mild zinc
Hi guys,
I have a Home Assistant Yellow and I am running out of eMMC memory.
This is what I see with the du command: https://dpaste.org/WJRzc
I have a folder on my NAS mounted for my backups.
I have round about 100 devices paired to HA.
Any idea what I can do to find the source?
Thanks Hoppel
mild zinc
Hi, thanks for link. I installed gdu. I have 32GB eMMC. Is what I see under backup the used storage or the remotely used storage of my NAS? Maybe NAS was nbot available for some time and the backups were stored locally instead on my NAS.
ncdu shows me the same result.
mild zinc
OK, I scp'ed all local backup files from my ha yellow to my nas. After that I deleted the local files. So far so good. my disks metrics look fine now.
But I am not ready, I made a test backup and haos backups to the local storage on the HA Yellow instead of using my NAS. The remote directory on my NAS seems to be connected.
mild zinc
Oh, this is interesting. I made some further backup tests and recognised that the haos backup procedure uses my local data disk instead of my remote nas folder. I reentered the credentials, but it didn't help (look at the screenshot 'backup all test 2'). After that I removed the backup folder and configured it again. Now it works (look at the screenshot 'backup folder test 3').
This is critical! I did not recognise that the backups to my nas didn't work.
Is there any way to analyse this any further?
trim garnet
I've heard about remote storage having issues. The issues on github seem to be unanswered. You can check the host and supervisor logs for hints. Gdu has a -x flag to exclude other mount points.
mild zinc
The log shows this:
Logger: homeassistant.components.hassio
Source: components/hassio/websocket_api.py:135
integration: Home Assistant Supervisor (documentation, issues)
First occurred: 3:11:35 PM (2 occurrences)
Last logged: 3:11:46 PM
Failed to to call /backups/bd7d00c8 - Unknown error, see supervisor
Failed to to call /backups/8ed2a7cd - Unknown error, see supervisor
Check the supervisor logs for hints then.
You can switch between core, host, supervisor, etc at the top right.
mild zinc
Yes, I am already checking that. Here is what I see in teh supervisor log: https://dpaste.org/yaALC
trim garnet
You only need the logs with the same timestamp as the other error
+- a few minutes or so.
I'd try to remove the remote storage, then add it again and immediately check the logs for issues.
mild zinc
There are some errors like this:
ERROR (MainThread) [supervisor.backups.backup] Can't read backup tarfile /data/mounts/backups/d423be7b.tar: "filename './backup.json' not found"
and this:
2024-12-03 15:11:35.721 ERROR (MainThread) [supervisor.backups.manager] Can't remove backup bd7d00c8: [Errno 2] No such file or directory: '/data/backup/bd7d00c8.tar'
IMO the latter ERROR is not relevant, because I scp'ed some files from haos to my nas and after that delted them via ssh.
mild zinc
I'd try to remove the remote storage, then add it again and immediately check t...
Its already working. The solution was to remove the remote storage and re-add it. But I do not want to be in this situation again. I really thought that all my backups were stored on my NAS and not locally until I realized that my storage is almost full.
trim garnet
On a normal linux system I would run chattr +i on the mount point to avoid that. With HAOS you are reliant on the devs to do the right thing. Apparently they don't do that. I can only suggest to report it in the existing issues.