I'd like to be able to use a tar archive as a backend for running 1-way sync to the cloud - i.e. an uncompressed tar being used as a filesystem like store to do a sync up to a cloud drive.
My use case is large data imports which come as archives but which should be restored to an S3-like bucket service as individual files - and I'd like to intelligently skip already present files (i.e. based on size/date/modification etc.)
imo, that is not a realistic, given how most cloud providers work.
in addition, rclone copy entire files, does not support delta copy.
what other cloud copy tools and cloud providers offer tar backend?
tho, i could be wrong.
rclone is open-source, you are welcome to add a tar backend or https://rclone.org/sponsor/
There was already attempt to implemented something similar:
But it is still waiting for somebody to make it happen. It can be you @wrouesnel if it is something you need. I suspect it is extremely niche feature very few people would use.
I saw that work and I think it's targeting a different problem? I have a separate afero based extension of treating a tar file as a filesystem since an uncompressed tar file can be relatively efficiently iterated by skipping through header offsets. This works well enough to read through using fuse.
Sounds like a patch would be accepted here, do I'll see what I can do.
EDIT: the use case is "files in tar" become "a bunch of files on a cloud drive" without needing my users to do more then run "rclone sync tar:somefile s3:some bucket"
How sophisticated are the users? Potentially they could mount the tar file to the filesystem using something like ratarmount and then rclone sync the visible files.