I’ve found SMART won’t automatically recover, only mark and report problem sectors.
Perhaps most importantly (I’ve found) is that you should re-run tests to see if the number of reported bad sectors increases or stays the same. I have one disk that has remained the same for many years, and although not heavily accessed is powered 24/7.
If numbers of bad sectors increases, then plan on replacing ASAP.
Sectors can also be marked bad by mistake.
If you want to try to recover those sectors, you will need to research recommendations by that disk manufacturer, some OEMs like Seagate make an image available for trying to recover those sectors. But, unless the number of marked sectors is very large, it’s probably not worth the trouble… Nowadays most disks are set up with a reserve of sectors to replace bad, so your total disk capacity isn’t affected.
It’s my impression though that most disk manufacturers don’t provide a way to recover bad sectors, the only way to do so is at the factory (You will need to send the disk to be “repaired”)
Take a look at what ‘smartctl --health /dev/sda’ and ‘smartctl --health /dev/sdb’ are reporting.
If the news is bad then, a new drive will need to be purchased.
[HR][/HR]The ‘smartctl’ man page offers some examples of how to run self-tests on the drives:For example, selecting LBAs 10 to 100 and 30 to 300: “smartctl -t select,10-100 -t select,30-300 -t afterselect,on -t pending,45 /dev/sda”.
As I posted earlier,
You can download and run Seagate Seatools to perform a low level disk re-format to try to recover bad blocks.
When you run a SeaTools scan, it will test each marked bad block to see if it can be written to and data read from it correctly, over-riding normal system behavior.
AFAIK this is the only publicly available tool that provides a way for a User to recover erroneously marked bad blocks, if it’s <really> bad (no mistake) then yes… In all other circumstances, it’s a problem that can be fixed only by the OEM Factory.
AFAICS, one of the most useful pieces of information for resolving the question “Is the disk nearing ‘end-of-life’?” is the following section of “smartctl --all” (following example is my SDD used for system directories – the values I consider to be important are highlighted):