Re: geting some file stat

[George Georgalis <george%galis.org@localhost>]

On Wed, Feb 20, 2008 at 03:02:14PM -0500, George Georgalis wrote:
>On Wed, Feb 20, 2008 at 02:14:00PM -0500, Thor Lancelot Simon wrote:
>>On Wed, Feb 20, 2008 at 10:47:49AM -0500, George Georgalis wrote:
>>> 
>>> actually in my shell code the restore case was missing
>>> quotes around the filename. This whole project is to
>>> bootstrap an offisite (rsync) backup where there is most
>>> of the files already but the uid/gid/mode are incorrect,
>>> so if I miss setting a few files, that's okay, just need to
>>> make most of them fixed to save bandwidth/time.
>>
>>Can't you just rsync again with -a?  That won't use any bandwidth, to
>>speak of.
>
>Normally yes, but in this case the backups are rsync
>hardlink snapshot based and managed with a script that
>does timestamp directory management, so a change in file
>stat is seen as a need to duplicate data and preserve
>data for each mode.
>
>Which is the box I'm thinking in, but indeed, a manual
>rsync -a may indeed be the easiest way to bootstrap
>proper uid/gid/mode on the backup side.

and indeed this worked for a massive saving of time.
(was using ignore mode/uid/gid before cause they kept
getting changed on big directories) got about 80GB of
data in sync over 100Kb link in under 6 hours! Thanks
for the tip!

// George


-- 
George Georgalis, information system scientist <IXOYE><