Hm. Good to know, back when i used to work on friends computers that was not the case. But those were mostly HP and Emachines, that sorta deal.
Maybe back in the Windows 2000 days. But SLP (aka SLIC) has been around since the Windows XP days.
You just had to have the correct install media provided by the manufacturer. Starting with Windows 8, you could use the generic ISO downloaded from the MS website to reinstall and activate OEM systems.
This was pre-2009, you know how many people actually held onto those manufacturer disks? (hint: nobody)
Don’t call me a nobody. Do you know who I am?
With a GPT drive, you might also want to wipe the secondary/backup partition table at the end of the drive as well.
Real answer:
Do an ATA Secure Erase[1] (use some live Linux from an USB, eg. Ubuntu) and install Windows 10 LTSB afterwards. Keep room on the SSD for Linux as well and install one afterwards.
[1] https://ata.wiki.kernel.org/index.php?title=ATA_Secure_Erase
ATA Secure erase is also a good practice when selling an SSD. It’s fast and simple, after you get used to the few commands it requires.
Hmm. I have not heard anyone call that “slack space” before; Wikipedia, for example, calls it a “reserve space” pool, and Ars Technica calls it “over-provisioning” space.
Anyway in the end, it’s still kind of a game of luck.
If I’m understanding correctly, your method assumes that the SSDs wear-levelling will work like so:
Initial state
|123456789| logical layout
|123456789----| physical layout
Pass A
|AAAAAAAAA|
|AAAAA6789AAAA|
Pass B
|BBBBBBBBB|
|BAAAABBBBBBBB|
so that by the end, all blocks have been overwritten.
However, there is no guarantee from the SSD firmware that all blocks will be overwritten. If the wear-levelling firmware decides that a certain block has too much wear on it, it might decide to not reuse that one block for some time; or if a block dies, it might never overwrite it again.
Granted, this is speculation on my part, but that’s the point each, SSD manufacturer might have different firmware, with different wear-levelling behaviour. It’s probably’s and maybe’s all the way down.
Yeah, that seems like the only real guarantee of anything, apart from software encryption, so I would go with that if possible. Although, as your link mentions, there can be problems if used over USB, Firewire/IEEE 1394, RAID controllers; basically anything that’s not a direct SATA connection.
its dell, you can just put the service tag on dell.com and get the latest drivers, Dell its the only OEM that i can get behind, they are a smooth sale when trying to look for drivers on the Page.
-Fallen
For hard drives, my method is probably pretty solid.
For SSDs, I think I have discovered a better way: Instead of overwriting the whole available area, just overwrite the same block over and over. The wear leveling algorithm will eventually distribute the writes to all physical parts of the SSD. However, it will be impossible to know exactly when that has happened. A couple full passes is probably a good idea too.
I think the ultimate take away is that you can never really know what is left behind. You either have to trust the manufacturer utility, or nuke it from orbit; it’s the only way to be sure.
1 Like
Ahh. I’m going crazy.
How would the wear leveling differentiate manually written data from what was intended as a wipe? Even all 0s is legitimate data.
So, after even more ruminating… I think my original method is still solid.
The only other method would be to use the manufacturer wipe utility, then rewrite random data to the same block over and over again.
Obviously, it’s anyone’s guess as to what really takes place with the wear leveling and storage algorithms. If it’s that important, you just have to destroy the media - literally pound it back into sand.
Sorry about the rabbit hole and about the tangent for everyone else in the thread.
Unless Windows 10 is trying to do forensic analysis when you reinstall 
END PARANOIAJust because you aren’t paranoid, doesn’t mean they aren’t out to get you.