Same issue I had
Its was trying to update and kept falling over.
But I removed the latest which is the it was pulling down.
I had mine back up and running last night and managed to get it to upgrade to 103.6. I then had ad guard and plex running this morning and itās sent it offline again. I noticed home assistant was using a whole cpu unit, but my system still had 2.2 cores free. Usually my system idles around 5% cpu. Trying to get it back up again.
It seems like a network issue within docker as the logs show home assistant running and I can see the associated containers for supervisor and home assistant and dns. But not entirely sure why its coming up.
Itās alive! So after a lot of attempts, it turned out something in the config was making it very flakey. Iāve still got to narrow it down exactly but my finds point towards:
- Konnected
- Yeelight
Also I noticed a few addons were now making my server run hot, as in so hot the thermal protection was kicking in on the cpu. I will say all of these addons and configuration were working happily together until I last tried to update to 0.103.6
- Visual Studio Code
- Ad Guard (running but using more cpu than normal)
- Motioneye (running but using more cpu than normal)
- Plex
VS Code was by far the worst offender for thrashing the cpu, with something like 2.86 cores being consumed out of 4.
My server is fairly capable of handling all of this with an i5 and 16gb of ram and SSD. Iāve also made no configuration changes for a while, and have only noticed this since my original failed attempt to upgrade to 0.103.6.
Iām now fully upgraded and my supervisor is running at 195. It does keep tagging images with latest and the version, but to me this is fairly standard practice. All my addons are kept up to date too so I am on latest versions of all.
At the moment Iāve removed plex entirely as that seem to absolute thrash the entire system. Iāve got vs code disabled and Iāve set yeelight back to auto discovery instead of the explicit configuration I had. Iāve also completely removed the Konnected configuration as I believe it was that which was sending my cpu over the edge.
I still need to do more testing to fully establish the route cause(s). One thing I will say, I went through from an empty config adding one thing at a time, when I eventually did get to the Konnected section, removing it was problematic as my ssh became unresponsive due to the load on my server. Even when I eventually managed to update the config after a restart, using the hassio homeassitstant check command always passed, even on my full original configuration. hassio homeassistant restart didnāt always seem to reset my home assistant after removing an offending config section. Even killing the homeassistant and hassio_supervisor containers didnāt fix the issue the first time. I had to kill the containers and restart the daemon a couple of times before it properly reloaded. Iām not sure exactly but it seems like some configuration state is stored or cached somewhere and doesnāt always get removed when restarting hassio with its cli or even killing containers.
This is glances now, you can see homeassistant itself is running around 2-3% cpu, when this issue was in play it was using a whole core, I guess it maybe doesnāt support running over multiple cores?
You can also see motioneye is running quite high, previously that was idling around 11-16% and adguard was also much lower.
The supervisor also randomly seems to spike up now which Iāve never noticed before, literally in the time I was writing the previous comment its jumped up to nearly 2 cpu cores and then fluctuates between 1 and 1.8 cpu cores.
Any resolution? On mobile right now but the best I can explain my issue is, hassio on 103.6 was having add-on problems so I tried to resolve to no avail. With the pop up @nedde stated above. Thought a complete new install would work, so removed all of docker off a Ubuntu laptop, everything, clean wipe of docker.
Reinstalled docker and hassio, see they are up and running in portainer (assuming latest portainer). But canāt ip:8123 at all. Nothing, no where can I find hassio but I see it in portainer running. Please send your thoughts and prayers to me right now, agh! Any ideas?
Read the log files and find out why itās not starting up!
Yeah check your homeassistant container logs in portainer if itās not starting. Then my next port of call was the config, did you clear all the share/config folders when you did your reinstall or is your old config there?
If your old config there then like I did you may have to go through section by section adding bits back in.
Iād recommend copying your configuration.yaml to configuration.bak.yaml. Then delete everything from configuration.yaml. And reboot, that should most likely get you up and running again or at the very least rule out a config issue.
If it is a config issue, just add each bit back, bit by bit till it breaks.