A drive-health or S.M.A.R.T. warning is a backup alarm, not an invitation to benchmark, repair, update, or keep working normally. If the computer can still read the drive, use that remaining access to preserve the hardest-to-replace data first. Microsoft explicitly recommends backing up after a Windows critical storage warning, while Apple says a fatal hardware error cannot be fixed by Disk Utility or another software-repair utility. [1] [2]
Guidance checked: September 19, 2026 at 1:19 p.m. EDT / 17:19 UTC.
The problem in plain English
A storage device can remain readable while reporting low spare capacity, degraded reliability, read-only operation, or a fatal hardware condition. Windows documents those first three critical-warning states for monitored NVMe SSDs. Apple exposes S.M.A.R.T. information for many disks and draws a firm line between hardware failure and software-repair problems. [1] [2]
The costly mistake is using the warning as a cue to run every available test. A long scan, firmware change, operating-system reinstall, filesystem repair, cloning attempt, or hours of ordinary work consumes time and drive activity before the important files are safe. The safer order is document → classify the data → copy or preserve → verify the copy → diagnose → replace.
Who this guide is for—and who should stop
Use this response ladder
- Windows shows “Reliability is degraded,” “Spare capacity is low,” or “Drive is set to read-only.” [1]
- macOS Disk Utility reports a S.M.A.R.T. or fatal hardware problem. [2]
- A computer maker’s preboot diagnostic or the drive manufacturer’s supported tool reports a failure or warning.
- The computer still starts and the drive is stable enough to copy files.
Stop ordinary DIY work
- The drive clicks, scrapes, chirps, smells hot, was dropped or wet, or repeatedly disappears.
- The only copy contains customer, legal, tax, medical, family, or business-critical records.
- The disk belongs to RAID, a NAS, a storage pool, or an encrypted system whose recovery key is uncertain.
- The machine asks to initialize, format, erase, clean, convert, or repartition the disk.
If the disk is missing or appears as RAW, Unallocated, or Not Initialized, use our separate external-drive preservation guide. If the computer cannot find a boot device, start with the no-boot-device diagnostic ladder. A degraded multi-disk system needs the NAS and RAID-specific workflow.
Important Windows limitation
Windows’ built-in critical-storage monitoring does not cover every drive. Microsoft says Windows monitors NVMe SSDs, not SATA SSDs or hard disk drives. A clean Windows health screen therefore does not prove every installed disk is healthy. [1]
Use the computer maker’s or drive manufacturer’s supported diagnostic for the exact model after the data is protected. Do not install a random “driver updater,” “PC cleaner,” or unknown S.M.A.R.T. utility just because it promises a one-click repair.
The backup-first response ladder
Freeze the evidence
Photograph the exact warning and any error code. Record the computer model, drive model and capacity, date and time, whether the system is slow or freezing, and what changed before the warning. In Windows 11, Microsoft directs users to Settings → System → Storage → Advanced storage settings → Disks & volumes, then the affected disk’s Properties page for details. [1]
On a Mac, open Disk Utility, select the physical disk, and use the Info button to view available health information. [2] Record the result; do not start First Aid simply because the button is nearby.
Decide whether the data is replaceable
Classify the files before choosing a tool. If everything is already in a tested backup, you can move more quickly to diagnostics and replacement. If this is the only copy, treat the drive as evidence rather than a repair project. Shut down and price professional recovery when the files are worth more than another round of experiments.
For a small business, include accounting exports, active project files, email archives, line-of-business databases, password-vault recovery material, scanner folders, and locally stored customer records. A green cloud icon is not proof that every folder finished syncing or that a deleted/corrupted file has a usable version.
Copy irreplaceable files first
If the drive is stable and readable, copy to a different healthy device or service. Start with small, irreplaceable folders; then move to larger replaceable data. Open several copied files from the destination and confirm their dates and sizes. Do not “move” files, because a move can delete the source after copying.
Keep normal use to a minimum while copying. Pause games, large downloads, creative renders, virtual machines, local indexing, and nonessential backups that read the warning drive. If the disk begins disconnecting, throwing read errors, or making new sounds, stop and reassess rather than hammering it with retries.
Understand the warning—without treating it as a countdown
Microsoft lists three Windows critical-warning types: low spare capacity, degraded reliability from significant media-related errors or an internal NVMe-subsystem error, and a drive forced read-only. It also displays estimated remaining life, available spare, and temperature when supported. [1]
Estimated remaining life is a manufacturer prediction, not a guaranteed number of safe days. Microsoft notes that reaching the displayed life limit does not automatically mean immediate device failure. That uncertainty is a reason to replace deliberately after backup—not permission to ignore the alert. [1]
Run only the supported diagnostic that answers the next question
After the backup is verified, use the PC maker’s preboot storage test or the drive manufacturer’s supported diagnostic for the exact model. Start with a short, non-destructive health or self-test. Save the result, failure ID, S.M.A.R.T. summary, firmware version, and warranty information.
