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.

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!