raid_recovery
github.com
- Free plan
- Linux
What raid_recovery is
raid_recovery is a free, open-source Linux toolkit for recovering RAID5 data from disk images. It is intended for situations where the RAID superblock is gone or the disk sequence is unknown. The command-line utilities can detect page size, identify images, check parity, determine disk order, and reconstruct data. After page size and image order are established, the restore command writes RAID pages into one file representing the virtual drive. A Python nbdkit plugin can instead expose a virtual drive from RAID5 images without a superblock, when disk order and page size are known. It also supports disks split into multiple image files described by a geometry file. The utilities require numpy and tqdm; virtual-disk operation uses nbdkit, nbd-server, and nbd-client. Detection relies on heuristics whose results are not guaranteed. Reconstruction may need dozens of terabytes beyond the raw images. The project supports EXT4 and uses GPL-3.0.
Who it suits
This suits Linux users recovering RAID5 images, including administrators handling missing superblocks or uncertain disk order. It requires command-line work and substantial storage for reconstruction.
Good to know, and what to watch
- Free and open source
- Recovers RAID5 disk images
- Includes parity checking and reconstruction
- Virtual disk export through nbdkit
- Supports split image files
- Heuristic detection is not guaranteed
- Reconstruction may need dozens of terabytes
- Requires Linux networking tools
- No repository releases are listed
The verdict
raid_recovery provides several paths for RAID5 recovery, from image reconstruction to virtual-disk export. Its heuristic limits and potentially large storage requirement are central planning considerations.
What the maker publishes
- Array identification
- The identification algorithm tests image file combinations against RAID5 parity and can report multiple matches. source
- Dependencies
- The recovery utilities list numpy and tqdm as dependencies. source
- Heuristic limitation
- Page-size detection and disk-order detection use heuristics, and the README says their results are not guaranteed. source
- Heuristic limits
- The README says page size and sequence detection use heuristics, so their results are not guaranteed. source
- Identification limitation
- Disk-image identification requires the array size and page size, and the selected page range should contain data rather than only zeroes for reliable identification. source
- Integration
- The virtual-disk setup uses nbdkit, nbd-server, and nbd-client; the README gives Ubuntu package installation instructions for the latter two. source
- Integrity check
- The parity check examines the full image length and returns a pass or fail result. source
- License
- The repository identifies its license as GPL-3.0. source
- Linux setup
- The README gives Ubuntu installation instructions for nbd-server and nbd-client and describes building nbdkit. source
- Page size detection
- Page size detection uses ASCII file statistics and requires a long ASCII or similar text file on disk; one disk image is sufficient. source
- Page-size detection requirement
- The page-size heuristic needs a long ASCII file on disk, ideally a few megabytes, and examines one disk image. source
- Purpose
- Command-line utilities recover RAID5 data from disk images, including when the RAID superblock has been deleted and disk order is unknown. source
- Reconstruction
- After page size and image order are known, the restore command extracts RAID pages into one output file containing the virtual drive content. source
- Reconstruction output
- Reconstruction extracts RAID pages in sequence into one output file containing the virtual drive content. source
- Recovery steps
- The tools support page-size detection, disk-image identification, parity checking, disk ordering, and data reconstruction. source
- Release availability
- The repository page shows no releases. source
- Resource requirement
- The README says file reconstruction may require dozens of terabytes of free disk space in addition to space occupied by raw disk images. source
- Split images
- The virtual-disk plugin supports disks split into multiple image files, described by a geometry file. source
- Storage requirement
- Reconstruction into an output file may require dozens of terabytes of free disk space in addition to the raw disk images. source
- Virtual disk option
- A Python nbdkit plugin can export a virtual drive from RAID5 disk images without a RAID superblock, provided disk order and page size are known. source
Checked against the maker's pages on 9 Oct 2026.