...
That's ...
Behaviour confirmed. That is *not* cool.
On Fri, Jul 15, 2022 at 04:18:18AM -0400, Kevin Korb wrote:
> Yeah, unfortunately rsync's --checksum doesn't even employ obvious
> optimizations that the man page says it does. There is no reason for it to
> checksum files that only
Yeah, unfortunately rsync's --checksum doesn't even employ obvious
optimizations that the man page says it does. There is no reason for it
to checksum files that only exist on one end or are different in sizes
yet it does. Just try doing an rsync of a huge tree to an empty dir.
You won't run
--checksum is only slower than re-copying if your network connection
between the hosts is similar in speed (or faster than) each host's
local disk access.
If local disk access is 10x your network link, it is definitely not
slower than re-copying.
Having said that, it really is *very* slow, and
You should almost never use --checksum. It is slower than just
re-copying everything. You should almost always use --times (or
--archive which includes --times). Without this rsync is almost as dumb
as cp. Also, ssh has been the default --ssh for a long time.
On 7/14/22 04:22, Fourhundred
Hello,
I want to sync local folder to remote server. When I run follwing
command repeatedly, it always transfers everything each time again and
again:
rsync --rsh='ssh' foo/ server:/foo/
does it mean I have to always use either --checksum or --times, to
prevent repeated transfer of files
