You can’t — because Duplicacy mixes metadata chunks and data chunks in the same pile. If it didn’t — we could set different access classes on prefixes, storing metadata in hot storage and data in glacier. Then you could list files, do all metadata work, and when you want to restore, duplicacy would request thawing only of necessary data chunks, and ask you to come back later and then restore. That’s what Arq is doing. It’s feasible. It’s doable. It exists in the wild. I don’t know why it is not done in duplicacy still (other priorities, “why don’t you submit PR then”, etc, I understand )
Good point.
Whoa. Where did those numbers creep from?
I downgraded my acceptable data storage policy for super-duper important data to 2-1. My main copy of data, and my offsite backup.
If stars align and I lose data so be it. Probably I will. have more pressing needs showering all that ash anyway to worry about some data.