Eliminate the temp copy

jplorier at gmail.com jplorier at gmail.com
Tue Jan 28 13:35:23 MST 2014


Indeed, it's a fuse mount using curlftp. I use rsync because of the
differential copy to move only new files. Besides I understand that rsync
performs a verification after the copy which is not the case of cp that
only uses crc (maybe I'm wrong).
I may use rsync to directly copy files, but I'm not allowed to install
rsync y the target server.
Do you know of any limitation of rsync over curlftp?
Regards
El ene 28, 2014 6:19 PM, "Matthias Schniedermeyer" <ms at citd.de> escribió:

> On 28.01.2014 08:35, Juan Pablo Lorier wrote:
> > Hi,
> >
> > I have to sync two directories and one is an ftp mount. I had to set a
> > different tmp dir as tmp files are not allowed in ftp mounts so I see no
> > point in copying things to a temp dir if they won't be used for the
> > transfer.
>
> What do you mean by ftp mount?
> fuse?
> Is that on Linux or something else?
>
> If you DON'T sync to a regular filesystem, there are (or can be)
> limitations on what kind of basic operations are allowed on a file. So
> follows the second questions: Can rsync cope with the limitations. Which
> would explain the I/O-error in the other mail, rsync --inplace does
> something that isn't allowed/supported.
>
>
> But IFF it isn't a regular filesystem you are syncing to, the value of
> using rsync for the job is largely reduced. Rsync's strength for remoted
> transfers is "delta transfer" which can't be used in such a setup. In
> that setup rsync is reduced to a glorified "cp" (ignoring the capability
> to delete files). You could just use cp directly. And with "-u" you also
> have the basic capability to not copy files that didn't change in
> between runs. (Maybe with an added "-a", or at least
> "--preserve=timestamps")
>
>
>
>
>
>
> --
>
> Matthias
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.samba.org/pipermail/rsync/attachments/20140128/64943f2c/attachment.html>


More information about the rsync mailing list