Raid Data Recovery

Raid disks are one of the marvel of hard disk technology. They are used by most of the fortune 500 companies for data storage needs, since they are tolerant to failure at several levels.

Raid disks are produced with an internal software that is capable of checking for errors and recovering of some of the possible errors. In the case of a crash, RAID disks are extremely valuable, because they can retain their data even if some of the disks fail.

The way RAID disks permit this recovery is through the redundancy of data storage. Based on the internal format used for RAID storage, RAID disks can become aware of what kind of failure happened, and what to do to recover from the failure.

RAID disk organization

RAID means Redundant Array of Inexpensive Disks. It means that the data is stored in disks that, in isolation, are pretty inexpensive. However, there is redundancy on that data, provided by the logical connection between several disks.

The data stored on a RAID disk can be organized according to several levels of redundancy. These levels are numbered from 0 to at least 4 (other more advanced levels or RAID exist).

RAID level 0 happens when data is logically divided between disks. This level is pretty basic, and there is no recovery once a disk is corrupted. The big advantage, however, is that since data is stored in different disks, transfer can be done at higher performance.

RAID level 1 is the simples level with redundancy. In that case, the disks in the RAID array are mirrored. This means that if one disk fails, there is a copy on another disk. RAID 1 has many of the advantages of RAID 0, at just the cost of additional storage.

RAID level 2 uses a simple parity checking system. The method used is a Hamming code. This means that for every bit store in two or more disks, the last disk will store a Hamming code that says if the sum of the other bits is odd or even. In that way, we can recover data if we lose at most one disk.

RAID level 3 is similar to, but data is divided in blocks that are completely paired together, so the disks have to rotate at the same speed.

RAID level 4 is also like RAID level 2 and 3, but it offers block level parity, which improves the safety of the data.

The Raid Data Recovery Process

Once a RAID array suffer a failure, there is a RAID data recovery process that starts automatically, defined by the disk access firmware.

Based on the data recovery algorithm, the RAID drive will detect what disk was demaged, and start request a new disk. Once a new disk is added to the RAID array, the firmware will imediately start making a copy of the missing data into the new disk.

The Raid data recovery process can take some time, depending on the size of the hard disk and the algorithm used to compute the redundancy data. After a few hours, on average, the RAID drive will have reconstructed the whole disk.

Tags: RAID, disks
Article created on 2011-06-02 08:06:30

Post a comment