SQL Server database recovery case analysis

The fault environment is 4 Servers , Every server 12 The disk is divided into 2 Group raid, common 8 Group raid. According to the customer's description, a total of 4 Nodes , One of the nodes is still in use after failure , After the second section failure , A series of re launch operations have been carried out , The management storage software cannot be used . To prevent unrecoverable changes to the original disk due to some operations during data recovery , This leads to secondary data loss , Mirror the original disk . North Asia engineers conduct a detailed analysis , Get 5 Bottom image of all hard disks on node servers . Through the analysis of , It is found that the underlying index bitmap is damaged . Analyze all image files , Reorganize according to the underlying data raid, And extract each group raid Medium map, Data pair map Analyze , Manually index the data according to the bitmap , Exclude some corrupt bitmaps . The main data of the customer is SQL server database , After preliminary testing , The index bitmap is partially corrupted , Therefore, if the data is damaged after extracting the data volume , The database can be repaired . 【 Data recovery process 】 1. restructuring RAID The engineer is right about RAID Stripe size 、 Disk sequence 、 After analyzing the key information of the verification direction , Judge the offline order of the member disk . Ten groups were divided RAID Restructuring , And generate RAID Image file .

2.  Data volume MAP check stay RAID To query and extract data volumes MAP Area . As shown in the figure ：

View according to the index bitmap structure and index to the data area , check MAP Completeness and correctness of information , Exclude error indexing bitmap . Programming , according to MAP structure , Data pair MAP Analyze and organize , Sort out all key bitmap information . 3. Code writing and data volume generation First of all, according to the Lefthand Storage correlation structure algorithm , Write the corresponding data volume extraction program . Then use the written program to extract all data volumes .

Cause data MAP Partially damaged , According to the file system structure 、 File structure and other information to repair the data volume and verify the integrity and correctness of the extracted data volume . 4. Data volume analysis and repair Engineers analyze file systems in data volumes , And according to the structure of the file system , Verify the integrity and correctness of the file system . Parse the file system in the data volume , Get all the files , After viewing the file integrity and time information, extract the data in the data volume . 5. Database repair Data file extraction completed , Try attaching , Because the index bitmap is partially corrupted , Lead to SQL server The database cannot be attached directly , according to SQL server The data structure and data record storage rules perform a series of repair operations on the database . After repair , Install the same version of the database , Insert the repaired data into the new library .

6. data validation Customer led validation of data , After verification , Data is normally available . The recovery rate is 100%. about LeftHand Storage , We have mature recovery experience and recovery technology , In case of storage failure , Try not to do anything else , To avoid irreversible structural damage , Please restore professional data as soon as possible .