Inside the lab follow MDrepairs on YouTube, TikTok & Instagram - visit our YouTube channel ↗

Data Recovery Guide

Windows Stuck 'Repairing Disk Errors'? Careful Now

Updated November 18, 2026
In this article
Short answer: "Repairing disk errors. This might take over an hour to complete." means Windows detected filesystem damage and is running an automatic repair against your drive at boot. Here's the uncomfortable truth: this screen very often appears because the drive itself is failing, and the repair process is hammering a struggling drive with hours of intensive reads and writes. If it finishes in under an hour on an SSD, fine, let it. If it's been stuck for many hours, or you hear clicking, whirring or grinding, or this screen keeps coming back every boot: stop. Pull the power. A drive with your data on it is being stress-tested to death, and the data-first path beats the repair-first path from here.

This screen puts people in a genuinely unfair spot: Microsoft's own advice is "wait," every forum says something different, and the thing you're waiting on might be eating the only copy of your files. Let's sort out what's actually happening, when waiting is right, and when it's the worst possible move.

What Windows is actually doing on that screen

At boot, Windows found the "dirty bit" set on your system volume, a flag that says the filesystem wasn't left in a clean state. So it launched its repair machinery (the same engine behind chkdsk) to scan structures and fix inconsistencies before loading the OS. On a healthy drive that was simply shut down mid-write (power cut, forced restart, crashed update), this is routine maintenance: it runs, it fixes bookkeeping, Windows boots, life goes on.

The problem is why filesystems usually get damaged on a machine that wasn't unplugged mid-update: sectors that failed to read or write correctly. Bad sectors. A drive that's starting to die. Windows can't tell the difference between "untidy after a power cut" and "bleeding out," so it runs the same aggressive repair either way, and against a failing drive that repair is a marathon of exactly the workload the drive can no longer handle.

When to wait, when to pull the plug

Waiting is reasonable when:

  • It's the first time you've seen this screen, and
  • There was a plausible trigger (power outage, forced shutdown, interrupted update), and
  • The machine has an SSD (no moving parts to grind), and
  • You have a backup of anything irreplaceable, and
  • It's been under an hour or two, with the activity light flickering like work is happening.

Pull the power when:

  • It's been stuck for many hours, or overnight, especially with steady rather than flickering activity.
  • You hear clicking, buzzing, grinding, or rhythmic seeking from a hard drive. Mechanical noise plus repair screen is the classic prelude to a dead drive; our clicking guide explains what that sound physically is.
  • The screen returns every boot. Repair, boot, repair, boot: that loop means the "repairs" don't stick because the hardware underneath is failing. Each lap makes things worse.
  • There is no backup and the files matter. This is the deciding factor that outranks all others. Hold the power button, shut it down, and stop booting from that drive entirely.
Why pulling power is the conservative move: it feels wrong to interrupt a "repair." But chkdsk-style repairs fix the filesystem, not the drive, and they fix it by writing. On a drive with failing sectors, the repair can relocate, truncate, and orphan your actual files based on misread data, doing permanent logical damage on top of the physical decay. Recovery labs see the aftermath weekly: drives where the failure was survivable but the repair pass wasn't.

Two related screens cause the same dilemma and follow the same rules: "Scanning and repairing drive (C:)" with a percentage, and endless "Attempting repairs" under Automatic Repair. Percentage counters that freeze for hours, or percentages that climb and then restart from zero, are the same signal as a stuck repair: the process is re-reading sectors the drive can't deliver. The decision tree above applies unchanged, and so does the data-first exit.

The data-first path (what a lab does instead)

The professional sequence flips the priorities: secure the data, then worry about Windows.

1. The drive comes out (or the machine stops being booted from it). Every boot attempt runs more repair against it.
2. It gets imaged, weakest areas last. On a hardware imager, the whole drive is cloned to a healthy one, reading the stable regions first and coaxing the damaged regions gently at the end. Your files come off the clone, never the patient.
3. Repairs happen on the copy. Filesystem fixes are safe when they run against a full image on good hardware; if a repair goes sideways there, we re-clone and try differently. Nothing is irreversible anymore.

That ordering is why "stuck repairing disk errors" cases that arrive powered-off recover so well, and why the ones that arrive after five days of boot loops sometimes don't. If the machine's already unbootable either way, the honest comparison is: a reinstall fixes Windows for free, but only recovery gets your files back, and only if the drive wasn't run into the ground first.

If Windows eventually comes back: your one job

Sometimes the repair finishes and the desktop appears. Great. That session might be a genuine all-clear after a power cut, or it might be a dying drive's last good afternoon. Treat it like the second one: copy your irreplaceable files somewhere else immediately. Photos, documents, tax records, the desktop folder everyone forgets. External drive, cloud, another computer, doesn't matter. Then check the drive's health (SMART data will usually show reallocated or pending sectors if this was hardware) and decide about replacement from a position of safety. What you shouldn't do is exhale and change nothing; this screen re-appears on failing drives, and the second showing is usually worse than the first.

Frequently Asked Questions

How long should "Repairing disk errors" take?

On a healthy SSD, minutes to an hour or so. On a large or slow hard drive with real filesystem damage, several hours can be legitimate, once. Past 8 to 12 hours, or into a second day, the odds that it's productively working (rather than endlessly retrying unreadable sectors) drop sharply, and a hard drive making seek noises the whole time is telling you where the problem really is.

Can I interrupt Windows while it's repairing disk errors?

You can, by holding the power button. On a healthy drive interrupted mid-repair, Windows just re-runs it next boot; the process is designed to be interruptible. The bigger risk isn't the interruption, it's what continued repair does to a failing drive. If you're pulling power because of noise, hours stuck, or a repeat loop, don't boot from that drive again; that's the moment to change strategy, not retry.

Does "repairing disk errors" mean my hard drive is failing?

Not always, but often enough to respect. A one-time appearance after a power cut on a drive that then behaves normally is usually just cleanup. Repeat appearances, appearances with no trigger, mechanical noise, slow file access, or SMART warnings afterward all point at the drive itself. Filesystems rarely corrupt themselves on healthy hardware without a reason.

Will the repair delete my files?

It can damage them, yes. Chkdsk-style repairs reconcile the filesystem's records, and when the underlying reads are unreliable, that reconciliation can truncate files, strand them as orphaned fragments (the infamous found.000 folders full of .chk files), or "fix" directories into emptiness. Your data's safest state is untouched on a powered-off drive, imaged by hardware that doesn't write anything back.

The repair finished but my computer is slow and files are missing. Now what?

Stop putting new data on the machine; missing-after-repair files are often recoverable if their space isn't reused. Copy what survived to another drive, then run recovery against the original (ideally against an image of it). If the drive also shows bad-sector symptoms, skip DIY scanning and have it imaged professionally. And going forward, that drive has disqualified itself from holding the only copy of anything.

Lost your data?
Let's assess your options.

Estimated turnaround: 4–6 weeks; timing may vary by case. Start with a free initial quote. The $100 lab diagnostic is credited on success; diagnostic and approved parts charges apply on failure.

Split the cost of your recovery with Affirm. Explore financing Subject to eligibility.

Devices recovered at our lab

These rows come from BenchKey, the ticket system we built for our own shop. Customer names, contact details and descriptions are left out.

Opened laptop hard drive with platter and heads visible at our Lincroft lab
An opened laptop hard drive shows its platter and heads at our Lincroft lab.
4.6 · 345 Google reviews$100 diagnostic, credited toward a successful recovery
Start your mail-in recovery