Recently I had a disk fail whose contents I was backing up to a local drive.
However there are two issues that are complicating the restore - the local database appears to be in need of a repair/rebuild and the backup set has a file that may be corrupt (based on the warning I get doing a command line restore).
I tried a command line restore, but based on the size of the backup (1.2 TB) and the time it takes for each “Downloading file” step at 200MB / chunk suggests that will take weeks/months.
Likewise, attempting a restore via the GUI of all files gets me in to the “recreating database” scenario where it looks like it’ll take weeks as well.
What’s odd is that Duplicati’s restore interface is aware of the files in the backup set and let’s me navigate them, so on some level the database is intact, yes?
Is there a more efficient restore method / option I’m missing that allows me to just restore everything or otherwise avoid a weeks-long process?
Happy to attempt any set of parameters and post relevant logs on such attempts. I upgraded to the latest canary build as part of my troubleshooting.
Do you recall the warning? Did it talk about a file from the backup drive or a source file you tested?
Is that about 10 minutes each? Seems rather long but it’s hardware dependent. Is this direct connect by USB on a reasonably powerful system? What do your system monitors say about CPU and disk loads?
Sometimes USB 3 connections to the drive will run significantly faster than USB 2. Do you have USB 3?
Are the files encrypted? If not, the performance you mention seems even worse, if it’s just a local zip file.
I think it only does this if the database is missing. The Database tab shows its path. Is that file still there?
If the database exists and is intact, that would be fastest. If not, direct restore is probably the next fastest, unless you need to restore many times (because it’s a partial temporary DB built for a particular restore).
Seems like yes (on some level), so getting conflicting signals…
2.0.4.28 would be good (but 2.0.4.30 is newer). If you managed to get 2.0.4.29 it had a performance bug.