Hi,
I would like to make a back-up from OneDrive to WebDav. So it’s possible to use OneDrive as destination, but can I use OneDrive as source? I know it’s possible to sync all OneDrive files to my desktop and then backup to WebDav, but my local harddrive is to small to sync every file to my desktop. I would like to backup directly from OneDrive to Webdav.
Kind regards.
No, Duplicati doesn’t support using the back end storage types as the source. There is a feature request to add this on the Github project page, but I don’t think it is very likely to happen.
It will be awkward, but you might be able to add folders a few at a time (based on how much room the hard drive has) to Duplicati Source Data configuration, let Duplicati cause downloads of files as side-effect of a backup, then you can manually free up space by making the folders online-only. Repeat until done. Once a backup is made, Duplicati will only need to read files if it gets new files or timestamps, but they may be few and Storage Sense might be able to free space up automatically at a rate that keeps drive space available.
Save disk space with OneDrive Files On-Demand for Windows 10
Use OneDrive and Storage Sense in Windows 10 to manage disk space
You will face a similar problem on restores because Duplicati will want to look at original source file to pick whatever it can out of it to make the restored file, so if this becomes a major problem, use no-local-blocks:
--no-local-blocks = false
Duplicati will attempt to use data from source files to minimize the amount of downloaded data. Use this option to skip this optimization and only use remote data.
Duplicati (and most backup software) uses a local drive as source, however sometimes what looks like the local drive is actually a network drive. Performance may be poor, but sometimes it may be tolerable to run. True local drives are not only quick, but may also have speedups such as usn-policy to avoid full scanning.
I agree, and one challenge is being able to do it in a generic way while still having acceptable performance. There’s also information like destination timestamps that code doesn’t get because Duplicati has no need.
(actually, it’s possible that LIST can get times, but there’s are other source features that destination lacks).
Because @kamokagi talks about “possibility”, if a little-by-little plan isn’t appealing, test an all-networked:
rclone mount uses a FUSE filesystem. Its backup performance may be poor, but there’s one way to tell…
Probably quite a bit faster is for rclone to copy or clone which isn’t a versioned backup but might suffice…
This isn’t as protective from accident or attack as an actual versioned backup, but it’s better than nothing.