7 Warning Signs Your SSD Is Closer to Death Than It Looks

The seven signs, in the order they usually appear

First come the file errors: a photo that will not open, a copy that dies halfway with a CRC error, a save that suddenly reports the disk is corrupt. Second, the freezes — seconds where the mouse moves but nothing loads, caused by the drive timing out and being reset; Event Viewer shows the resets as Event 129 from stornvme or storahci, and a drive that gets reset weekly is a drive on its way out. Third, the disappearing act: the SSD intermittently vanishes from Explorer or the BIOS, usually after waking from sleep — sometimes that is a firmware or power-state bug, but on a failing controller it is a countdown, not a quirk.

Fourth, writes far slower than the drive's old self: worn NAND and a shrunken SLC cache throttle sustained writes dramatically, though a drive stuffed above 90 percent full does the same thing, so retest with room to spare before panicking. Fifth, the SMART warnings — on NVMe, Percentage Used climbing and any nonzero Media and Data Integrity Errors; on SATA, Reallocated Sector Count and Uncorrectable Sector Count drifting up. Sixth, file system corruption that chkdsk keeps rediscovering. Seventh and terminal: the drive goes read-only, refusing writes entirely — wear-out protection kicking in, and the final courtesy before the end.

What to do the day you spot one

The order matters more than the list: the backup is step one because every diagnostic after it stresses the drive you are trying to save. The tools are free, take minutes, and read the same sensors the manufacturer trusts. Firmware deserves special mention — a noticeable share of "dying" NVMe drives, especially ones that vanish after sleep, are running firmware with known bugs, and a vendor firmware update has saved more drives than any tweak ever will.

• Back up the data you care about immediately — before any diagnostics, because scans are reads and writes on a sick drive
• Read SMART with CrystalDiskInfo: Caution or Bad status, or nonzero Media and Data Integrity Errors, means plan the replacement
• Check the vendor tool — Samsung Magician, WD Dashboard, Crucial Storage Executive — for health percentage and TBW consumed
• Stop write-heavy work on that drive: no defrag, no full-format "repair", no torrent reseeding
• Update the firmware if the vendor offers one; several "dying" NVMe drives were cured by firmware, not surgery
• Replace on a schedule once Available Spare approaches its threshold, not when the drive dies on its own terms

Why SSD death is more sudden than HDD death

Hard drives usually wear out mechanically, with months of audible warning; SSDs more often die by controller or firmware stroke, which takes seconds and skips the warning phase entirely. NAND wear-out itself is gradual — that is what Percentage Used and TBW ratings measure — but in a typical consumer drive the controller gives up long before the cells are exhausted. Practically, that means SMART monitoring catches maybe half of real failures, and the other half arrive unannounced.

Which leads to the conclusion nobody markets but everyone should hear: an SSD does not need a cleanup routine, it needs a backup routine. Check the health numbers quarterly, replace when the reserve drops, and treat any sign from the list above as a fire alarm rather than a suggestion. The drive that dies tomorrow usually showed you nothing today — and the one that showed you something today is telling you exactly how much runway is left.

Questions and Answers

Read next