Storage of RAID information loss data recovery solution in Linux RAID6

Data Recovery Failure Description:

The original storage is a 12 2T hard disk composed of Linux RAID6, the file system is EXT3, this storage is zoned 3 LUNs, each 6TB size, one day after the raid failure, maintenance personnel in order to salvage the data, the failure of the storage to allocate raid, and initialized.
After a long period of initialization, the maintainer detects that the situation is different and forces the initialization to stop, but the initialization has reached more than 50%. The data section has been irreversibly destroyed.
Data Recovery Failure Analysis:
The cause of the failure is only the raid failure, the maintenance personnel subsequent rescue data process with 11 hard disk to redistribute RAID5, and for a long time initialization, which is irreversible damage to the original data, after the proof, only the third LUN can be used to recover data from the normal RAID6 method, But the third LUN does not have the important data that the customer wants, and the important data is focused on the first LUN.
Due to the complexity of this case, our company received the repair of the customer has been in the domestic data recovery company changed hands several times, including several well-known data recovery company, still unresolved.
Data recovery process:
The recovery process is divided into 4 steps:
1. Analyze the RAID and disk organization of the original 12-block disk RAID6.
2. Analyze the organizational structure of RAID and disks when redistributing RAID5.
3. Algorithms that determine recoverability and how to implement a recovery program.
4. Recovery and repair.
The structure of the original RAID6 is quickly analyzed, but it is difficult to analyze the structure of the RAID6 and RAID5 because of the large amount of information coincident with the underlying, which takes 1 days.
The first and second steps have been completed, after analysis, the initialized data can be destroyed by other methods to restore, to develop a recovery algorithm, spend a day to write programs and program algorithm correction, the program to 12 disk in the original data of the first and second LUN mirrored to two 7TB of storage.
Verified that the second LUN data is completely normal, but the most important first LUN is about 10MB data destruction before, this 10MB data is very deadly, EXT3 root directory and the first block Group I node all in this former 10MB inside, then use data recovery commonly used software UFS Explorer and The recovery effect of R-studio is quite unsatisfactory, and may be the cause of large storage.
In this case, we have to repair the corrupted EXT3 file system, write a program to do the EXT3 orphan directory lookup, found the root directory has 3 directories, the root directory and I node, the file system resolver to open is completely normal, but in order to ensure some of the original data permissions and properties, In linux simple fix, Linux has been able to mount properly, and then the Linux file with the CP command to copy the formatted EXT3 of a single disk partition. When the customer uses the data, no other settings are needed, and the file directory structure and properties are exactly the same as the original CP.
Figure One:

Figure II:

Might

Data Recovery conclusion:
Spents 3 days, data 100% is successfully restored.

Storage of RAID information loss data recovery solution in Linux RAID6