hi, there’s one thing I wondered for long time, let’s say there are 2 hosts sharing a same folder, and one host somehow has a corrupt file (e.g. ReFS integrity streams found a mismatch checksum), in this scenario, how syncthing will download good file from another host to recover the bad file?
This question decided whether we can use single harddisk with remotely backup via syncthing instead of using harddisk mirrow for important data.
Thanks.
One more thing to confirm, won’t syncthing delete good file from remote host?
Syncthing doesn’t know what’s corrupt or what’s not, it has no way of deciding that.
If the file has changed (which is decided by observing mtime/size/permissions change) it will assume the user changed the file and propagate that to everybody. If the filesystem index goes corrupt and the file disappears, it will also delete the file on every other device.
Syncthing is not a backup tool, because if you accidentally delete your backups yourself, it will go ahead and delete the data on the source.
Yes, different folder types might help with this, but you are using a tool which primary purpose is not backups, for backups, so there are quite a few rough corners where you might accidentially set something up incorrectly, or misunderstand what is being asked, and nuke your data.
This is not to say that it’s impossible to use syncthing for backups, but to say that when red buttons start appearing, you better understand pretty darn well what they do before pressing them.
Rsync on a cron is a pretty good backup tool, and has less rough edges where you can screw something up.
Thanks for your kindly reply 
Now I’m studing on ReFS terrible corrupted file policy, I didn’t notice that file system will remove corrupted file automatically. It’s totally unacceptable.
Thanks again, wishing you a good day!
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.