Case:
The server is a RAID5 composed of 6 hard disks (ST3300657SS). Two of the hard disk damage causes the server to run normally. The server stores a large number of database and database backup files, as well as some daily work.
Solution:
1. Detects the two damaged fault disks, respectively. After the test, one of the hard disks are not used caused by the firmware problem. The hard disk firmware is repaired through a dedicated firmware repair program. There are a small number of bad disks in the hard disk. The other is the damage to the magnetic head. It opens a hard disk with a magnetic head under the dust -free platform. It is found that the plate has been severely scratched. The plate has become a substance that has become black and has a powder state. The data in this hard disk cannot be obtained. 2. Analyze and view 4 normal hard disks to determine that 4 hard disks are not problems, and mirror processing is done. 3. By analyzing the mirror analysis of 5 hard disks, the specific parameters of RAID5 are obtained, and the data area is preliminarily verified to compare the data area. It is found that the data is normal. The rear partition status is normal and can be opened normally. The database file that the final use of the final use can be hung normally, and there is no error in DBCC checking the database. As a result, the hard disk of damaged firmware is the last offline hard disk. It is damaged because RAID5 allows normal work and data integrity when a hard disk is damaged. Read RAID data to the backup hard disk provided by the user.