Error: Another job is running for job group home_assistant_core

Installation Home Assistant OS
1

Running HA OS 10.5 on a NUC, Core 2023.10.0b

I was on Core 2023.9.3 and OS 10.4 Wednesday afternoon, updated to the latest versions at the time and Home Assistant was no longer accessible. I couldn’t access it by my Cloudflare domain or HA’s IP address, so I plugged into the NUC and poked around with the CLI.

Checking the Supervisor logs, I see an error that shows up repeatedly.
Error on call http://172.30.32.1:8123/api/core/state: [Errno 104] Connection reset by peer

Checking the Core logs, I see a warning referencing a database called home-assistant_v2.db.

INFO: Home Assistant Core process received signal 4
... A bunch of custom integration warnings ...
WARNING (Recorder) [homeassistant.components.recorder.util] The system could not validate that the sqlite3 database at //config/home-assistant_v2.db was shutdown cleanly
WARNING (Recorder) [homeassistant.components.recorder.util] Ended unfinished session (id=176 from 2023-09-28 05:07:28.966858)
INFO: Home Assistant Core finish process exit code 256

I want to try and get rid of that database file, but I can’t because when I delete it, it instantly reappears. I assume this is happening because the core is still running, so I try to run core stop. This throws an error:
Error: Another job is running for job group home_assistant_core

I tried core options --watchdog=false and restarting the system, hoping the core wouldn’t run at startup, but this did not work.

I can’t restore a backup because it gets interrupted by the same Error: Another job is running for job group home_assistant_core error.

So I’m at a loss. I don’t know how to do… anything with this persistent core error. Any ideas would be greatly appreciated!

1 Like
2

same here with 2023.10.0b0

stuck

3

Oh damn, found this thread right now.

Looks like the same problem as me… Any hints are welcome…

4

Same here. This is not funny

5

throw me on the pile. How do you paginate in the cli? I can’t see the backup names to try restoring. (i.e. backups | less)

Edit: I forgot I’m running on proxmox with backups. :facepalm:

6

I was able to roll back to 2023.9.3 thanks to @adamminer’s suggestion on @deadrabbit87’s thread concerning the same issue. Here’s my reply with exactly what I did.

Holy cow, this did it for me. From the CLI, I ran su restart, then immediately typed core update --version=2023.9.3 and kept pressing up arrow, then enter to repeatedly submit the command. I guess one of them went through and the display stayed on Processing... for a while until eventually it said Processing... Done. and successfully rolled back to 2023.9.3.

1 Like
7

Where did you enter su restart? In my installation the command is not available.

I login directly with the root account.

This is how i did:

  1. reboot with reboot now
  2. immedatly login with root
  3. enter ha core update --version=2023.9.3

but no luck…


# ha core update --version=2023.9.3
Error: System is not ready with state: setup
Error: System is not ready with state: setup
Error: Another job is running for job group home_assistant_core
# ha core update --version=2023.9.3

Error: Another job is running for job group home_assistant_core
# ha core update --version=2023.9.3
Error: Another job is running for job group home_assistant_core
# ha core update --version=2023.9.3
Error: Another job is running for job group home_assistant_core
8

I’m woefully unfamiliar with Linux and more advanced setups, but I ran the command to restart the supervisor directly from the HA OS CLI (with the splash screen). I think you should have access to it from wherever you’re able to run core commands.

I chose to restart the supervisor instead of rebooting the host because I imagine there’s less time to react. I can spam the core update command while the supervisor is in the process of rebooting, so I have a better chance of squeezing one through before the supervisor starts the core.

9

I had luck rolling back using two different terminals

From one terminal running rollback continiously by using watch command with no time delay.

watch -n0 “ha core update --version=2023.9.3”

From second terminal asking supervisor politely to restart

ha supervisor restart

Terminal one started update and completed with no problems in first try.

When update is done kill the terminal window with the watch command as that will keep running even after the rollback is completed.

s

6 Likes
10

Thanks, my HA is accessable again :slight_smile:

1 Like
11

@munklinde if I run that command I get;

Every 0.1s: “ha core update --version… HomeAssisant: Thu Sep 28 15:17:44 2023

sh: 1: “ha: not found

12

I’m getting the same thing… I feel dumb, but I’ve tried everything I have found, thus far.

13

it doesn’t sound like you’re in the CLI. I think? I’m pretty dumb with this stuff.
it should look something like this, I think: image

14

Thanks. I was definitely in the CLI. However, after the 4th reboot, I got it working. Whew!

15

Doesn’t work for me aswel. Looks like it’s stuck in startup or it’s saying the another job is running

16

It’s the same with my setup.
Try to Upgrade to 2023.10.5

Error: ‘Supervisor.restart’ blocked from execution, system is not running - startup

17

watch will repeat the action, don’t you want to upgrade only once?
so, may be you should:
ha core update --version=2023.9.3
and do a watch on the progress:
watch ha core info

if you remove the ", it will recognize the ha command.
which is:
which ha
/usr/bin/ha

By default, watch will repeat the command every 2 seconds