Where do ditto'd files go if target disk is missing?

Question

Hi folks,
I've got a series of scripts on a server which do nightly/weekly/monthly backups to an external firewire drive using ditto:
ditto -rsrc /path/to/original /path/to/target
I had to remove the target disk for a while, and didn't bother to turn off the backup script. I'm pretty sure that the reason that my hard drive was filling up rather quickly was that ditto was storing the stuff in some temp files somewhere. One reason I suspect this is that when I cloned the drive, huge amounts of space - hundreds of gigs - were "miraculously" recovered, yet everything continues to function as before. And when I disconnected the external again with the new larger drive running, the usual daily backup amount (about 20 gigs) again 'disappeared'.
So, my question: Where might those phantom files be? I'd prefer to get rid of them manually without having to clone the drive if I ever forget to turn off the backup script again when the external is removed.

Answer 1

So if I just delete the original entry for that hard drive, that'll do it right?
If you mean can you delete the "false copy" - a large item in /Volumes on your source volume which is not an alias and has a name like your target - then yes.
When I reconnect my external, I'll be able to return the script paths to the originals too, I suspect.
As long as your external wasn't really re-named - you'll have to see.
re saving space by cloning-
Since I was cloning the actual drive's contents instead of a false volume as explained above, it does make some sense.
The actual drive DID contain the extraneous files - if you check with Disk Inventory X you will probably see a big /Volumes folder. I think the reason the space got reclaimed is that Carbon Copy Cloner is "smart" enough not to copy those files on the source that are not needed on the clone because they will get re-created there anyway. If you look at CCC's "Incremental backup" mode with "everything" selected to be copied, /Volumes is not on the list.
Mike Bombich has a page [here|http://www.bombich.com/mactips/image.html] for guidance to those who want to clone manually rather than use CCC. That page contains the following:
Some directories are populated by the System. For example, the Volumes directory is populated with directories corresponding to the names of Volumes you have on your system. If you insert a Zip disk, a new directory in /Volumes is created with the name of the Zip disk. These directories are called "mountpoints", and are created "on-the-fly" by Apple's autodiskmount utility. Because these directories do not contain data on your boot volume, they do not need to be copied during a clone operation. The Volumes directory is just a placeholder (and the OS will recreate the Volumes directory on bootup, so it is unnecessary to recreate).
Since the /Volumes directory in your case does mistakenly contain "real" files, not copying this directory onto the clone would "recover" the lost space.
Next question, though, is why not also use CCC for your scheduled backups instead of your ditto scripts? You can choose which files you want to copy. I don't think CCC will execute a scheduled copy session if the target is not connected (I'm not sure of this), and it's current copy engine is said to be better than ditto in preserving metadata.
Message was edited by: jsd2