Bad Sectors on a Disk: When to Shut Down Your Computer Immediately and What a Service Does (Including CHKDSK)

A bad sector rarely announces itself outright. First, copying a single file hangs once. Then Windows freezes for a few dozen seconds when you open a specific folder. Sometimes a quiet click that wasn’t there before comes from the case. And one morning, the accounting database won’t open because the disk keeps spinning up and down repeatedly.

In a company, that disk holds invoices, project documentation, emails, or the only copy of customer files — and every attempt to “get it working somehow” can damage exactly the data you need. The good news is that a large proportion of cases can be rescued if the disk is taken offline in time and nothing unnecessary is done to it.

The short answer: If the disk is clicking, freezing, or reporting read errors, shut down the computer and stop using the disk. Do not run CHKDSK, any “sector regeneration” tool, or formatting, and do not copy anything onto it. First, a sector-level copy is created on healthy storage, and only then is the recovery performed on that copy. It’s better to avoid running chkdsk /r on a dying disk altogether.

This article concerns traditional platter-based hard drives (HDDs). The explanation of bad sectors, scratched surfaces, heads, and mechanics, as well as procedures like CHKDSK, cloning, and laboratory intervention, applies to HDDs. SSDs have no platters or heads — “bad blocks” there are related to NAND memory wear, controller errors, or firmware. The external symptoms (disappearing disk, read errors, freezing) and first aid (shut down, don’t write anything to the disk) are similar, but the causes and diagnostics differ — it’s the state of the NAND, controller, and firmware that matters, not the mechanics. If you have an SSD, treat it as an SSD data recovery case, not according to the mechanical procedures below.

Bad Sectors Are Not a One-Command Fault

A bad sector is a spot on the disk from which the stored data can no longer be reliably read. On a traditional platter-based hard drive (HDD), this is usually due to physical damage to the magnetic surface, worn read/write heads, or a mechanical problem. The trigger is often an impact, vibration, overheating, or unstable power supply.

A logical bad sector is something different — the data or its checksum is corrupted, but the surface itself might still be intact. However, the difference between a physical and logical fault cannot be determined without diagnostics, and they behave similarly outwardly.

A modern disk continuously replaces bad sectors with reserved ones, so it may appear to be fine for some time. However, as soon as values like Reallocated Sector Count, Current Pending Sector Count, or Offline Uncorrectable start rising in the SMART attributes, remapping hasn’t solved the problem — it has only postponed it.

There is one key point for a business owner: “repairing the disk” and rescuing files are two different goals. When data is important, rescue comes first. The disk, which in most cases ends up being replaced anyway, is dealt with afterwards.

When to Shut Down the Computer Immediately

Shut down the computer or external drive as soon as any of these symptoms appear:

• The disk is clicking, knocking, repeatedly spinning up, or making scraping sounds.
• Windows hangs for dozens of seconds when working with the disk.
• Copying fails with a read error, freezes on a specific file, or the speed drops to zero.
• The disk is sometimes visible in BIOS and Windows, other times it disappears.
• The computer stopped booting, and disk errors were appearing beforehand.
• Folders are disappearing, files cannot be opened, or the system itself prompts you to “format the drive.”
• The disk has been dropped, bumped, or had liquid spilled on it.
• The Current Pending Sector Count or Offline Uncorrectable SMART value is growing between checks.

Do not power on a mechanically damaged disk repeatedly “to test it.” Every additional spin-up can worsen the condition of both the surface and the heads. Even routine browsing of folders in Explorer, an antivirus scan, or a file-by-file backup is risky — these operations stress the disk with random reads and constantly return to damaged areas.

If this involves accounting data, a database, email, or the only copy of customer data, it is safer to turn off the device and have the type of fault diagnosed. Diagnostics roughly start at a few hundred crowns — but even that doesn’t automatically promise that all data can be recovered.

What You Can Safely Check Yourself

If the disk isn’t making unusual sounds, hasn’t been mechanically damaged, and the system is still responding, you can perform a short check without writing to the problematic disk.

First, rule out whether the fault is caused by something other than the disk itself:

• For an external drive, try a different USB cable and a different port.
• For a desktop computer, the problem could be the SATA cable or power connector.
• Verify that the disk is correctly detected in the BIOS or UEFI.
• Read the SMART attributes using a tool like CrystalDiskInfo or via smartctl from the smartmontools suite.
• For a NAS, check the disk status in Synology DSM or QNAP QTS, but do not run an automatic volume repair.

A one-time SMART query is relatively gentle. In contrast, a long surface test means reading the entire disk and is inappropriate for a failing unit — just like a scan that takes several hours with a program that retries the same sector again and again on every read error.

