How to Recover Data from a Failed RAID Array
Learn how to recover data from failed hardware and software RAID arrays, NAS devices, DAS systems, and server storage with Hetman RAID Recovery. This guide covers the complete recovery process: preparing the original disks and choosing a safe connection method, reconstructing the RAID automatically from metadata, rebuilding the array manually when its configuration is missing or damaged, and analyzing the resulting volume to find, preview, and save recoverable files.
The program can work with directly connected physical disks, disk images, and remote block devices added over SSH. It reconstructs the storage configuration virtually without modifying the original data on the source disks. To begin, download and install Hetman RAID Recovery on a Windows, Linux, or macOS computer that you will use to access and analyze the RAID member disks.
Prepare for RAID Data Recovery
and Connect the Source Disks
Before starting the recovery, stop using the affected RAID, NAS, DAS, or server storage system. Any rebuild, initialization, formatting, file system repair, or other write operation can overwrite RAID metadata or recoverable file data.
Do not rebuild or reconfigure the original array. Do not initialize its disks, create new partitions, format volumes, or run CHKDSK, FSCK, or other repair utilities.
Record the Original RAID Configuration
Before removing or reconnecting any drives, collect as much information as possible about the original storage system:
- Photograph the disk bays and the current order of the drives.
- Label every disk with its bay number or original position.
- Record the model and serial number of each drive.
- Record the model of the NAS, RAID controller, motherboard, or server.
- Write down the RAID level, number of member disks, disk sizes, and stripe size if known.
- Note which disks failed, were replaced, were used as hot spares, or were removed from the array.
- Record any recent rebuild, migration, expansion, disk replacement, or configuration change.
Keep every available disk related to the array, including failed drives, replacement drives, hot spares, and disks removed during an interrupted rebuild. They may contain different versions of RAID metadata or file data that can help reconstruct the storage correctly.
Connect the Disks Directly
Power off the original device before removing its drives unless the hardware documentation explicitly supports safe hot removal. Connect every available member disk to the recovery computer through compatible SATA, NVMe, SAS, SCSI, or Fibre Channel hardware.
The controller or adapter must expose every drive as an independent physical device. Use an HBA, JBOD enclosure, or controller operating in HBA or JBOD mode. Do not combine the source disks into a new RAID array and do not initialize them when prompted by the operating system.
Create and Mount Disk Images
Create complete sector-by-sector images before analysis when a disk is unstable, reports read errors, disconnects unexpectedly, or may fail during repeated scanning. Disk images are also useful when the recovery computer does not have enough ports to connect all RAID members simultaneously.
In this case, connect the disks one at a time, create a complete image of each drive, and then mount all resulting images in Hetman RAID Recovery. The mounted images can be used instead of the original physical disks.
Connect to a NAS or Server over SSH
If the NAS or server still starts and provides SSH access, you may add its physical disks, partitions, or other block devices remotely without removing them from the system. Administrative or root privileges may be required to access the source devices.
Physical disks, mounted images, and remote SSH devices can be used together in the same recovery session. This is useful when only some RAID members can be connected directly.
Prepare Storage for Disk Images and Recovered Files
Use a separate healthy storage device for disk images and recovered files. The destination must not be a disk that belongs or previously belonged to the damaged RAID.
When creating complete disk images, prepare free space approximately equal to the total capacity of the source disks being imaged. Also make sure that enough separate storage is available for the files you expect to recover.
Verify the Connected Sources
Start Hetman RAID Recovery and check that every available physical disk, mounted image, and SSH device appears in Drive Manager with the expected model, serial number, and capacity.
If a disk is missing, has an incorrect capacity, disconnects, or reports read errors, do not proceed with repeated scans. Check the connection or create a complete disk image first. When all available sources are ready, continue to automatic RAID detection and reconstruction.
Detect and Reconstruct the RAID Automatically
After all available physical disks, disk images, and remote SSH devices are added, Hetman RAID Recovery automatically reads the available RAID metadata and compares information found on different sources.
The program uses this metadata to determine the original storage configuration, including the RAID level, number and order of member disks, stripe or block size, start offsets, block order, parity layout, disk groups, missing members, and other controller- or manufacturer-specific parameters.
Wait for Automatic RAID Detection
Check that all available source devices appear in Drive Manager, and allow the program to complete its initial analysis. When sufficient metadata is available, Hetman RAID Recovery automatically reconstructs the array and displays it as a separate logical storage device.
The reconstructed RAID may contain:
- one or more partitions;
- a file system volume;
- an LVM, Storage Spaces, ZFS, Btrfs, or other logical storage layer;
- multiple volumes created on the same storage pool;
- one or more missing member disks represented virtually.
