Re: [rdiff-backup-users] why does exclude move files to increments?

[Brandon Saxe]

Thanks for the responses! Thanks a lot, Chris. Your
post was very informative.

Yes, rdiff-backup does do basically what I want. I
would **love** to be able to tell it to only do
increments for the files I tell it has changed using
find -mtime 0. This would be nice for me to do on a
nightly basis. I would run full backups probably
weekly or monthly (this would capture the files
created with "cp -a").

After thinking about Chris' post. I guess I **could**
do my baseline, and then do an exclude ** to move all
files to the increments folder. **Then** I could run
subsequent runs only including those files using mtime
0. I haven't tested this yet, though. I am assuming
that this would indeed backup the files that have
changed, but would also be difficult to restore a
whole system to a certain state at a point in time as
all the increments would have different dates. I would
need to manually go through the increments tree to
find individual files to restore if the need arose.

Basically, it seems rdiff-backup maintains a mirror
(similar to a snapshot in time) of the includes, with
the differences being stored in increments for true
backup recovery (deletes, changes).

It would be nice to be able to control the mirror
aspect a little more with includes and explicit
processing.

I am thinking a directive such as --assume-stdin-only
or --assume-includes only or something like that would
be nice to have. Better yet, being able to specify the
--mtime 0 on rdiff-backup command line itself would be
nice. I guess in this case, the mirror aspect would
potentially be lost as deletes and moves would not be
one to one with the source tree. This is okay for me.
As long as I have at least one copy of all my files
from all time, I am happy. I can easily create a
backup of my filesystem tree into a sqlite database
and a script on a nightly basis to capture the system
'state'.

I am not a python programmer, but I would be
interested in learning where I could change the source
to control this behavior. Rdiff-backup is very nice in
one it does and I'd like to see if I can customize it
rather than write something else from scratch.

I looked at backuppc, but it sees too overkill for
something like this. I want simple and effective.

Any ideas? Which source files should I check out?

Thanks!

--- "Marcel M. Cary"
<address@hidden> wrote:

> If rdiff compares checksums of files to see whether
> they've changed, maybe 
> Brandon is looking to speed this up by assuming that
> files whose 
> timestamps haven't changed have indeed not changed. 
> In otherwords, trust 
> the filesystem timestamp and skip the checksumming.
> 
> Although there is certainly a risk that files will
> be created and have 
> their modification timestamp set to the past (like
> with "cp -a"), maybe 
> it's an acceptable risk in some situations?
> 
> If so, then rdiff-backup is doing almost what
> Brandon wants, but more 
> slowly.
> 
> Marcel
> 
> 
> _______________________________________________
> rdiff-backup-users mailing list at
> address@hidden
>
http://lists.nongnu.org/mailman/listinfo/rdiff-backup-users
> Wiki URL:
>
http://rdiff-backup.solutionsfirst.com.au/index.php/RdiffBackupWiki
>