Case:
The server is IBM3850. Use 5 146GB SAS hard disks to build a RAID5 storage architecture. One of them is because of physical faults. Only four disks have been working on the line for a long time. At this time, RAID5 does not have an error redundant function.When the second plate was faulty, the array collapsed.
Solution:
The engineer cleared the earliest offline disk from five disks.Back up the remaining 4 disks and skip the bad sector that cannot be read.Use these four disk mirrors to build virtual RAID5.For the structural verification of the EXT3 file system for the generated storage, modify and fill the data missing in the position of the bad sector.Export the Oracle database, perform the verification of the database itself, and confirm that there is no error.Use a new disk to build a new RAID5 storage in the original environment.Back the repair mirror clone back to IBM3850.