I am quite convinced there are “dead” files in my system.
I run HAOS with ~20 add-ons in a Proxmox VM.
Since a few days the “free disk-sensor” reports about 28 GB less free disk than usual. The nightly snapshots (that are moved to other storage by script after completion) are usually around 14 GB so it could indicate 2 hidden/failed snapshots somewhere.
I have probably restarted the server while the snapshot is being created/stored without noticing.
I suppose they are created in a temp folder somewhere and then moved to /backup?
Anyway, I have checked all the accessible folders, including /config and all other samba shares and the hidden folders are just a few hundred MBs all together. And there is nothing strange there. There are no files in /backup except last night’s snapshot.
My guess is that there are failed InfluxDB snapshots stored somewhere since they consume about 12-14 GB each which match the increase - and nothing else comes even close to that amout of data.
I read that I can access everything if I run SSH & Web Terminal and disable Proctected Mode, but I don’t want to do that since I have no idea what I’m doing in a Linux/Docker/whatever environment.
Any tips where an incomplete snapshot would be stored and how I could safely check if that is the case?
And as step two if my guess turns out to be correct, how to delete them?
Edit:
Found answer to my own question, saving the answer so I can find it next time:
There were ~40 GB of failed snapshots (backups) in /mnt/data/supervisor/tmp/.
- Access via console
- login
- ls /mnt/data/supervisor/tmp/
If there are files you don’t want: - rm /mnt/data/supervisor/tmp/*
With those removed, everything is back to normal.