[LUNI] Incremental Tar

[Martin Maney]

On Sun, Jun 08, 2003 at 05:31:10PM -0500, Patrick R. White wrote:
> So isn't this a good reason to use the dump/restore utilities to begin 
> with?

Maybe, but dump/restore is no panacea.  Back in 1991, at LISA V,
Elizabeth Zwicky of SRI presented a fascinating paper comparing the
performance and problems of then-extant version of tar, cpio, pax,
afio, as well as dump.  dump did well on most of the tests, but by its
design it is capable of really frighetening errors if the filesystem is
not quiesced (in practice, unmounted or mounted r/o would appear to be
necessary) during dump.  Also worth noting is that dump is quite
filesystem-specific, and I seem to recall hearing that the ext2 version
was interestingly broken a while ago.  Since I don't employ dump, I
can't tell you any more than that, sorry.

The only link I can find to the paper is from here:

  http://www.phys.washington.edu/~belonis/

There's a postscript file and jpegs of the printed document.  I thought
there used to be something less cumbersome, but Google isn't finding it
for me.  It did find a good number of now-dead links, though.  :-(

Ah, "zwicky backup torture" is a better search key.  Still mostly
passing mentions of this seminal work.  Here's a more recent survey
paper about *nix backup techniques:

  http://citeseer.nj.nec.com/chervenak98protecting.html

Here's another useful compendium that seems to be currently maintained

  http://www.cybertiggyr.com/gene/htdocs/unix_tape/unix_tape.html

OTOH, "cpio: Some Linux folks appear to use this" seems... odd.

Ah, google-diving!

-- 
A delicate balance is necessary between sticking with the things
you know and can rely upon, and exploring things which have the
potential to be better.  Assuming that either of these strategies
is the one true way is silly.  -- Graydon Hoare