Case:
The IBM 3650 server, the bottom layer uses 5 SAS 500GB disk RAID5, one of which is a heat -reserve disk.The server operating system is Windows 2003 X64. The entire storage is divided into three partitions. The first 40GB is the Windows system area, the second is about 700GB, which is the partition where the DB2 production library is located.Division.Customers use the IBM DB2V8.2 database. The overall data is not large. The DB2MOVE backup file is about 16GB.However, multiple tables in the database store external file pointers, pointing to more than 1 million scanned parts.
Solution:
1. The engineer began to mirror all the disks. It was found that two disks had serious physical bad ducts. Maximize the data through the maximum mirror of the dedicated equipment, and then set the RAID analysis and reorganization process.2. Analysis found that the heat -preserved disk has already topped the RAID disk, and then analyzes the first two disks to the first offline. At this time, the three disks required to restore the data are determined, but there is a large number of bad cavities in the front of a disk.In the Windows system area.3. According to the analysis of the original raid parameters, reorganize RAID and repair the Windows system.Use the normal Windows system file to replace the Windows system file in the damaged RAID, and then the mirror back to the customer to use the new disk to make a new RAID, start the server, and enter the system, but one is normal.4. Verify DB2 database: The database cannot be started. It is found that the Windows service has problems. The repair is well repaired. The database cannot be started. The database instance can be restored normally to activate the two databases under the instance, which is normal.5. Customers verify the data in the table in the database. The latest data exists and the operation is normal.
