Thanks, and indeed my investigations into possible a/m/ctime problems were a red herring. Something learned.
Regards, Charles On Sun, 2 Mar 2014 18:48:03 -0500 Nathan Stratton Treadway <natha...@ontko.com> wrote: > On Tue, Feb 25, 2014 at 10:53:07 +0100, Charles Stroom wrote: > > I than repeated the whole sequence, with the > > "--atime-preserve=replace" added, and there is no change: again > > "full" incremental backup. > > > > And than again but now with "--atime-preserve=system" and now this > > works! > > > > Right. If you include "--atime-preserve=replace", tar actually has to > explicitly set the atime back to its original value -- and in the > process it changes the ctime for the file, thus triggering that file > to be included in the next incremental dump... > > On the other hand, ""--atime-preserve=system" causes tar to tell the > OS essentially to pretend that particular access never happened in the > first place, and thus there's no need to go back and reset the atime > later. > > (This topic is discussed on the File Attributes page of the tar > manual: > http://www.gnu.org/software/tar/manual/html_node/Attributes.html#SEC137 > ) > > Note that it's an updated ctime that would cause the file to be > included in the next level 1 dump. Changed atimes don't trigger the > problem -- otherwise just looking at a file would cause it to get > included in the next dump.... > > > Is there an easy way to get this option added to tar in amanda, or > > do I need to recompile? > > I'm fairly certain that Amanda will never send > "--atime-preserve=replace". > > You can find out what your version of amanda is using by searching for > the the '/bin/tar...' line in the > /var/log/amanda/client/[CONFIG]/amgtar.*.debug log files (or > equivalent). (In my Amanda 2.6.1p1 installation, I see it's sending > "--atime-preserve=system".) > > In any case, (as you've already realized later in the thread) in the > case of actual Amanda runs, your always-full-sized-dumps problem is is > probably not related to --atime-preserve. > > Nathan > > ---------------------------------------------------------------------------- > Nathan Stratton Treadway - natha...@ontko.com - Mid-Atlantic > region Ray Ontko & Co. - Software consulting services - > http://www.ontko.com/ GPG Key: > http://www.ontko.com/~nathanst/gpg_key.txt ID: 1023D/ECFB6239 Key > fingerprint = 6AD8 485E 20B9 5C71 231C 0C32 15F3 ADCD ECFB 6239 -- Charles Stroom email: charles at no-spam.stremen.xs4all.nl (remove the "no-spam.")