Some unwanted behavior might occur in scenarios like the one you describe. Most probably, the change performed by one client will silently overwrite the other. If there is however a small gap between the operations, the earlier one will have a better chance of being picked up by Drive and detected by the other file system instance. In this lucky case there might be no data loss.

I would set the `sync_interval` configuration parameter to a low value to improve the chances of detecting changes as soon as they appear, but I would also try to make sure that only one client works on a certain file/directory at one time.

This case looks like a good area for future improvement. Thank you for addressing the issue!