Do not enable a repair option, secure erase, sanitize, firmware update, performance benchmark, full-surface write, or sector-remapping operation until replacement and recovery decisions are settled. “Diagnostic” does not always mean “read-only,” and a model-specific tool may expose both safe tests and destructive maintenance on the same screen.
Separate hardware health from filesystem repair
CHKDSK, Disk Utility First Aid, Startup Repair, and reinstalling an operating system address logical structures or software startup paths. They do not make worn flash cells, failing heads, unstable electronics, or a fatal hardware error healthy. Apple explicitly says a fatal disk hardware error cannot be repaired by Disk Utility or another utility that repairs software problems. [2]
If no hardware error is reported and the data is protected, Apple says Disk Utility may be used for other disk problems. [2] Keep that boundary intact: software repair belongs after hardware-risk triage and backup, not before it.
Replace the drive and prove the restore
Replace a drive that fails the supported diagnostic or continues reporting a critical health condition. Match interface, physical size, capacity, encryption plan, warranty rules, and the computer maker’s service requirements. For a business device, schedule downtime and preserve the old drive until the new system, applications, files, encryption, and backup jobs are verified.
Completion means more than “Windows starts.” Open representative documents, confirm current accounting or project data, verify email and line-of-business applications, check encryption and recovery-key custody, run updates, and perform a fresh backup plus a restore sample from the replacement system.
Good response versus dangerous shortcuts
Good
- Photograph the warning and record the exact drive.
- Copy irreplaceable data to a different device first.
- Verify files from the destination.
- Use the exact maker-supported diagnostic after backup.
- Save failure IDs and warranty details.
- Replace and test the restore before retiring the old drive.
Bad
- Run benchmarks to see whether the drive is “still fast.”
- Update firmware before the only copy is safe.
- Initialize, format, clean, convert, or repartition.
- Assume CHKDSK or First Aid repairs hardware.
- Keep rebooting an intermittently detected drive.
- Trust one dashboard or one successful boot as proof of health.
Two realistic examples
Readable NVMe warning
A home PC shows “Reliability is degraded,” but Windows still starts. The owner photographs the warning, copies tax files and family photos to a separate drive, verifies them, records the supported diagnostic failure ID, and replaces the NVMe SSD. The warning created a recovery window instead of a data-loss event.
The expensive wrong turn
A bookkeeper’s laptop begins freezing and reports a storage warning. A long benchmark, firmware change, repeated CHKDSK runs, and a reinstall attempt come before the only local accounting export is copied. The drive disappears during the process. The safer first move was to stop normal work, preserve the records, and choose recovery or replacement before write-heavy repair.
Frequently asked questions
Does a S.M.A.R.T. warning mean the drive will fail today?
Not necessarily, and it is not a trustworthy countdown. Microsoft’s estimated-life value is manufacturer-derived and does not guarantee an immediate failure at the displayed limit. Treat the warning as a reason to back up and plan replacement now. [1]
Why does Windows say the drive is read-only?
Microsoft includes “Drive is set to read-only” as a critical-warning state: existing data may remain accessible while new writes are blocked. Use the readable window to copy important data; do not try to defeat the protection before the files are safe. [1]
Can CHKDSK or Disk Utility fix the warning?
They may address some logical filesystem problems, but Apple says a fatal hardware error is irreparable by software-repair utilities. Back up first and separate hardware diagnosis from filesystem repair. [2]
Should I clone the drive?
A clone can be useful when the source is stable and the operator has a correct destination, enough capacity, and a failure-aware plan. It can also spend hours reading a deteriorating device or overwrite the wrong destination. If the drive is unstable or the only copy is valuable, decide between professional imaging and ordinary backup before starting.
What if the health screen says everything is fine?
That result is limited to what the operating system and device expose. Microsoft says its Windows critical-storage monitoring covers NVMe SSDs, not SATA SSDs or hard disks. Persistent freezes, I/O errors, disappearing disks, or unusual HDD sounds still justify model-specific diagnosis and backup. [1]
Actionable next steps
- Photograph the warning, drive identity, and error code.
- Stop nonessential use and classify the value of the data.
- Copy irreplaceable files to a different healthy destination and open a restore sample.
- Stop if the drive is mechanically distressed, repeatedly disappears, or holds the only valuable copy.
- After backup, run the exact PC-maker or drive-maker supported non-destructive diagnostic and save its result.
- Replace a failed or persistently warning drive, then verify applications, files, encryption, and backup restoration.
Need help before the drive stops cooperating?
Practical referral: if a Port Saint Lucie-area home or small business has a drive-health warning, freezing computer, missing disk, or uncertain backup, contact The IT Guys before running write-heavy repairs or replacing the wrong device.
Sources
- Microsoft Support — What to do about a critical warning for a storage device (accessed September 19, 2026; Windows NVMe monitoring scope).
- Apple Support — Check if a Mac disk is about to fail (accessed September 19, 2026).
Scope: General backup-first triage, not a universal diagnosis or recovery guarantee. Drive model, interface, encryption, RAID/storage-pool membership, device-maker documentation, and the value of the only copy control the safest next step.