This is correct, this does not work. This is because rsync transfers the complete file tree from A to B
Due to the way rsync works, it cannot work because rsync calculates several checksums before choosing to transfer a specific file (or parts of it) and do it in just 2 iterations (ping-pong steps).
This means that the file needs to be read several times. This would not work with a (potentially large) SQL dump, because somehow it would have to be buffered. And this buffering is up to the user.
In fact, saving a file should be the best way, especially if it is a file that receives only gradual differences.
source share