Cause storage environment to deploy and recover data stored faults: a company NetApp FAS-8200 memory, using 96 disk set up two storage pools, storage pool mirror each other. Storage pool volume is divided and mapped to ESXI used as data storage, the number of virtual machines within the volume to about 300+. During operation due to unknown causes loss of volume, the volume inside the virtual machine is not accessible. The company's advanced storage administrator for a simple checking and data recovery but without success, due to the company's important data within the storage administrator not frivolous, but to contact Beijing store data recovery company specializing in data recovery.
First, the data recovery backup
To prevent damage to the customer within the raw data disk, first separately for each copy disk mirroring (determined during the hot standby is not enabled, remove hot spare disks and disk storage systems after analysis) small series of past data recovery cases to be backup introduced, the process will not repeat them here.
Second, the analysis of the underlying data and reorganize disk RAID data recovery
The first step in data analysis Data Recovery i.e. where the underlying array, according to the underlying meta information recording disk, the disk is determined sequence and function (data / parity) where each disk, no off-line while the disc is determined without checking information, excluding parity disk.
Third, the data extraction volume
NetApp file system used for the WAFL, in this case the file system uses a high-version mode. Fill in the configuration file, use a data recovery company developed parser to parse: 
After completion of the extraction of the data by the data recovery engineers extracted data file self test, test file data found during data file exceptions, data recovery engineer for data two secondary analysis found that some of the data block is filled since the pointer exception. Below: 
the process has not appeared in the past, such pointers recovery of data, after the advice of experienced data recovery engineers said business experience in data recovery have not found this type of pointer, and there is no readily available data recovery solutions solve this failure, the data recovery technology officer unconventional business technology research group refer cases to the data recovery center for technology research.
Fourth, data reduction
After data recovery technology research team concluded after analyzing the test is: This pointer is occupied compression flag, and gives decompression algorithms.
Write data decompression program based decompression algorithms to verify the extracted data to decompress. At any time on the part of the abnormal situation in the decompression process of adjustment program, feel free to improve decompression algorithms, and ultimately get the full available decompression program. Proven program is available, after extracting VMDK virtual machine can be resolved properly and export the file. Data recovery engineers extracted from the samples sent to the customer file test, the test result is normal.
Data recovery engineers continue to adjust data extraction program according to this idea, add directory block analysis module and a decompression module, extract all the files in user volume, bulk data recovery operations.
Fifth, data validation; data recovery success
After the completion of all data extraction, data recovery engineers to recover complete data handed over to the store to verify the customer, validated by the customer, this data recovery a success.