Bad Sectors

A bad sector is any individual sector on a storage device that can no longer reliably hold or return data. The drive’s firmware detects bad sectors when a read fails the ECC (error correction code) check, when a write fails verification, or when the drive’s self-monitoring routines flag a sector as questionable. Some bad sectors arrive at the factory; most develop during the drive’s working life through wear, physical damage, or interrupted operations.¹

The defining property of a bad sector is unpredictability. The drive’s firmware retries failed reads multiple times (typically 3-5 attempts at the firmware level, plus additional retries from the OS) before declaring a sector bad. Some bad sectors return data successfully half the time and fail the rest; some return correct data after retries; some return wrong data that happens to pass the ECC check but is silently corrupt. The “bad sector” label is a verdict by the drive firmware, not a binary state of the underlying media.

Common causes of bad sectors

The cause determines whether the sector is recoverable and whether more bad sectors are coming:²

• Manufacturing defects. Every drive ships with some bad sectors flagged at the factory. These are mapped to spare sectors via the P-list (Primary defect list) before the drive leaves the production line. Users never see them; the drive presents a clean sector range from LBA 0 to the last LBA. P-list bad sectors are normal and don’t indicate a problem.
• Wear and tear over time. Magnetic media gradually loses signal strength; flash cells gradually fail after many program-erase cycles. As areal density rises, the margin for error shrinks, and sectors that worked yesterday can fail tomorrow. This is the most common cause of bad sectors on aging drives.
• Head crashes (HDDs). The read/write head normally floats nanometers above the platter on a cushion of air. If something disturbs that cushion (sudden shock, dust intrusion, oil contamination), the head can contact the platter and physically scratch the magnetic surface. The result is a sequence of bad sectors corresponding to the affected track.
• Power loss during writes. If power fails while the drive is writing a sector, the sector ends up with partial old data and partial new data. The ECC fails and the sector is flagged. This is a soft bad sector that can be fixed by rewriting.
• Physical shock. Dropping a running HDD can cause head contact and platter damage. Dropping a powered-off drive is much less risky because the heads are parked safely. Even on SSDs, severe shock can damage solder joints and fail the whole drive, though individual cells are usually unaffected.
• Heat. Sustained high temperatures (over 50°C / 122°F for HDDs, over 70°C / 158°F for SSDs) accelerate wear and increase bad sector rates. Drives in poorly ventilated enclosures fail faster than drives in cooled environments.
• Dust contamination. Hard drives are not perfectly sealed; they have breather filters that allow air pressure equalization. Dust intrusion through damaged filters reaches the platter and causes head crashes. Helium-sealed drives (used in many high-capacity enterprise drives) eliminate this risk.
• Magnetic decay. Over decades, the magnetic charges on platters slowly weaken. Drives left unpowered for years can develop bad sectors purely from charge decay, without any physical damage. Less of an issue for active drives.
• Flash cell exhaustion (SSDs). Each flash cell can be written and erased a finite number of times (typically 1,000-3,000 cycles for consumer TLC, 10,000+ for enterprise SLC). Once cells exceed their endurance, they retain charge unreliably and fail their ECC check.
• Firmware bugs. Rare but real: drive firmware bugs can flag healthy sectors as bad, or fail to remap genuinely bad sectors. Manufacturer firmware updates fix specific bugs; if a drive develops bad sectors shortly after a known firmware issue, updating may help.

The CERN study and why bad sectors compound

The most cited research on bad sector behavior is a 2007 CERN study that monitored 1.53 million hard drives across 30 models for 32 months. The study found that 3.5% of drives developed at least one unreadable bad sector during the study period, but the more significant finding was about correlation: drives that developed one bad sector were statistically much more likely to develop more, and the additional bad sectors clustered both spatially (within about 10 MB of an existing bad sector) and temporally (within hours or days of the first one).³

This is the practical mental model for users: a single bad sector is a warning that more are coming, often soon, often nearby. Treating it as an isolated incident is the wrong response; treating it as the first sign of progressive failure is the right one.