Do not scan the individual RAID member disks when the reconstructed array is available. File system analysis must be performed on the resulting RAID volume or on the logical volume located inside it.
Review the Reconstructed Array
Select the automatically detected RAID in Drive Manager and review its configuration. Compare the reconstructed storage with the information collected before recovery.
Check the following parameters when they are available:
- RAID level and storage layout;
- total number of member disks;
- order and original position of the disks;
- number and position of missing members;
- stripe or block size;
- start offset of each member disk;
- block order and parity layout;
- total reconstructed capacity;
- partitions, volumes, and logical storage layers found inside the array.
A disk marked as missing does not necessarily indicate an error. If the original RAID can operate with one or more failed members, the program adds virtual placeholders for the unavailable disks and reconstructs the remaining data from the available members and parity information.
Verify the Automatic Reconstruction
Before starting a full file system analysis, open the reconstructed array and check whether its structure is consistent with the original storage system.
The automatic reconstruction is likely to be correct when:
- the total capacity is close to the expected RAID capacity;
- the expected partitions and logical volumes are displayed;
- the correct file system is detected;
- the volume starts at a plausible offset;
- the folder structure appears consistent with the original storage;
- known folders and filenames are visible;
- files shown in Preview open correctly.
The reconstruction may be incorrect or incomplete when:
- the reconstructed capacity is significantly different from the expected size;
- the wrong disks are included in the array;
- partitions are missing or have implausible sizes;
- the expected file system is not detected;
- folders contain corrupted or meaningless names;
- known files have incorrect sizes or cannot be opened in Preview;
- data appears correct only at the beginning of the volume and becomes corrupted later.
Decide What to Do Next
If the correct RAID, partitions, and volumes are detected automatically, continue to the final stage and analyze the reconstructed storage to find and recover files.
Proceed to RAID Constructor when:
- the required array is not detected;
- the program displays only an outdated or incomplete configuration;
- the member disks or their order are incorrect;
- the reconstructed capacity does not match the original RAID;
- the expected partitions or file systems are missing;
- files cannot be opened correctly in Preview;
- RAID metadata is missing, damaged, or inconsistent.
RAID Constructor allows the storage configuration to be found automatically by testing possible parameters, selected from manufacturer-specific profiles, or defined manually.
Reconstruct the RAID Manually with RAID Constructor
Manual reconstruction may be required after an interrupted rebuild, disk replacement, storage migration, array expansion, controller failure, accidental reconfiguration, or partial overwriting of RAID metadata. It may also be necessary when disks from several old configurations are connected at the same time or when one or more member disks are unavailable.
RAID Constructor creates a virtual representation of the original array. It does not rebuild the physical RAID, write metadata to its disks, or modify the data stored on the source devices.
Open RAID Constructor and Choose a Configuration Method
RAID Constructor provides three methods for reconstructing the array:
- Automatic Configuration Search — tests possible disk orders and RAID parameters;
- Search by Manufacturer — uses configuration profiles associated with specific NAS devices, controllers, chipsets, or storage platforms;
- Manual Configuration — allows every RAID parameter to be defined directly.
Search for the Configuration Automatically
Use Automatic Configuration Search when you know which disks belonged to the RAID but do not know their exact order, stripe size, parity layout, start offset, or other parameters.
To search for the RAID configuration automatically, select all available member disks and specify the total number of disks originally included in the array, including any missing or failed members.
During the search, the program can test different combinations of:
- member disk order;
- positions of missing disks;
- RAID level and storage layout;
- stripe or block size;
- data start offsets;
- block order;
- parity position and rotation;
- parity delay;
- disk groups used by nested RAID layouts.
Review the configurations found by the program instead of selecting the first result automatically. Several variants may produce a recognizable partition, but only the correct configuration will provide consistent file data across the entire volume.
Search by Manufacturer
Use Search by Manufacturer when the original RAID was created by a known NAS device, RAID controller, motherboard chipset, operating system, or storage platform.
Select the corresponding manufacturer or storage technology and add the original member disks. RAID Constructor will apply known configuration patterns and narrow the search to parameters commonly used by that hardware or software.
This method is useful for arrays created by NAS systems, hardware RAID controllers, motherboard RAID utilities, and other vendor-specific storage implementations.
A manufacturer profile helps identify likely parameters, but it does not guarantee that every device from the selected vendor uses the same configuration. Always verify the resulting array by checking its capacity, partitions, folder structure, and file contents.
Define the RAID Configuration Manually
Use Manual Configuration when the original RAID parameters are known or when automatic search does not produce a correct result.
