Wipe/nuke limitation on SSD

[Stilez]

Worth noting that Wipe/Nuke is limited on SSD drives (which are increasingly being used, especially by skilled users who are the kinds of people who use XYplorer). Because of sector allocation and wear levelling on SSD's, wipe won't overwrite the same sectors the file occupies.

I don't know if anything can be done (probably not) but maybe there should be an option or warning, that wipe shouldn't try to do anything if the storage device is an SSD - it won't do anything except increase wear and won't make data any more secure.

[admin]

Thank you, good point!

I'll add a note of warning to the Wipe Now? prompt.

I currently do not know of a way to programmatically detect a SSD, sadly.

[PeterH]

In other words: wiping on SSD should be disabled, isn't it so?

The user might not always know which drive is an SSD.
And whiping there never seems to make sense...

Don, just saw your update. As I read: SSD is detected by number of rotations = 0. Do you happen to have access to that info?

[admin]

As I read: SSD is detected by number of rotations = 0. Do you happen to have access to that info?

A quick research brought up nothing. I will look deeper later.

[Stilez]

Forum posts on Google suggest that Windows detection can be unreliable (may not detect when it should).

Microsoft seems to use two detection methods:

• zero rotation ("Windows 7 will set Defrag off as a default, when detecting non-rotating media")
• "[R]andom read performance characteristics above the threshold of 8 MB/sec" ("few SSDs on the market today properly identify themselves as SSDs. 8 MB/sec is a relatively conservative rate. While none of our tested HDDs could approach 8 MB/sec, all of our tested SSDs exceeded that threshold").

It looks like read speeds are the main one used and more reliable, but it could require actual testing of the medium each time. Rotation is already known as part of the device information. So rotation might be the one to use...?

One feature that might work even on SSDs is "wipe spare space". Since spare space can be wiped by writing a file that uses all the spare space, wear leveling shouldn't have any effect. Over-provisioning might though, so research required. SSD manufacturers and similar might be able to comment on the issue technically if emailed - people like SandForce, Corsair, Anandtech, Intel.


Update - technical method how to detect rotation: "Windows 7 detects SSDs by using ATA8-ACS identify word 217: Nominal media rotation rate, with value 0001h as Non-rotating media like solid state devices. But not all SSDs adhere to the ATA8-ACS1 spec section 7.16.7.77" (source: MyDigitalLife.com, see also Google SSD detection word 217). The MyDigitalLife post didn't load for me but I could view it via google cache "text only".

Word 169 is also useful, see blog post "Windows, TRIM and other SSD Mysteries".

[admin]

Thanks for researching this!

However, since there is no safe way of detection (as it seems) I rather leave this task to the human brain. I assume a good majority of those who use wiping will know a hard disk from a flash card.