The single most important distinction for understanding bad sector behavior. The two types have completely different recoverability and require completely different response strategies.⁴

Hard bad sectors: physical damage

The media itself is damaged. On HDDs this means a section of the platter’s magnetic coating is unreadable, often because of a head crash or oxidation. On SSDs it means flash cells in a particular area no longer hold charge reliably. Hard bad sectors cannot be repaired; the data they held is gone. The drive’s only response is to mark the sector unusable and remap it to a spare sector from the spare pool. SMART attribute 5 (Reallocated_Sector_Ct) tracks how many sectors have been remapped this way.⁵

Soft bad sectors: ECC mismatch on intact media

The sector’s bytes don’t match what the ECC expects, so the drive flags it as bad, but the underlying media is still good. The classic cause: power loss during a write, leaving the sector with partial old data and partial new data. The ECC fails; the read returns an error. But if the OS later writes fresh data to the sector, the write succeeds and the sector returns to normal use. Soft bad sectors can be repaired by writing, which is what tools like CHKDSK /R do at the file system level. Drive firmware also handles this: a sector flagged as “pending” gets removed from the pending list when a successful write happens.

P-list vs G-list

HDD firmware tracks bad sectors in two lists:

• P-list (Primary defect list). Bad sectors flagged at the factory before the drive shipped. Mapped to spare sectors during initial manufacturing. Invisible to users and to most diagnostic tools. The drive ships with a “clean” sector range from LBA 0 to the last LBA, even though some physical sectors are spare-pool entries.
• G-list (Grown defect list). Bad sectors discovered during the drive’s working life. Updated as the drive runs. Each entry consumes one slot in the spare sector pool. SMART attribute 5 reflects the G-list size; once the spare pool is exhausted, the drive cannot remap further bad sectors and starts returning errors to the OS.

The spare pool is finite; on consumer drives, typically 2,000-5,000 spare sectors out of billions. Drives that exhaust their spare pools are at end-of-life regardless of how many “good” sectors remain.

SMART (Self-Monitoring, Analysis, and Reporting Technology) is the standard way drives expose health information. Three attributes specifically track bad sectors, and reading them is the first step of any drive health assessment.⁶

The three critical attributes

• ID 5: Reallocated_Sector_Ct. Number of sectors that have been remapped to spare sectors. Healthy value: 0. Any non-zero value indicates the drive has experienced bad sectors. Rising values over time indicate active degradation.
• ID 197: Current_Pending_Sector. Number of sectors flagged as questionable but not yet remapped. The drive is waiting for a write to either succeed (sector goes back to normal) or fail (sector gets remapped). Healthy value: 0. Non-zero indicates bad sectors waiting to fail.
• ID 198: Offline_Uncorrectable. Number of sectors that failed even with full ECC retry. The data in these sectors is gone. Healthy value: 0. Any non-zero value indicates permanent data loss has occurred.

Two related attributes are worth knowing:

• ID 196: Reallocated_Event_Count. Number of remap events (each event may remap multiple sectors). Useful for distinguishing many small events from one large failure.
• ID 1: Raw_Read_Error_Rate. Aggregate read error rate. Vendor-specific encoding; some drives report it as a number of errors per total reads, others as a normalized score. Read in conjunction with the bad sector attributes.

Real smartctl output from a drive with bad sectors

To check SMART on Linux/macOS, install smartmontools and run:

Reading this drive: 4 sectors have already been remapped (Reallocated_Sector_Ct value 4 in the rightmost RAW_VALUE column), 12 sectors are flagged as pending (Current_Pending_Sector value 12), zero sectors are uncorrectable. The drive is degrading. The remap event count matches the reallocated count (4 events for 4 sectors), suggesting the failures happened separately rather than one large cascade. This drive should be backed up immediately and replaced; the 12 pending sectors will likely fail soon and add to the reallocated count.⁷

Checking SMART on Windows

Windows has no built-in SMART display, but several free tools work well:

• CrystalDiskInfo: the most popular free tool; shows all SMART attributes with color-coded health indicators.
• smartctl for Windows: the same smartmontools tool from Linux, available as Windows binary; same command-line syntax.
• HD Tune: shows SMART plus benchmark and surface scan; commercial with a free version.
• Manufacturer tools: SeaTools (Seagate), WD Data Lifeguard (Western Digital), Samsung Magician (Samsung) all show SMART for their respective brands.

