(tl;dr and questions at the end)
I recently migrated from Resilio Sync (old Bittorrent Sync) to Syncthing. However, I did things wrong: I had a huge folder which wasn’t synchronized by Resilio and tried to synchronize with Syncthing. Once I noticed that I changed back to Resilio, synchronized the folders and went back to Syncthing expecting that my two computers realized they were already on sync and that they would build an accurate Syncthing database.
That’s not what happened at all.
Even though the folders on both computers were at 55GB, and supposedly synchronized by Resilio, in Syncthing the “global state” is set at 80 GB, and both folders are increasing in size.
Furthermore, I realized that I hadn’t ignored the “.sync” folder (which includes old versions of deleted files in Resilio) the second time I started Syncthing, so I deleted the folders in both computers as the sync was going on. Somehow, those files are being downloaded, although they have no origin.
My first thought was that my files were being shared with a third party that I don’t know of and that’s where they are coming from. Trying to find the origin of those files I’ve found in the interface that the “Out of sync items” were being obtained from “Copied from elsewhere”. I tried to look for the meaning of that, but the only thing I found was an offhand mention by @AudriusButkevicius that
Copied from elsewhere means file was reconstructed locally from content available on other files.
To me it’s astonishing that what currently is more than 8GB of data (and apparently there are still 17GB in the way) were obtained from files in my computer, as I assume the information to reconstruct those files should be at least 1 or 2 GB (we are speaking of thousands of files, with their original metadata and full names), and I couldn’t find those files neither in the Syncthing folder (".stfolder") nor in my OS. It’s more like it is a thing of magic.
tl;dr: Some deleted files are reappearing on the system without a clear origin. Syncthing says that they are “copied from elsewhere”, but it’s as if they are been copied from a third, unknown source.
My questions are:
- How does that “file reconstruction” actually work?
- Where is the data to achieve the “file reconstruction” stored?
- Would it be possible to inform of what is happening more clearly to the user? As I looked for more information the 3rd party theory died a bit, but it was very creepy nonetheless.