Did these "free space" erases work?

Local Disk

New Member
I used Eraser to overwrite about 180 GB of unused space on a cheap laptop running Windows 7, first using the pseudorandom one-pass then the Russian two-pass. On the pseudorandom it took about 3 1/2 hours to get down to about 520 MB left of free space, then the program appeared to stall and accomplish little or nothing for the next hour or two, so I stopped it. On the Russian the results were similar: It took about 4 hours to get down to 28 MB of free space left, then progress slowed to a virtual standstill for the next 7 hours, dropping to about 20 MB of free space, then rising to 33 before I gave up and stopped the erase again. In both cases there were a few hundred "cluster tip" errors, and when I stopped the erases the random-character Eraser file in drive C disappeared on its own and that drive showed about 180 GB of free space again.

My question is did these erases actually overwrite everything but the remaining 33-520 MB? If so, then it would seem to be a virtual success. Also, did the Russian erase do both passes, or just the first before it stalled?
 
Unfortunately, you might as well ask if a tyre with a puncture is not in fact 99% sound ... :)

The thing with a free space erase is that it needs to complete fully before you can be sure that everything that was previously deleted has in fact been erased. The erasing method used makes no difference to this, nor does it affect the likelihood (which has always been a problem of the otherwise rather effective 'fill all the free space with random files' approach) that windows will decide to shut Eraser down when the drive is nearly full. If this happens, transfer anything you want to keep temporarily to another drive, and quick format the target drive before erasing the free space on it, and finally (if required) restoring your data.

The variant of this process I use quite successfully with laptops is to use whatever facility the laptop manufacturer provides to restore the machine to factory condition (obviously backing up any user data first), then installing Eraser and doing a free space erase before I do anything else. Also, on a system drive, I typically disable the option to erase cluster tips; there are so many files on a system drive that are locked by Windows, that erasing cluster tips is simply not worth all those error messages. But Joel might disagree ...

David
 
You didn't really answer the point of my question: Are you saying that in the examples I provided, the amount of free space not overwritten might not have been 33 to 520 MB as indicated by my computer before I stopped the two erases, but well beyond that amount--possibly 5, 10, 50, or a 100 GB? In other words, unless one lets Eraser fully complete free space erases, you can't go by what the computer says has been overwritten?

This problem of Eraser slowing down to a slug's pace towards the end of free space erases and taking forever to finish appears to be such a common and confusing situation that it seems the authors of the PDF Help manual should address it better.
 
Local Disk said:
You didn't really answer the point of my question: Are you saying that in the examples I provided, the amount of free space not overwritten might not have been 33 to 520 MB as indicated by my computer before I stopped the two erases, but well beyond that amount--possibly 5, 10, 50, or a 100 GB? In other words, unless one lets Eraser fully complete free space erases, you can't go by what the computer says has been overwritten?
I did answer the question, but, it seems, not clearly enough ... :)

What I said was that your question was, in my view, based on a false premise. Leaving aside the question as to how accurate the progress bar may or may not be (I think that it's pretty good as a general indication), the point is that even a small percentage of unerased space can contain just the data you wanted erased. The only way of knowing whether this is the case is to run a file recovery program and/or a disk editor on the drive, which is likely to be another time-consuming process.I believe that the only secure way to look as the issue is to regard any failure to complete as a 100% failure rather than as a partial success (hence the tyre analogy).

Local Disk said:
This problem of Eraser slowing down to a slug's pace towards the end of free space erases and taking forever to finish appears to be such a common and confusing situation that it seems the authors of the PDF Help manual should address it better.
Improvements in the documentation are under discussion within the team. But what really needs to happen is that the problem is more fully resolved. In so far as it is currently understood. It does seem to be system specific, because different systems seem to be tolerant to differing degrees of the drive being filled up with erasing files. How far this is a function of the state of the drive (how full, how fragmented etc.) and how far it depends on things such as drive firmware I really do not know. There may well be several factors involved.

Joel has made a number of changes to the development (6.1 beta) builds to address the problem; I don't know to what extent these changes were back-ported to 6.0.8. On the machine on which I most frequently test the free space erase, I am running 6.1.0.2284 without problems of this kind under Windows 7 x64. That build is available from the Eraser downloads page.

David
 
I've got a more aggressive fix in mind for this, when I do code it for 6.1 I will definitely back-port this to 6.0. Backporting is done only for major bugs like these, however (so features won't "trickle down" versions)
 
Back
Top