The bad sector concept applies to both HDDs and SSDs, but the failure modes and recovery strategies differ enough that they deserve separate treatment.⁸

HDD bad sector behavior

HDDs fail gracefully. Bad sectors typically appear one or a few at a time, often clustered near each other (the CERN study’s spatial correlation), and the drive remains usable for weeks or months as the failure progresses. Symptoms become noticeable: the drive starts making clicking noises (the “click of death” as the head retries failing reads), specific files take ages to access, and occasional read errors appear in system logs. The window between first bad sector and total failure is usually long enough for a careful user to image the drive and migrate.

HDD bad sector recovery is well-understood: tools like ddrescue on Linux, HDD Raw Copy Tool on Windows, and lab-grade hardware imagers all handle bad-sector-tolerant imaging. The drive’s read retries can be tuned via the SCT Error Recovery Control feature, allowing users to set short timeouts (so reads fail quickly and recovery moves on) instead of the default minutes-long retry sequences.

SSD bad sector behavior

SSDs fail more abruptly. Flash cells degrade gradually, but the drive’s controller hides the degradation by aggressively remapping failing pages to the spare pool. Users see no symptoms until the spare pool is exhausted, at which point the entire drive often goes read-only or fails completely with little warning. Some SSDs handle this gracefully (read-only mode allowing data extraction); others lock up entirely.⁹

SSD recovery is harder for several reasons. The flash translation layer (FTL) maps logical sectors to physical pages dynamically, so reading raw flash chips requires knowing the FTL’s state at the moment of failure. TRIM commands tell the drive that certain logical sectors are no longer in use, and the drive may zero out the underlying physical pages, making deleted data unrecoverable. SSD firmware encryption (default on most modern consumer SSDs) means recovering raw flash without the encryption key returns nothing useful. Lab-grade SSD recovery requires specialized hardware and deep knowledge of specific controller chipsets.

SMART attributes on SSDs

SSD SMART attributes are similar but not identical to HDD attributes. The same Reallocated_Sector_Ct (ID 5) applies, but SSDs add SSD-specific attributes:

• ID 169 / 232: Available_Reservd_Space. Percentage of spare pool remaining. Approaches zero as the drive fills its remap pool with failing pages.
• ID 173: Wear_Leveling_Count. Average program-erase cycles across the flash cells. Compare against the drive’s rated endurance to estimate remaining lifespan.
• ID 233: Media_Wearout_Indicator. Vendor-specific wear estimate, usually starting at 100 (new) and decreasing toward 0 (worn out).
• ID 241 / 242: Total_LBAs_Written / Total_LBAs_Read. Total data volume written and read since drive birth. Compare against the rated TBW (Total Bytes Written) endurance specification.

Several tools claim to repair bad sectors. Most are useful in specific scenarios but dangerous in others. Knowing which tool fits which situation, and which to avoid on a failing drive, separates successful recovery from compounded failure.¹⁰

CHKDSK /R on Windows

Microsoft’s built-in disk check tool. The /R switch reads every sector on the drive and marks bad ones in NTFS’s $BadClus metadata file (or FAT32’s bad cluster chain), preventing future writes to those locations. What CHKDSK /R does: file system level marking of damaged regions so the file system avoids them. What CHKDSK /R does NOT do: repair the underlying hardware; recover data from the damaged regions; fix the drive’s spare sector pool.

