Is there any value in running a drive-level (not partition or filesystem-level) "surface test" on an SSD?
These are the kinds of tests which look for bad sectors.
Super User is a question and answer site for computer enthusiasts and power users. It only takes a minute to sign up.Sign up to join this community
A test originally designed for magnetic HDDs would not produce meaningful or reliable results on a SSD. Trying to perform a "low-level" test through the high-level SATA/ATAPI interface will still employ wear-leveling to render results meaningless.
Typically a surface test would write a particular data pattern to each sector, and then verify each sector. The cycle would repeat with a different data pattern to confirm that each sector can be written, read, and retain data. Unless bad-block remapping kicks in, you are assured that when you write to sector X five times you have actually accessed the same physical sector on the platter of the HDD. Since you're actually rewriting the same physical sector each time, you'll actually test that sector with different data patterns.
The NAND flash used in SSDs require block erasures prior to writing. The number of erasures that the flash can tolerate is finite. So wear-leveling schemes are employed to distribute erase and write operations across all erase blocks of the device.
That means that a second (or subsequent) write request to the "same" sector of a SSD will not test the same physical sector of NAND flash that was written the first (or prior) write. The surface test will not be able to actually write different data patterns to the same sectors.
The NAND flash in a SSD would have to be tested without wear-leveling involved. The embedded controller of the SSD already monitors bit errors while performing routine operations and garbage collection, and can also perform bad-block management.
A surface test (intended for HDDs) on a SSD would probably only provide you with a false sense of reliability (and a reduction in the life of the SSD).
If you want to perform a test of the drive, then use the extended/long SMART test. You can run that test as often as you want without affecting the life of the SSD.
Yes, there is every reason to run a surface scan, just like on an HDD. The electronics or software may not know there is a problem until you read or write every sector on the drive.
I’ve discovered bad or failing SSDs like this.
SMART is not a reliable indicator that a drive is GOOD. It is, however, a reliable indicator that a drive is BAD.
I do, however, feel that surface scans should be used sparingly on an SSD. They are much less prone to failures of this type. So do it if you feel there might be a problem, rather than as a maintenance task.
I would recommend to just check the S.M.A.R.T values of the SSD to see if anything is out of the manufacturer stated normal values.
Manufacturers generally design a SSD to give indications of it's life span, such as 'wear-leveling' counts.
A good tool to check this is CrystalDiskInfo: https://crystalmark.info/software/CrystalDiskInfo/index-e.html