Moreover, SMART is not a definitive verdict. A “Good” status does not mean the disk is healthy — the electronics can report that everything is fine even when the disk freezes, has a head problem, or cannot stably read a certain area.

Can CHKDSK Fix Bad Sectors?

CHKDSK checks the file system (NTFS, FAT, exFAT). It can find inconsistencies in directories, indexes, and metadata. It will not repair a scratched platter surface, a faulty read head, or worn mechanics — it is the wrong tool for that job.

The switch used makes a big difference:

The /r switch reads a large portion or all of the volume’s used space. On a healthy disk, it helps mark unreadable areas so the system avoids using them. But on a disk that is clicking, freezing, or has a growing number of pending sectors, it is precisely this prolonged reading that worsens the damage.

Furthermore, CHKDSK works with the volume structure, not the value of individual files. After the operation, damaged fragments can end up as found “file chunks” in the found.000 folder, and the original directory structure can be scrambled. That is why CHKDSK is not run as a first step on the only copy of important data. The safe order is: diagnostics → sector-level image on healthy storage → file system repair only on the copy.

Why a Dying Disk Is Cloned First

Standard disk cloning — for example, when moving Windows to an SSD without a reinstall — assumes a readable source disk. With a defective disk, the goal is different: to retrieve as many sectors as possible with the least stress.

Rescue cloning typically works like this:

1. Reads easily accessible areas first.
2. Skips defective and slow spots and records them in a map.
3. Returns to the problematic areas later.
4. Limits the number of retry attempts for reading a single spot.
5. Performs file system recovery on the copy, not the original.

Software tools like GNU ddrescue or HDDSuperClone / OpenSuperClone are used for this. However, it’s not just about installing a program — it’s critical to correctly identify the source and destination drives, prevent automatic mounting of the damaged volume, and set the reading order. Swapping the devices can overwrite the only source of data. The destination disk must also have at least the same genuine capacity as the source; relying only on a “1 TB” label isn’t enough, because the exact number of sectors varies between models.

What a Service Does and What You’re Paying For

First, it’s determined whether the problem lies in the cable, electronics, file system, platter surface, or mechanics. Diagnostics should distinguish a disk that can be safely read using software from one that belongs in a specialized lab.

For a readable disk, the service typically creates a sector-level copy onto healthy storage and only then reconstructs partitions, the directory structure, and files on that copy. The result is saved onto a different disk — the repaired original HDD should not be returned to service as business data storage.

If the disk is clicking, not spinning up correctly, or has damaged heads, a standard computer repair shop should not open it. Opening a hard drive outside a cleanroom environment practically means its end — a single dust particle between the head and platter is enough. Such a case belongs with a specialist who has the appropriate equipment and compatible donor parts. The same rules apply to data recovery from RAID arrays and NAS, which also involves calculating the array parameters.

The rough cost is affected by the type of intervention:

• Software recovery from a readable disk: technically the simplest, roughly from the lower thousands of crowns.
• Laboratory intervention (head replacement, working in a cleanroom cabinet, sourcing a donor): starts approximately in the higher thousands and can be significantly more expensive depending on the fault.
• The price is further increased by capacity, BitLocker encryption, the state of the heads, and the number of read errors — not just the number of requested files.

We perform diagnostics before we do anything else, and we tell you the approximate price and realistic chance of success upfront. We never proceed with an intervention without your consent.

How to Recognize a Quality Offer

Before handing your disk over to anyone, ask these specific questions:

• Will a sector-level copy be created first before any recovery attempt is made?
• Will anything be written to the original disk during diagnostics?
• Will I receive a list or preview of the found data before the job is completed?
• Is the diagnostic service paid for, and what exactly does it include?
• Where will the recovered data be stored, and will the original disk be returned?
• What happens if laboratory recovery is necessary?

A warning sign is the automatic launch of CHKDSK, a “sector regeneration” tool, or formatting before a copy is made. A promise of 100% recovery before diagnostics is also suspicious — the result always depends on the physical state of the media and whether the damaged areas hold parts of the needed files.

What to Do After Data Recovery

Do not use a disk with confirmed bad sectors for accounting, backups, or standard operation again. Even sector remapping does not restore its original trustworthiness. Replace it and restore your data only onto verified storage.

For a standalone computer, a new SSD usually makes sense. For a NAS, choose disks designed for continuous operation — such as the WD Red Plus, Seagate IronWolf, or Toshiba N300 series — and verify compatibility with Synology or QNAP. At the same time, remember: RAID is not a backup. It mainly protects availability against a single disk failure, not against deletion, ransomware, or damage to the entire device.

For ongoing operation, set up regular backups according to the 3-2-1 rule. And after every data recovery, verify not just that the backup exists, but crucially, that the files can actually be opened from it.