On Thu, Feb 26, 2026 at 13:37:55 -0500, [email protected] wrote: > Thanks to all who replied! > > One followup question and then some questions on rsync operation (which I > probably don't really need to know). > > On Thursday, February 26, 2026 11:22:54 AM Greg Wooledge wrote: > > On Thu, Feb 26, 2026 at 10:49:12 -0500, [email protected] wrote: > ... > > Assuming a one-time interactive copy, what I would normally do would be: > > > > cd /rhk > > rsync -a . /back/rhk/ # or cp -a > > If I know that /rhk exists, is there anything wrong with: > > rsync -a /rhk /back/rhk/ # or cp -a
That will create /back/rhk/rhk. > But, overall, it seems like a lot of overhead, including reading both the > source and destination file, calculating checksums, and then (partially) > rewriting the destination file. Seems worth it for a remote system, but I'm > not sure about on the same machine. In the most common case (a backup that's performed multiple times), a lot of the files will have the same name, size, owner, group, permissions and timestamps. Rsync will skip those entirely. The checksum algorithm is only used for files that have been altered. I'm not intimately familiar with the algorithm, so I don't know at what point it says "Hey, this is a whole new file, so I'm just gonna copy the whole thing", but I assume there is such a point.
