S.M.A.R.T. information view

S.M.A.R.T. (Self Monitoring, Analysis, and Reporting Technology) is a method for hard drives (and SATA SSDs) to report their health to the host PC. It does not apply to NVMe SSDs. NVMe SSDs have their own protocol, covered on its own page.

S.M.A.R.T. data consists of two parts, attributes, and logs.

S.M.A.R.T. attributes

S.M.A.R.T. attributes describe the condition of the drive. Each attribute is composed of the following data:

• Attribute ID, a number defining the meaning of the attribute.
• Attribute Name, which is the textual description of the attribute.
• Value (normalized value), ranging from 0 to some maximum (often 100, 200, or 255). The higher the value, the better.
• Worst, the lowest normalized Value ever recorded over the lifetime of the drive.
• Threshold. If the Value is below the Threshold, the drive has failed.
• Raw, or Raw Value, a 64-bit integer from which the firmware derives the normalized Value. Keep in mind that Klennet Recovery shows the Raw field in hexadecimal. There is no uniform specification for the Raw fields, which are vendor-specific. However, as a general rule, if the Attribute Name ends with "Count", the Raw field contains the number of events described by the attribute.

Some attributes are designated as life-critical. The drive itself provides this distinction. The attribute considered critical by one drive vendor may not be considered critical by another vendor.

Klennet Recovery highlights some attributes:

• in red, when the Threshold is not zero, and the Value is below the Threshold, or
• in amber, when the Raw value is not zero for certain attributes that I personally consider important:
  • Reallocated Sector Count, Reallocation Event Count, Current Pending Sector Count, Offline Uncorrectable Sector Count - all related to the growing bad sectors.
  • Spin Retry Count - indicating failure of the motor to bring the platters up to speed on the first try.

Sample S.M.A.R.T. attribute view for a SATA SSD. Note the raw value of the Reallocated Sector Count attribute, which is highlighted. It indicates 15 (0F hex) bad sectors were replaced with spares.

S.M.A.R.T. logs

S.M.A.R.T. logs provide some historical information about the drive. The logs are not available on AMD-based systems due to compatibility issues.

Different data may be available depending on the vendor and model of the drive. In general, you can expect to see some of the following:

• Device statistics log, including
  • Power-on data - total hours and number of restart cycles.
  • Temperature data - current and past drive temperature and sometimes time spent in overheat condition.
  • Disk media statistics - number of read errors, corrected or uncorrected; number of write errors and such like.
  • Flash media statistics - remaining or expending endurance and number of errors.
• S.M.A.R.T. self-test logs, showing the history of self-tests, if any. Note that self-tests must be explicitly initiated by the operating system. So it is likely to see the self-test logs on a drive from a NAS, but rare to see any on a drive from a desktop PC.

Interpreting S.M.A.R.T. data

Interpreting S.M.A.R.T. data is not an exact science. It is better than reading tea leaves, but it still is made in a significant part of educated guesses based on experience. There are only a few hard and fast rules:

• Having less than ideal Spin Retry Count or Calibration Retry Count is very bad. Make an image of the disk immediately. The total mechanical failure may be imminent.
• Having non-zero Raw displayed for Reallocated Sector Count, Reallocation Event Count, Current Pending Sector Count, or Offline Uncorrectable Sector Count is just simply bad. It means there are bad sectors on the surface, and the situation tends to deteriorate over time. I recommend you make an image of the disk if this is the case. However, if you do not want to make an image, then at least keep an eye on the values. Bad sectors tend to grow in number, but it isn't such an immediate threat as a failure to spin up.