Specify the original storage layout and add the disks in their expected order. Depending on the RAID type, configure the following parameters:
- RAID level — such as JBOD, linear RAID, RAID 0, RAID 1, RAID 5, RAID 6, RAID 10, or another supported layout;
- Member disks — all physical disks, disk images, and SSH devices belonging to the array;
- Disk order — the original position of every member in the array;
- Missing disks — virtual placeholders for unavailable members;
- Start offset — the position on each disk where RAID data begins;
- Stripe or block size — the amount of consecutive data written to one member before continuing on another;
- Block order — the sequence in which data and parity blocks are distributed between disks;
- Parity layout — the location and rotation pattern of parity blocks;
- Parity delay — the interval used before the parity position changes in layouts that use delayed parity;
- Disk groups — the arrangement of member disks in RAID 10, RAID 50, RAID 60, and other nested configurations;
- Sector size — the logical sector size used by the original storage when it differs from the automatically detected value.
Preserve the original member order whenever it is known. Do not assume that the physical bay order always matches the logical disk order used by the controller or storage system.
Do not include a hot spare as an active RAID member unless it replaced a failed disk and participated in a rebuild. After an interrupted rebuild, both the failed disk and its replacement may contain useful but different portions of the array, so test the relevant configurations separately.
Preview and Verify the Reconstructed Array
Use the preview available in RAID Constructor to inspect the resulting storage before adding it to Drive Manager or starting a file system analysis.
A likely correct configuration should display:
- the expected total RAID capacity;
- a valid partition table or logical storage structure;
- partitions and volumes with the expected file system;
- recognizable folders in the root directory of the reconstructed volume;
The configuration probably requires adjustment when:
- the expected partitions are missing;
- the file system is not recognized;
- no recognizable folders appear in the root directory of the RAID disk;
Return to the configuration and test another disk order, missing disk position, block size, offset, parity layout, or disk grouping when the preview is inconsistent. Do not start a long file system analysis until the reconstructed storage passes these checks.
Add the Reconstructed RAID
When the configuration appears correct, add the reconstructed RAID to Drive Manager. The program will display it as a virtual storage device together with the partitions, volumes, and logical storage layers found inside it.
Do not write the reconstructed configuration back to the original disks and do not create a new physical RAID from the source drives. Continue the recovery by analyzing the resulting virtual RAID volume and saving the required files to separate storage.
Analyze the Reconstructed RAID
and Recover Files
After the RAID has been reconstructed automatically or created with RAID Constructor, analyze the resulting virtual storage device to locate its partitions, file systems, folders, and recoverable files.
Work with the reconstructed RAID, storage pool, or logical volume displayed in Drive Manager. Do not scan the individual member disks unless you specifically need to investigate data that was stored outside the RAID area.
Select the Correct Volume for Analysis
Expand the reconstructed RAID in Drive Manager and examine the storage layers found inside it. Depending on the original system, the array may contain a partition table, one or more file systems, an LVM volume group, Storage Spaces pool, ZFS pool, Btrfs volume, encrypted container, virtual disk, or another logical storage structure.
Select the lowest-level logical volume that previously contained the required files. For example:
- select the NTFS, ReFS, EXT4, XFS, Btrfs, APFS, HFS+, ZFS, or other file system volume when it is displayed directly;
- select the required logical volume inside LVM or Windows Storage Spaces;
- select the appropriate dataset, ZVOL, subvolume, or storage volume when the original system contained several logical storage objects;
- unlock an encrypted volume before analyzing the file system stored inside it;
- select the reconstructed RAID itself when the expected partition or logical volume is missing and must be found during analysis.
Check the volume capacity, file system, partition position, and label before starting the scan. If several similar volumes are displayed, compare them with the original storage configuration and expected folder structure.
Choose the Analysis Method
Right-click the required RAID volume and start the analysis. The program offers different scanning methods depending on the condition of the file system and the cause of data loss.
Fast Scan
Start with Fast Scan when the RAID configuration is correct and the file system remains mostly intact. This method reads the existing file system metadata and can quickly display:
- the original folder structure;
- existing files and folders;
- recently deleted files;
- files removed from the Recycle Bin;
- data lost after minor file system damage.
Review the results and check the required files in Preview. If the expected folders or files are missing, return to Drive Manager and run Full Analysis.
The Ultimate Guide to Recovering Data from RAID 6 on a Non-Operable NAS Qnap TS-439U-RP/SP
Full Analysis
Use Full Analysis when the RAID volume was formatted, its partition was deleted, the file system is damaged, the storage layout changed, or Fast Scan does not find the required data.
Full Analysis searches the entire selected volume for file system records, lost partitions, deleted folders, and file contents. When configuring the scan, select the file systems that may have been used on the original storage. Limiting the search to relevant file systems can reduce unnecessary results.
