App restore not successful

[tecbox]

Hi,

yesterday my Immich-App crashed after aborted a backup process.
Since the app was not responding anymore, I tried to restore. After 6 hours (or more), I received mails about running out of disk space.

Server Details:
Cloudron 9.0.17
2 disks: main (nvme) = 240 GB now (doubled after first crash) and a second hdd-disk on which immich is home (changed data directory)
Immich package version 1.96.0
about 360 GB of data, backup from Hetzner Storagebox

The restore proceed as expected, after upload the json and fill private key and encryption key, it downloads the backup and stores on the hdd. But since this is finished, it seems to beginn from start and stores data on the main ssd (where not enough space is).
In my case it creates some additional folders in appdata like "U+JFw4C6ztAgbImq73rqNenmK4orcV3TMLuNPiamLNI".
In that are many, many folders and files (photos?).
So the end is, the restore download until the main disk is full and crashed then.

The data directory is, as mentioned, already moved to hdd disk.

Thanks for help!

[james]

Hello @tecbox

Thanks for reporting.

@tecbox said in App restore not successful:

The restore proceed as expected, after upload the json and fill private key and encryption key

I am gathering from this section that you have installed a new @immich app and imported the backup config JSON into that new @immich app.
If that is the case, did you set the home of that new @immich app also to the second disk?

[tecbox]

@james Yeah, first I tried backup from the app-menu, this crashed. It was a bit difficult to remove the app / data, but after some time I get it it uninstalled.
After reinstalling, I changed all settings to previous settings manually - RAM and also data directory.

Confusing is, the restore goes to the data directory, but after then, it created the mentioned other folders in appdata.

[tecbox]

Hello @all,
short feedback on this.

After digging some days around, it seems the problem is related to the configured sshfs-backup to my storagebox.
The backup take very long in the past, but since process must backup aroun 360 GB, I didn't think about it.
But, after searching and looking for the actual problem, I found this topic: https://forum.cloudron.io/topic/14848/extremely-slow-backups-to-hetzner-storage-box-rsync-tar.gz-replacing-minio-used-on-a-dedicated-cloudron

Long story short: I found that I configured the storagebox exact in the wrong way and the restore-process seems restored the right data (going to extra hdd), but also restores encrypted data (they look exact same as in the storagebox) and so the main ssd went full...
With another backup (puuh I had!) from another location, the restore went well and all is good.
I setup a new location to storagebox with recommendation to use the prefix and see, the first backup needs long time (as in the past), but the next backup, went successful in about 20 minutes! oO

Not so short, sry

[joseph]

@tecbox thanks for the update!