1
What can be done to find out if an SSD is bad, i.e., in bad health to the point that it's already corrupting data or about to do so?
There's smartctl
and its graphical version gsmartcontrol
, which visually highlights bad drive attributes. However, those tools were originally written for HDDs. SSDs work differently. Are there special tools for SSDs?
Background: I came up with this question when I saw that, after creating a backup image of an SSD and generating a checksum of that SSD, the checksum didn't match the image file, suggesting that the SSD might be bad. Link.
SMART status is the same between SSDs and hard disk drives. – JakeGould – 2018-08-11T23:48:07.190
1At the present time there is no way to reliably predict SSD failure. A failing SSD may have SMART warnings or corrupt data. But SSDs typically fail without warning or apparent cause. – LMiller7 – 2018-08-12T00:11:50.170
An error can occur on imaging a drive due to a cosmic ray, voltage glitch or other source. A checksum error, in itself, does not mean the drive is bad. I'd make another image and verify it. – DrMoishe Pippik – 2018-08-12T03:28:37.290
1
The ways SSD fails may differ from the way HDDs fail, and current tools do typically correctly report failure state. SSDs also offer other ways to not just see whether they are alive or dead, but at what point in their life they are at. I high recommend reading the SSD Endurance Test, particularly the discussions of the metrics being reported by the drives as they proceeded through the accelerated lives of the test: https://techreport.com/discussion/24841/introducing-the-ssd-endurance-experiment
– music2myear – 2018-08-14T17:19:10.580