SMART attribute 231 thestery: Mushkin SSD

smartsmartctlsmartmontoolsssd

Following from this question: How do I convince smartmontools that my SSD is not a hard drive?

… I now have the Mushkin MKNSSDCR120GB-MX's SMART attributes reading as "SandForce Driven SSDs". However I am still getting weird values from SMART attribute 231:

ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
231 SSD_Life_Left           0x0013   001   001   010    Pre-fail  Always   FAILING_NOW 21474836481

Googling around, it looks like this normalized value starts at 100 and works its way down to 0 (when the drive is at the end of its life and goes read-only). However this drive is < 1 year old and has had little use. Unless it's defective, I believe that this is incorrect, probably from interpreting the attribute incorrectly. I don't believe the drive is defective because the other relevant attributes (Reallocated_Event_Count, Program_Fail_Count, Erase_Fail_Count, Retired_Block_Count, etc) are all reading 0, and the drive is not malfunctioning in any way I can discern.

The other clue is the raw value. From what I've read, on SandForce-driven SSDs, the raw value for attribute 231 should always be 0. But one day, out of the blue, on my drive it changed to 21474836481 and stayed there.

In binary, I see that 21474836481 is two 32-bit words, with bits 0 and 2 set in the most significant word and bit 0 set in the least significant word. In other words, 5 and 1 in decimal.

In a Kingston document (under "raw usage" of attribute 231), they say that they are using it in much this way, i.e. they are using some bits in the field as flags. But this is a Mushkin-brand drive that may or may not have a SandForce controller, not a Kingston drive.

Does anyone know how I can interpret this value for this drive?

Best Answer

The Mushkin spec page says it is SandForce powered.

The SSD life left if that is indeed the attribute has no direct relation to the other values. There may or may not be another problem.

The thing is that the SSD life left only counts the number of writes done to the NAND and compares it to the specified limit for that NAND, so if the NAND is specified for 10,000 writes and it had gone 9900 writes it will shows 1% remaining regardless if the NAND is still ok.

The NAND while being specified for 10,000 writes may work much more than or it may fail soon thereafter. It can also fail before that since no hardware is perfectly manufactured so the spec is mostly a guideline.