Shulou(Shulou.com)06/01 Report--

Fault description:

5 pieces of 2T hard disk to build RAID5, divide LUN for windows server to use. The Sql Server2008 database is installed in the windows server. There are three logical partitions in the storage space, with sizes of 500G, 800G and 2.3T, respectively. Database files are missing, mainly involving five databases, and the number of tables is about 6000. The cause of the loss is unknown and the location of the data storage cannot be determined. The sizes of the three databases are 8G, 15G and 20G, respectively. The server is still powered on after the file is lost, but a large amount of data is not written.

Initial inspection process:

1. Use RAID information and internal data block information to reorganize RAID.

Recombinant RAID

2. Extract three partition images in LUN.

3. Scan the missing files in the file system, and no deleted database files are found.

4. The result of the initial check is that the database file is missing and cannot be recovered from the file system point of view.

Recovery process:

1. Make a recovery plan. After the database file is deleted and determined that the file cannot be recovered, it can only be restored by scanning the data page and extracting the records within the page.

2. Use the North Asia self-compiled data page scanning program to scan and extract the data pages in the partition. After scanning the two partition images respectively, it is found that the number of data pages in 500G system disk is very small and the data pages are seriously broken, and the number of data pages scanned in the other partition is more. Tentatively, this partition is the database file storage space.

Scan data page

3. Reorganize the system table. Sql Server database uses system tables to manage all user tables, in which the number of columns, data types and constraint information of each table are recorded. In the process of parsing the system table, it was found that the system table in the extracted data page was damaged and could not read the information normally. After communicating with the customer, we know that there are backup files, and there is no large number of changes to the table structure after the backup is completed, and the system table is available.

4. Restore the backup.

Restore backup

5. Extract the structure information of each table from the three libraries.

Extract table structure information

6. Parse the table structure script. The column information of each table is stored in the database for subsequent use.

Scan script file

Table structure information is stored in the database

7. Parse the system table to get the id information of user table, the structure of associated table and data page. (in order to protect customer privacy, the following steps involve no screenshots of the table name of the user table and the data part of the data page)

8. Create a new database, use North Asia to independently write parsing records and import them into the recovery environment.

9. Sort out the recovery results. There are several backup files in this partition in addition to the database files, so there may be duplicate data after the record is exported and must be de-duplicated. Write SQL stored procedures to remove duplicates.

Database deduplication

10. Customer verification data. After checking the data, the customer indicates that the data is acceptable, transfers the data to the customer storage device, and recovers it successfully.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.