noob question re:pre-existing folders

When adding a new folder, syncthing will first index it by reading all files, hashing them and adding this information to a database. This initial index (called scanning) can take some time on large folders. Subsequent scans are a lot faster. At some point the indexes will be send to other devices in the cluster - then the actual synchronization begins.

Essentially, the behaviour for each new file is as follows (with both devices having their folders configured as send-receive):

  • If a file exists on A, but does not on B, it gets created on B.
  • If a file exists on B, but does not on A, it gets created on A.
  • If a file exists on both A and B, the following applies:
    • If the content, the permissions and the timestamp are identical, nothing happens.
    • If the contents are identical, but either permissions or timestamps are not, this soft-conflict gets automatically resolved.
    • If the contents are not identical: One of the two versions remains as-is, the other one is renamed to a .sync-conflict file.

(This also works with more than two devices. The above assumes two devices for clarity. A new file means a file newly seen by syncthing, e.g on the initial synchronization every file is “new”)

After syncthing has finished the initial synchronization for a file, a deletion of that file will be counted as a deletion and will be propagated to other devices.

Syncthing also has addtional modes, such as send-only folders. Those won’t accept any changes advertised by remote devices. Such folder modes obviously behave differently than what is described above.

Also, while syncthing generally tries to be safe from data loss, having a backup is always advised.