Hello Community , We will be using rsync to keep data flowing from a data-collection computer (data belongs to user "det") to a data-processing computer where the data is to be owned by whichever account is doing the experiment. I would like to run rsync and be able to specify the target UID/GID to use on the command line. For security reasons, the user doing data processing does not have direct access to the computer doing the data-collection so rsync must run as root which then is able to read the data transfer it and change its owner ship. I found the message below (from 10/2001) addressing this issue and it looked like the patch had not found resistance. Unfortunately I cannot get this to work (of course the man page doesn't suggest it should...). http://groups.google.com/group/mailing.unix.rsync/browse_thread/thread/50f6957086b822d/c113be869c374a63?lnk=gst&q=chown&rnum=2#c113be869c374a63 Can someone clarify if this is in the pipeline for rsync 3.0.0? Right now I am using a python script which will be doing the chown but I really think this feature would be very usefull in general, not to say *much* faster when dealing with lots of files. Thanks in advance, Ezequiel keywords: transfer superuser chown ownership chgrp
On Wed, May 02, 2007 at 11:36:58AM +0200, Ezequiel Panepucci wrote:> I would like to run rsync and be able to specify the target > UID/GID to use on the command line.It would be nice to have something like this. There are also some times where it would be nice to give rsync some more powerful hints for how to map the uid/gid values. So a more general solution might be to have some mapping options akin to this: --usermap=foo:bar --usermap=873:1001 --usermap='*:0' --groupmap=foo:879 --groupmap='*:baz' That would not be hard to implement given the existing ID mapping that is already happening. A daemon that was running as root could be configured to disallow those options for incoming data, if desired, so I don't see this as having any negative security implications. (Especially since an intrepid user could run their own custom rsync and supply whatever ID values they wanted attached to file-list data.) Thoughts? ..wayne..