handling renamed files

Ben Escoto bescoto@stanford.edu
Mon, 13 May 2002 10:31:47 -0700

Content-Type: text/plain; charset=us-ascii

>>>>> "DG" == dean gaudet <dean-list-rdiff-backup@arctic.org>
>>>>> wrote the following on Mon, 13 May 2002 10:03:32 -0700 (PDT)

  DG> have you considered a less bandwidth intensive method of
  DG> handling renamed files?
  DG> one scheme which might work would be to have a database on the
  DG> mirror which has a list of the source inode numbers for each
  DG> file in the mirror.  then when a file isn't found at its current
  DG> name in the mirror, an extra check is performed to find its old
  DG> inode ... and if it looks like a rename occured (i.e. similar
  DG> file size) then run the rdiff on it.

Hmm, one thing I've heard discussed on the rsync list is a --fuzzy
option, so that files with similar names would be investigated to see
if they are similiar.  It seems this would have some advantages (might
be able to catch similaries of some files that weren't renamed, like
maybe backup copies of files, and wouldn't require as much memory
(a complete inodes -> filenames structure could get big)) but also
some disadvantages (wouldn't catch renames across directories
presumably, filenames must be similar).

    Anyway, I haven't really thought through this - just thought I'd
mention it because the "fuzzy" stuff seems to be aimed at similar

  DG> i assume you've already got the inode listing on the master
  DG> because you're handling hardlinks now...

Yes, but only for hardlinked files (link count > 1).

Ben Escoto

Content-Type: application/pgp-signature

Version: GnuPG v1.0.6 (GNU/Linux)
Comment: Exmh version 2.5 01/15/2001