--full-remote-verification
Checks the internal structure of each file instead of just verifying the file hash.
but that likely gets some complaints about Extra (better than Missing though). A known issue.
Pull request has been made, but pull request processing has unfortunately been flowing well.
This only goes so far and is not a replacement for an actual restore test or database recreate.
There is a recent forum topic on recreate having trouble that the usual self-tests didn’t predict.
Good practices for well-maintained backups has some other suggestions.
I’m testing with a script with 1 GB of data at default 50 MB dblock-size, changing 10% of the blocks on each run, using --threshold=5 to make it hyper-sensitive to wasted space. This churns quite effectively:
2023-08-15 10:43:21 -04 - [Information-Duplicati.Library.Main.Operation.CompactHandler-CompactResults]: Downloaded 29 file(s) with a total size of 1.12 GB, deleted 60 file(s) with a total size of 1.14 GB, and compacted to 4 file(s) with a size of 62.17 MB, which reduced storage by 56 file(s) and 1.08 GB
I was going to try 250 MB next, but I suppose I could go the other way. You can have the batch file if you’d like to play with it. The data-changer is a separate Python script similar to the one posted here.
Main loop is just 1) change 2) backup 3) pause so I can kill it if desired. No hang yet. I did see retries.