Predicting SSD Failures:
Specific S.M.A.R.T. Values

SSD drives don’t have some parameters that are specific to magnetic hard drives. Instead, they have other variables representing overall health of the disk. S.M.A.R.T. tools calculate SSD health by analyzing the following variables: Reallocated Sectors Count, Current Pending Sectors Count, Uncorrectable Sector Count, as well as Percentage of the Rated Lifetime Used (or SSD Life Left, whichever is available).

SSD disk
  • Wear Leveling Count. This variable is vendor-specific. It decreases with time. When it reaches a certain manufacturer-defined threshold, S.M.A.R.T. reports the drive’s overall health as FAILED.

  • Erase Fail Count. The number of failed attempts to erase the content of a flash chip. Increase in this number may mean that flash chips are dying prematurely (before reaching their rated number of erase/write cycles).

  • SSD Life Left. Supported by few manufacturers, this parameter represents calculated lifespan remaining in the disk based on certain equations. When normalized, it reads 100 (100%) for healthy drives to 1 (1%) for dead SSD’s. Sometimes replaced with Percentage of the Rated Lifetime Used.

  • Percentage of the Rated Lifetime Used. This is the opposite of SSD Time Left. 1 means the drive is 100% healthy, while 100 means that 100% of the drive’s lifetime is used up, and the drive can be used as a small doorstop.

  • Grown Failing Block Count. Manufacturer-specific value representing the number of reallocation events. A rise in this value represents a problem with the drive.

Crystal Info: SSD Parameters

Generally speaking, lifespan of SSD drives should be easier to predict compared to traditional HDDs as there are no mechanical elements prone to unpredictable wear. With SSD’s, one can simply analyze wear leveling count to figure out how many write cycles are left, or read calculated variables such as SSD Life Left/ Percentage of the Rated Lifetime Used. Unfortunately, it’s not as easy. Early SSD drives were known for abrupt, premature failures with close to zero chance of successful data recovery.

The situation improves with newer models, but sudden, unpredictable failures still happen. SSD drives of all manufacturers (e.g. Sandisk, Transcend, etc.) go out of order unexpectedly: it worked just fine yesterday, but appears to be dead today. Ironically, it may be easier to predict a failure of a mechanical HDD by listening to unusual noises made by the drive or looking at certain other S.M.A.R.T. parameters. Either way, no degree of monitoring and S.M.A.R.T. analysis can replace a good backup policy. Make sure you always have a recent backup, and you may never need a data recovery tool.

Author: Michael Miroshnichenko

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)