The news was released that the people behind the development of HDDSuperClone, a utility for copying information from faulty hard drives, they made the decision to release the source code of the utility.
Lack of time to maintain the project is cited as the main reason why the source code is released. The code is open under the GPLv2 license (the license is mentioned inside the code files).
For those who are not familiar with HDDSuperClone, you should know that it allows you to extract the saved data from the disk in case of read errors or unstable operation of individual magnetic heads.
Initially, the project was supplied andn binary builds under the EULA agreement and was developed as a advanced alternative to the GNU ddrescue utility, extending its capabilities previously only available on hardware systems for cloning failed drives, such as DeepSpar DDI4 and RapidSpar.
HDDSuperClone supports import and export of ddrescue log files, allowing you to switch between the two utilities as needed.
The free version uses SCSI pass-through commands to communicate with the device. This offers more feedback than traditional communication methods, and it may be possible to detect when a unit is no longer responding correctly. Ddrescue can have trouble knowing when a drive is no longer responding, as the OS just reports a read error with no way of knowing what's wrong.
The PRO version has the ability to use direct I/O for IDE and SATA attached drives, along with some other advanced features.
Another feature of HDDSuperClone is that It has an advanced self-learning head skip algorithm that will attempt to skip a bad head in about 7 bad reads. This can make getting data from good heads much faster. Hardware imagers can "turn off" a head and only read data from good heads, which is ideal, but requires the use of special vendor-specific commands. HDDSuperClone's head-jumping algorithm is as close as you can get without special commands.
About the released code, it is mentioned that it also includes the paid Pro version, which includes the following features:
- “Direct AHCI” and “Direct USB” passthrough modes, which allow adaptive resetting or rebooting of unstable drives in the event of a freeze.
- The «Virtual Controller» mode, which allows you to copy not the entire disk, but only those sectors in which the necessary data is located. In this mode, the data source and destination are connected, and a virtual disk (/dev/sdv by default) is created and must be read by third-party software to rebuild file systems (for example, TestDisk, DMDE, or R-Study). In this case, all sectors, the reading of which is requested by the software, will be read from the source disk only once and immediately written to the destination disk or image.
- The ability to access a drive using a virtual disk that has a different logical sector size, which can be useful for returning the original logical sector size on a disk that is accessed through the SATA interface, but was originally It was used with a USB interface.
- Support for cloning and imaging of SAS (SCSI) drives that use 520 bytes per sector and saving the result to an image or other drive that uses 512 bytes per sector.
- The function of analyzing the state of the disk.
- Support for sector sizes other than 512 bytes and the ability to adjust the block size to optimize for bad sectors on disks that use 4096 byte sectors.
- Fast Skip Mode, which uses a more aggressive algorithm to prevent disk corruption and bypass faulty magnetic heads.
- Support for using the data returned by the ATA interface to mark bad sectors.
For interested in being able to know the source code, you must know that it is published in a tar file, but the enthusiasts have already created a repository on GitHub, in which they plan to continue the development of the project.