Every single SSD carries two vastly different types of risk. There’s a gauge that indicates how much writing endurance an SSD has: this type of risk is visible, slow, and transparently communicated by each and every health monitoring tool available. However, there’s also a fuse known as the controller. This component doesn’t report anything until the moment when it fails catastrophically. To monitor the health of your SSD effectively, you need to read the gauge while demonstrating respect for the fuse.
Health percentages provide wonderful insight into how well flash is wearing down. Unfortunately, they don’t indicate whether the firmware or controller is faulty; these issues typically cause the majority of SSD failures. Read the gauge, but do not make the assumption that it represents some sort of warranty.
Windows Users: Install CrystalDiskInfo and you’re good-to-go; all SATA and NVMe drives will appear with their respective health percentages, temperatures and fully populated attribute tables. Your brand-specific dashboard goes one step further than CrystalDiskInfo on your branded products: Samsung Magician, Crucial Storage Executive, and WD’s Dashboard all include total writes over time and firmware updates presented in plain English. Mac Users will find the one-word SMART result in Disk Utility, and the entire story behind the drive using smartctl (smartmontools), installable via Homebrew, including the internal SSDs in your Apple silicon-based systems.
Regardless of the window that you have open to view your results, the four numbers that represent significant health metrics remain the same. Percentage Used: officially the NVMe wear counter, counts upwards from 0; alternatively, you may see this metric represented by an “inverse” friendly version commonly referred to as ‘Health’ Score. Total data written: measured against the TBW Endurance Value provided on your specific SSD’s specifications page; most drives are replaced before reaching anywhere close to the vast majority of their TBW Endurance Value. Available Spare: a reservoir of replacement blocks within which the controller can retire the old worn-out flash; Available Spare should always be at or near 100%. A visible decline in Available Spare will be your clearest indication of wear and tear that is being exhibited by your SSD. Media & Data Integrity Errors: a count with only one acceptable value, Zero, regardless of what the headline percentage states regarding your SSD.
Unfortunately most failures on a recovery bench are SSDs that have been delivered dead but not worn. Most SSDs have arrived at the bench with brand new wear statistics and a controller that was unable to respond, because of a firmware failure or electronic chip failure that caused the drive to go from being good to gone between two boots. The failed component in this instance is the same component that is used to write the status report (the health report) so none of the gauges indicated anything was wrong. This is simply how the product works; the Health Report indicates wear, but only a backup can protect against a fuse blow.
There is some good news after the failure. When the controller fails, the NAND Flash (and therefore your files), usually remains physically intact, locked away in an unresponsive drive. Direct Chip Reads and Firmware level repairs do exist for this purpose: recombining data that the Controller cannot read any more. This is what we use for our SSD not detected service and why losing an SSD is a recovery opportunity rather than an obituary.
Clean numbers: document today’s data-written number and note to check again in six months; checking an SSD is very easy-going. Spare slipping or integrity errors above zero: for the first time ever the flash tells the truth; copy all your data while the drive will still read, swap the unit out, and avoid the temptation to ‘verify’ using stress testing, as each full disk scan burns away even more of the exact resource you’re running low on. Slowdowns, freezes, disappearing-and-returning: these are symptoms related to the controller, and repeatedly cycling power on it is only going to make things worse; back up now in one smooth motion. Already vanished: cease reflashing, stop trying different cable connections hoping something works, and have a professional read it on the bench; our SSD & NVMe recovery service begins with a complimentary diagnostic and a written quote. The following is a single footnote related to SSDs specifically: On internal SSDs, TRIM deletes the block associated with deleted files within moments, so if you’ve accidentally deleted a file on an SSD, the backup has always been the plan.
No. Physical flash degradation (wear) and the state of the controller determines health; formatting simply writes the new file system on top of what existed before, and can use a fractionally greater amount of write budget. When you appear to “correct” an SSD by formatting after it exhibited problems, it was likely a logical error (and worth being taken as a warning sign, rather than correction).
Never; defragmentation is a routine for hard-drives which also damages SSDs. Defragmentation has no moving part to optimise for, therefore there is absolutely no benefit from performing defragmentation on an SSD; however, it does burn through some of your remaining write endurance to move around the blocks. Windows recognises this and will automatically run TRIM on SSDs instead of defragmenting them; don’t interfere.
It’s not a timer-based issue, but its service life is over. Typically at that point the SSD will start to behave as if it were read-only and there will be an increase in error rates. Any data or applications remaining on the SSD should be copied off immediately. Although reads are usually fine during this time frame, it is not recommended to continue using a failing SSD ‘until it actually fails.’ The last failure mechanism of an SSD may not always fail gracefully.
Controller, firmware and chip level SSD recovery on bench: free 48-hour diagnostic, written quote first, and in most cases you only pay if we recover your data.