Error taking backup: Unknown error, see supervisor logs

Configuration
1

I’m using Hassio:

  • Core: 2021.5.5
  • OS: 6.0

Trying to back up the system before attempting to upgrade. But when I try to run a backup I get this error:

  • Unknown error, see supervisor logs
    image

In the logs I have the following details:

Error 1

Logger: homeassistant.components.hassio.handler
Source: components/hassio/handler.py:213
Integration: Home Assistant Supervisor (documentation, issues)
First occurred: 10:07:22 AM (2 occurrences)
Last logged: 10:07:37 AM

/snapshots/reload return code 404
/snapshots/new/full return code 404

Error 2

Logger: homeassistant.components.hassio
Source: components/hassio/websocket_api.py:109
Integration: Home Assistant Supervisor (documentation, issues)
First occurred: 10:07:22 AM (2 occurrences)
Last logged: 10:07:37 AM

Failed to to call /snapshots/reload -
Failed to to call /snapshots/new/full -

Further, when I attempt to upload an existing backup. I get this:

  • 404 Not Found
    image

I have plenty of storage space available. Any ideas?

2

Is it something to do with the path? Snapshots were renamed in Summer 2021. There’s a blogpost about it here:

1 Like
3

thanks @Stiltjack that looks pretty likely. There are environment variables that need updating.

So the next question is, where are these environment variables and how do I get to them?

4

Hi,
I have the same issue on a RPi4 - but not on a Proxmox VM version of the same system.

I believe the backups are still being made, but the browser times out. Certainly when I look in the backup directory, it is still growing, and eventually it appears in the list of backups.
Interestingly, in looking for this issue I found a tmp directory under the backup folder that has 10s of GB of files, which is why my RPi version had filled the disk! Now it has been cleared out, backups have dropped from 4GB to 500MB and things run much easier, but the backups still time out.

1 Like
5

Yes they’re still being created into the new folder name, but you can’t access them through ha. They can’t be restored if you can’t get to them. So they’re as useful as in they didn’t exist.

1 Like
6

This sometimes happens because of insufficient permissions of files being prepared for backup:

7

Hi,
Have you found a solution to this problem ?

8

Resolute on my side, I made multiple partial backups to find what was wrong, it was “Assist Microphene”, I wasn’t using it, so I deleted it and everything came back to normal.

1 Like
9

In my case I ended up rebuilding a brand new instance and migrating over. It was because the folder structure and backup/restore being different between versions, but I had missed an earlier update that restructured things to make them compatible. So there was no other way.