Restoring a single file takes a long time

I have a backup of around 180GB (volume size 50MB).
To test out the system, I tried restoring a single text file.
It’s currently still “Recreating database” as part of the restoring process, but essentially it has been about an hour and still no file.

Is there something I can do in order to be able to access a single (or a few) files from my backup in less time ?

Normally a database recreate isn’t needed as part of a restore, was there maybe some other error before the restore was started?

Depending on how many dblock (archive) files were used to store the version of the file you’re restoring along with your dblock size there may be a bit of download and decompression time involved, but otherwise it shouldn’t take very long.

Did you browse to an existing backup, or point to the backup target?

I pointed to the target

When you point to the backup target (from another machine), it has to recreate the database. I suspect that is what is taking so long, not the actual restore.

1 Like

I’m not sure I understand the idea of “restoring” a file from an existing backup.
Isn’t that basically copying a local file in this case?

For example, If you backup your Pictures, and you want to restore one of them, surely you wouldn’t restore it from your local computer (as there’s nothing to restore), but from the target backup destination.

The way Duplicati does deduplication, no - restoring isn’t just copying a file. It has to gather all the deduplication blocks together to form the file.

And as for how you did the restore…

  • “Direct restore from backup files…” means Duplicati has to rebuild the local “deduplication database” so it knows which blocks to re-assemble for your file.
  • “Restore from configuration…” process varies depending on whether or not the local “deduplication database” file as referenced in the configuration is found (and if not found, re-create it from the destination)
  • “HP-XW4600” (which in this example is a specific backup job I created) would likely see the the “deduplication database” file exists and use that for the restore (and if not re-create it from the destination)

image

1 Like

Adding to the other excellent explanations: you can use the backup job and choose “Restore” from there, as that will use the already existing database.

However, due to the way the data is stored, you may need to download several 50MB files to reconstruct a single file.

If your usage pattern requires that you frequently need to restore a single file, maybe you want a sync tool instead? Something like SparkleShare or rclone ?