On: How does rsync work?

[orangepurple]

rsync uses a heuristic based on file times and sizes to compare files. to compare file content use the --checksum feature (computationally expensive to run)

[srvmshr → orangepurple]

Yes but it doesn't answer why rsync & checksum pass the set of files as same, but diff reports them different.

[pronoiac → srvmshr]

It isn't clear from the conversation so far: do you use "--checksum"?

[Dylan16807 → orangepurple]

Rsync can checksum a lot of megabytes per second. In general I'd say the disk IO is much more expensive than the computation.