Re: Using the --snaptime option

[Albert Peschar <albert@peschar.net>]

> No, not at all.  As Albert said, it's more of a --assume-modified-if-newer-than
> option (although "--assume-potentially-modified-if-newer-than" would be closer).

--assume-potentially-modified-if-newer-than-or-equally-new-as? ;)

On Fri, 24 Jan 2014 11:29:22 -0800
Colin Percival <cperciva@tarsnap.com> wrote:

> On 01/23/14 08:56, Daniel Staal wrote:
> > --As of January 23, 2014 8:32:14 AM -0800, Colin Percival is alleged to have said:
> >> That will effectively disable the "recognize when files haven't changed"
> >> functionality, which will force Tarsnap to re-read files which it might
> >> otherwise have not bothered to re-read.
> > 
> > So, essentially for this purpose it's the same as the `--newer-than` option, and
> > could be replaced with any of the --newer options, right?
> 
> No, not at all.  As Albert said, it's more of a --assume-modified-if-newer-than
> option (although "--assume-potentially-modified-if-newer-than" would be closer).
> 
> > (Of course, we are talking about ZFS snapshots here, which is already have an
> > atomic creation and can be browsed like any other filesystem.  I'm not sure if
> > that applies to other forms of snapshots.)
> 
> Atomic creation doesn't solve the problem of timestamps being too coarse grained
> to distinguish between a time just before and a time just after the snapshot was
> created.
> 
> -- 
> Colin Percival
> Security Officer Emeritus, FreeBSD | The power to serve
> Founder, Tarsnap | www.tarsnap.com | Online backups for the truly paranoid
>