The risk: CHKDSK /R reads every sector on the drive, which on a failing drive means hours of intensive I/O on hardware that’s already struggling. The stress can accelerate failure: sectors that were borderline pending can become permanently uncorrectable during the scan. If the data on the drive matters, image first with ddrescue, then run CHKDSK /R against the image (if at all), not against the original.

fsck and e2fsck -c on Linux

The Linux equivalent of CHKDSK /R. e2fsck -c /dev/sdX runs badblocks internally to test every sector and marks bad ones in the ext file system’s bad block list. Same trade-off as CHKDSK: useful on healthy drives for proactive sector verification, dangerous on failing drives because it stresses already-degraded hardware. The -cc double-c variant uses non-destructive read-write testing, more thorough but even more stressful.

HDD Regen and similar “magic” tools

Tools like HDD Regenerator claim to “regenerate” bad sectors through specific access patterns. What they actually do is force the drive to remap bad sectors by writing to them; if the write succeeds, the sector returns to normal use, and if it fails the drive remaps it. This is identical to what writing fresh data to a soft bad sector would do; HDD Regen just automates the process.

The hazards are real: forcing remap consumes spare pool entries, which is fine on healthy drives but accelerates exhaustion on failing ones. Lab-grade recovery engineers generally avoid HDD Regen on drives sent in for recovery because it can interfere with subsequent professional recovery attempts by changing the drive’s state in ways that complicate firmware-level imaging.¹¹

ddrescue: the one tool that’s safe

GNU ddrescue is the gold standard for bad-sector-tolerant imaging. Unlike CHKDSK or HDD Regen, ddrescue’s only goal is to copy what’s readable from the source to a destination, without trying to repair anything. It reads sectors in large chunks, marks failures in a log file, and skips back later for retries on failed sectors. The result is a complete image of everything readable, with a precise log of which sectors failed.

The standard ddrescue workflow:

The first pass gets the easy 95-99% of data quickly without stressing the drive on hard sectors. The second and third passes spend more time on the hard sectors. The drive is read-only during ddrescue; nothing is written back, so there’s no risk of making the drive worse. The image file can then be processed with any recovery tool.¹²

Warning signs of bad sectors

Bad sectors produce distinctive symptoms before they cause data loss. Watch for these in combination:

• Specific files take much longer to access than they used to. Read retries on bad sectors block until the firmware times out.
• “I/O error” messages in dmesg on Linux or in Windows Event Viewer (Disk source).
• Click of death sounds from HDDs. Repeated clicking as the head retries failed reads.
• SMART warnings from the BIOS or OS. Ignore at your peril.
• Files reporting wrong sizes or returning errors when opened. Often metadata stored in bad sectors.
• Slow boots on system drives. The OS retries reads as it loads boot files.
• Random freezes during disk-intensive work. Kernel I/O stalls waiting for read retries.
• Drives appearing as RAW in Windows Disk Management. Often partition metadata damage caused by bad sectors at LBA 0.
• System hangs at the BIOS POST stage. BIOS may retry reading drive identification sectors.

👥 Researched & Reviewed By

Marcus Whitfield

Data Recovery Software Analyst & Senior Writer

All Articles X (Twitter)

Marcus has evaluated data recovery tools for more than six years across Windows, macOS, and Linux, from free utilities to enterprise-grade platforms. He writes the technical reference content on Data Recovery Fix, with particular interest in SMART attribute interpretation and the practical realities of bad-sector-tolerant imaging across modern drive generations.

B.Sc. Computer Science 6+ years data recovery evaluation SMART diagnostics

Rachel Dawson

Technical Approver · Data Recovery Engineer

All Articles X (Twitter)

Rachel brings over twelve years of cleanroom data recovery experience, including hands-on bad-sector remediation via ddrescue and DeepSpar Disk Imager, head crash recovery on HDDs, and SSD spare pool exhaustion scenarios. She validates terminology and ensures published reference content reflects actual lab outcomes.

12+ years data recovery engineering PC-3000 certified Bad sector specialist