[SOLVED] Hassio on docker --- failed to start after 106.5

Does supervisor 209 address this ?
I can see that the latest tagged is 209, but https://version.home-assistant.io/stable.json has 208 so its getting downgraded.

I’m running .106.2 and now reluctant to upgrade to .106.5 - I’m also running on a NUC / Ubuntu / Docker setup. Believe I saw yesterday that my system had upgraded to the 209 supervisor version, but now back at 208 - maybe there was an issue there?

Current snapshot:
image

I’m on a NUC and Ubuntu 18.04.4 LTS, supervisor 208 and 0.106.5 without any issues. The only thing i noticed that I had to restart my NUC after the upgrade to 0.106.5 to get all containers running again.

i think 207 is definitely the problem… how do i upgrade? what do i need to delete first?

I’m on 208 and I’m still not functional. The Home Assistant container simply won’t start. I tried ha core start and I just get

Processing... Done.

Error: Unknown Error, see logs

I simply don’t know what to try next. @Sweetpants I wish I knew what was different between your system and mine - I tried restarting my NUC as the first attempt at a solution but it didn’t change anything.

Edited to add: the error in the log after I use ha core start is

ERROR (MainThread) [supervisor.homeassistant] Home Assistant can't write pulse/client.config: [Errno 21] Is a directory: '/data/tmp/homeassistant_pulse'

I used this procedure https://community.home-assistant.io/t/absolute-beginners-guide-to-installing-ha-on-intel-nuc-using-docker/98412 when i setup my NUC for HA. I’m at Ubuntu 18.04, HA 106.5, and 208 supervisor. All is well. A new container was created for hassio_audio. It’s likely a permissions issue people are experiencing when the update attempts to create the new audio container.

I ignored the manual 207 Supervisor update, and let the update occur automatically with the HA update to 106.5.

In my case… i used portainer just now. Clicked on hassio_supervisor container. Chose Duplicate/edit and recreated with latest image. restarted all containers and i think im up and going again

think i lost hacs though

Hello everyone ,
Thanks for the tip , it’s ok for me with recreation with last image :slight_smile:
Thanks a lot

@Anwen I had this exact same error and it was the only error in the log. The Core container wouldn’t start, but everything else was up. I am running on a raspberry pi. I just found that folder and renamed it, then rebooted and now the Core container starts up.

1 Like

Oh, thanks for letting me know! I’m going to try to find that folder.

Edited to add: I’m not finding that folder…could you tell me how you found it, and exactly how you renamed it? Thanks so much!

I’m getting the same issue.

Client error on Websocket API Cannot connect to host 172.30.32.1:8123 ssl:False [Connection refused].

Tried the above steps, removing the supervisor docker image, trying to re-install, rolling back to supervisor 206. But still no go.

Is it safe to delete the supervisor container AND the image ?
Thats about the only thing i havent yet tried!

Out of interest, i rebooted the host again, as last time i did it was on supervisor 207.
Now that its on 208, im now getting an actual useful error:

ERROR (MainThread) [supervisor.homeassistant] Home Assistant can't write pulse/client.config: [Errno 21] Is a directory: '/data/tmp/homeassistant_pulse'

To me this looks like an issue others have had where the audio group doesnt exist, but in my case it does. Seems atleast a step in the right direction.

I tried that, it didn’t help me.

Everything was recreated, but the error still persists. The Home Assistant image is created, but is shown as unused in Portainer.

I pulled the SD card out of my pi and mounted the “data” partition, and it was under /tmp. Not sure that helps you…

2 Likes

I have success!

I found the folder in /usr/share/hassio/tmp … i moved the folder homeassisant_pulse out of the way.
I restarted the supervisor, but it didnt do anything.

Last ditch attempt – i restarted the host, and now its all back again!

2 Likes

This also worked for me.

1 Like

And for me! I’m a happy camper right now! I think we were all finding the same solution at the same time! Just FYI, there is GitHub issue here that helped me figure out exactly where to find the directory.

Thanks very much to all of you for your help in figuring this out! :star:

As part of this i rolled back to a snapshot on 0.106.2 — am i game to update to 0.106.5 now? :smiley:

There was obviously some kind of underlying supervisor issue too (or my nuc needed a reboot) as ha restarts were taking up to 5 mins before this all happened, and i just restarted in under a minute.

Yes, I’m on 106.5 and supervisor 208. All seems fine.