled hls config, restart happened. Before that, while add-on and integration were disabled and all were healthy.

Next steps: disable integration but keep add-on with disabled hls config and share updates.

I am trying to pinpoint if the issue is with add-on or the integration. If it is add-on, i have to sort this out with image owner, if not, python code in custom component is causing the memory leak.

Update: I had moved away raspbian based OS to Ubuntu based OS. All add-ons are working healthy and there are no restarts for last week. I am using supervised installation.