Content-aware analysis can identify files by their internal signatures even when their names, folder paths, and file system records are no longer available. Such files are grouped by type and may receive generated names because their original metadata cannot always be restored.
For a large RAID, Full Analysis may require significant time because the program must read the entire reconstructed address space and calculate data from the available members and parity. Avoid disconnecting source disks or interrupting access to mounted images and SSH devices while the scan is running.
Work with Encrypted Volumes
If the reconstructed storage contains BitLocker, LUKS, VeraCrypt, FileVault, encrypted APFS, encrypted HFS+, or another supported encrypted volume, unlock it with the correct password, recovery key, key file, or other required credentials.
After the encrypted volume is unlocked, select the file system inside it and run the appropriate analysis. RAID reconstruction restores the original block order but does not bypass encryption, so valid decryption credentials are still required.
Files protected individually with Windows EFS or Linux fscrypt may also require their original certificates, keys, passwords, or user credentials before their contents can be accessed correctly.
Review the Scan Results
After the analysis is complete, browse the reconstructed folder tree and compare it with the original storage structure. Depending on the condition of the file system, the program may display:
- existing files and folders in their original locations;
- deleted files marked accordingly;
- found or reconstructed partitions;
- files grouped in system or recovery folders;
- content-aware results grouped by file type;
- files from snapshots, previous file system states, or older storage structures when available.
Use search and filters to find data by filename, extension, size, creation date, modification date, or deletion status. This is especially useful when the reconstructed RAID contains several volumes or millions of files.
Save Files to Separate Storage
Click Recovery and choose a destination located on a separate healthy storage device. Save the recovered files to another physical disk, external storage device, network share, or other destination that does not belong to the source RAID.
Never save recovered files to the reconstructed RAID or to any disk that belongs or previously belonged to the original array. Writing data to a source disk may overwrite file system metadata, deleted file contents, RAID metadata, or other information that has not yet been recovered.
When configuring the recovery, preserve the original folder structure and filenames when they are available. If the destination file system has limitations on filename length, unsupported characters, maximum file size, sparse files, permissions, or alternate data streams, choose a destination that can store the recovered data correctly.
For large recovery jobs, save the most important data first. Start with files that are difficult or impossible to recreate, such as databases, business documents, projects, virtual machines, photographs, and unique archives.
Quick Start
Watch this video to find out how to recover data from a RAID array with the help of Hetman RAID Recovery.
-
What should I do if the RAID controller reports a foreign configuration?
Do not clear, initialize, import, or rebuild the foreign configuration until the member disks have been imaged or their condition has been evaluated. Depending on the controller, importing a foreign configuration may modify RAID metadata or start operations that write to the disks.
Record the configuration shown by the controller, including the RAID level, disk order, missing members, stripe size, and virtual disk capacity. Then shut down the system and analyze the disks without modifying them.
-
Can I recover data from only one disk of a RAID 1 mirror?
In many standard RAID 1 configurations, each member contains a complete copy of the stored data, so one healthy and up-to-date disk may be sufficient. The disk can be analyzed separately when its partitions and file system are accessible.
This does not apply to every mirrored storage system. RAID 1E, nested RAID, vendor-specific layouts, partially completed rebuilds, encryption, or additional storage layers may require reconstruction with all available members. When the mirror disks contain different versions of the data, analyze them separately and compare the results.
-
Can data be recovered from an SSD-based RAID after TRIM or discard?
The RAID configuration may still be reconstructed, but files stored in blocks that were already discarded and physically erased by the SSD may no longer be recoverable. RAID parity cannot restore data when the corresponding changes were intentionally propagated across the array as valid discard operations.
Stop using the storage immediately, avoid mounting its volumes in read-write mode, and create images of the available SSDs before additional background cleanup or garbage collection occurs.
-
Can Hetman RAID Recovery repair files that remain corrupted after recovery?
The program reconstructs the RAID layout, analyzes the file system, and copies all readable file data. It cannot recreate sectors that were overwritten, discarded, or lost beyond the redundancy capabilities of the original RAID.
If a recovered file contains all required data but its internal structure is damaged, it may require a specialized repair tool for that file type. Databases, archives, virtual disks, documents, and multimedia files should be validated after recovery with their native applications or integrity-checking utilities.
-
Can I recover data after a NAS factory reset, firmware reinstall, or storage pool recreation?
Recovery may still be possible if the reset or reconfiguration did not overwrite the required file data. A factory reset or newly created storage pool may replace partition tables, RAID metadata, file system structures, or the beginning of the original volumes while leaving other areas intact.
Stop using the NAS immediately, preserve every original and replacement disk, and create complete images before analysis. The program may detect older RAID metadata, or the previous configuration may need to be reconstructed manually. The amount of recoverable data depends on how much information was written after the reset.