Each time I download on my PC a lot of new photos from my camera, I backup them immediately. Later, gradually, I set metadata on the photos (title, location, etc…). And then I backup them again.
The average size of my photos is 6-10MB.
With Borg I can set the chunk strategy so this new backup doesn’t increase the backup size so much. With Kopia, the backup size increases as if I had added new photos. In others words, if I modify the metadata of a 8MB photo, the backup size increases by 8MB.
Yes, but I suppose that if the chunk size is larger than my photo, it creates a new chunk, even if I have modified only a few Kbytes of metadata into the photo.
What splitter algorithm are you using with Kopia ?
I don’t understand this question : I didn’t find any “splitter algorithm” choice into KopiaUI.
You can see the splitter algorithm option under Repository. It is set when you originally create a repository, so you likely are using DYNAMIC-4M-BUZHASH if you did not change it. You can change this to a smaller chunk size, but you will need to recreate the repository as far as I know. There are also performance tradeoffs with using smaller chunks (see the discussion at Deduplication efficiency), but if most of your files are small, then you may benefit from a smaller chunk size.