Refining log entries

Bob von Knobloch bob at vknobloch.de
Sun Apr 7 11:18:53 MDT 2013


On 07/04/13 17:49, Francis.Montagnac at inria.fr wrote:
>
> On Sat, 06 Apr 2013 14:52:59 +0200 Bob von Knobloch wrote:
>
>> Thank you, it wasn't timestamps, but group id inconsistency (the source,
>> in this case is a Windows box, ids are always tricky).
> ...
>> I must rewrite the group ID after transfer, because the destination gid
>> (special group), is neither the source (users) nor the invoking id
>> (root) and I see no option to rewrite it in rsync.
>
> There is none, effectively.
>
> If you are using ssh, what you can do is to:
>
>    - define a dedicated account for your backup at the destination with
>      a uid of 0 (ie: root) and a gid of the special group you want
>
>    - change you backup script:
>       - to use this account
>       - to use not -a but -rlptoD (ie: -a without -g)
>

Yes, that looks good.
I run a second command at present to chown and chgrp the results, your 
suggestion might be faster.

> If you are not using ssh, but a mount of your destination on the
> source machine, look at the possible mount options and use -rlptoD
>
>      Francis
>
> PS: Using "rsync -a $SRC/* $DEST" is unsafe:
>
>        - it will fail if $SRC contains a file whose name contains a space
>        - it may fail if you have a lot a files under $SRC:
>          Argument list too long
>        - will not copy dot files
>
>      Use instead: rsync -a $SRC/ $DEST
>
Thanks, I had already realised this and have amended my script 
accordingly (not that I usually create filenames with spaces, but 
someone else might).

Many thanks Francis,

Robert


-- 
The Sun is out, the sky is blue, it's time to drive the MR2.


More information about the rsync mailing list