[Data recovery failure description]
An HP DL380 G4 server that uses the HP Smart Array controller to mount a homemade disk array consisting of a set of 14 146G SCSI hard drives in a group of RAID5.
Operating system for Linux, built Nfs+ftp, as the company's internal file server use.
Because the room moved, the administrator by the way to clean the machine, to the new room after the line, unable to identify raid, did not initialize.
[Data Recovery analysis]
The HP Smart Array controller originates from Compaq, and the redundancy in RAID uses a dual-loop checksum.
This example is actually caused by a loss of RAID information.
[Data recovery process]
1. Connect the SCSI hard disk cabinet directly to the SCSI expansion card without the RAID function.
2, in the dedicated (windows2003 modified) repair platform on a single disk to connect all the hard drives.
3, all hard disk read-only way to do a full mirror, while mirroring is also stored on the device with redundant protection.
4, from the mirror analysis of the original RAID double cycle check parameters.
5, in the virtual RAID platform to remove early offline disk, interpretation of the file system, can be exported data.
6, in the customer's original HP server connection disk array, reconfigure RAID.
7, through the network DD, NFS, SAMBA, FTP, SSH and other ways to pass data back to the new good disk array.
[Data Recovery Results]
All work lasted 2 days and 100% data was restored successfully.