I’ve tried to set up several ‘send only’ folders.
The folders are different shares on my unRAID server. (Syncthing running in a docker).
To do this, I mount the server root into syncthing as ‘/sync’ (readonly)
It will send my data to another offsite server.
I don’t want to allow rw access because that’s very bad practice.
Why does syncthing require rw access? For send only, any database files syncthing needs could be stored in /config.
Mounting server root is insane, it has nested never ending folder structrues exposed via virtual filesystems etc.
Syncthing needs to create a marker file to be able o distinguish between a mount failing and everything being deleted.
You can rename marker name in advanced settings and point it at some well known file or directory that always exists.
I can’t see another way but to mount root into the containers filesystem, then create a ‘send-only folder’ for each ‘share’ within it that I’d like to duplicate on the receiver server. A share is just a folder somewhere within root.
I’ll take a look at advanced settings to change the marker file location - thanks.
With send folders: Can syncthing not check if a folder is mounted, verses if a folder is empty? If the source folder is missing (not mounted), then that would be an error and nothing should be sync’ed, verses if it exists but is empty, then the destination should be sync’ed (emptied) to match.
This is what the “marker” is for. Its presence indicates the folder is mounted and healthy.
So the marker must be in the root of the folder that’s being sync’ed? I can’t allow rw access to my data.
The point of being able to chose the name is that you can point to something that already exists.
Ahhhh… I see. So I can manually create the required file and/or choose a name in advanced settings.
Syncthing doesn’t need to create the file it’s self. Correct?
Correct.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.