Feature: Archive old files to a different volume
 

SuperDuper! is currently very good at creating clone backups of the hard drive to provide an emergency disk to boot off of if the main drive crashes but doesn't really protect me from my own stupidity such as deleting a file then realising after I've made a backup.

To solve this problem it would be nice if any files that are being modified on the backup could first be copied to an archive folder elsewhere.

E.g. If I backup to /Volumes/Backup it could archive files to /Volumes/Data/Archive

This would allow the archive drive to be used for other things as well.

Standard archives would probably be one for each day of the current week, one for each week of the current month and one for each month (probably rolling for this one, so up to the last 12 months). Auto deleting older archives as they become obsolete.

This would lead to a layout such as:
/Volumes/Data/Archive/Daily/[Mon,Tue,...,Sun]
/Volumes/Data/Archive/Weekly/[Week1,...Week4]
/Volumes/Data/Archive/Monthly/[2006-12 Dec, 2007-01 Jan,...,2007-11 Nov]

The files would then be stored under the archive directory in a folder structure relative to their original path.

This archiving system would probably only be possible when performing smart updates.

Also, would probably need to be able to filter which directories are archived, as most people wouldn't want to archive everything, just perhaps their home folders.

This isn't something we have planned in the short term, KOTS, but note that it's exactly what Time Machine does... so using the two tools together, building on each's strengths, is something you might find useful.

I was thinking of getting Leopard for Time Machine but so far I've been unable to find any documentation on how it structures its folders when archiving and if you can use the archive drive for anything other than Time Machine.

The questions I've been unable to answer are:
1) Can you use the archive drive for anything other than Time Machine
2) Can you access the archived files through finder without using Time Machine
3) Can you tell Time Machine not to perform a full backup but instead use a different source as a base line (i.e. the SuperDuper! backup).

The thing I would most like is if there were a demo of Leopard available that I could try.

This might be something you should ask Apple...

Hi,
I agree completely! I'm looking for a backup solution for my Macbook which does that too. A perfect utility would provide me with a means to get my book running as fast as possible again while still allowing to save my backup _anywhere_ (external disk, network share, SMB share, FTP site, ...) and perhaps even provide a bootable rescue CD which can access the network share and reload the image. This could potentially be a simple Linux system with a menu-based backup / verify / restore options.

It would allow versioning (ie. multiple versions / revisions of files) while only storing the differences (so it would be possible to back up large files that change frequently, like VMware Fusion images, without running out of disk space) between file revisions. And it would ideally be a service that runs in the background all the time watching for file changes and storing these changes, for example, every hour.

Arconis DriveImage (for Windows) does all that, unfortunately I haven't found a tool for OS X that does something similar. Any ideas?

1) Yes. You need to tell TM which drive to use, but after you did that you can access it normally as well.
2) I think so, since TM does not archive differences, but only whole files. You might need to test this.
3) I don't think the SD backup can be used as a baseline.

Indeed, after having upgraded to Leopard time machine does store its data in a sensible way. It can't use another copy as a base line, but other than that everything is fine.

it stores its data in ./Backups.backupdb/<ComputerName>/<DateTime>/<files with original path>

So the files are perfectly browsable without time machine if required.