RAID refers to a system of several hard drives that are needed for sharing or copying of data. The data is spread across the drives through one of the many ways called ‘The RAID levels’, depending on the mode of performance required. RAID 5 uses a “block-level striping” where the parity data is distributed across various disks. A very low redundancy cost has made RAID 5 immensely popular. In order to achieve the given number of capacity, comparison of the drives is done. If a disk numbered 0 has 4 blocks of information, A1, B1, C1,Dp and a disk numbered 1 has A2, B2, Cp, D1 then disk0 would service the read-request for block type A1 but block B1 cannot be serviced till A1 has been completed, but disk 1 can read the request for “B2” and service it at the same time.
Performance and Parity handling of RAID 5
In parity handling, stripe is a name given to a continuous strand of blocks, one on each disk in the array. If one more block or a part of it gets written on the same stripe then the whole or a portion of it gets rewritten. It requires reading the older parity and data block, comparing and modifying it with the new request and then writing the new data or parity block. “Distributed parity block” is the term used when they are shifted from one of the stripes to another. When it comes to performance, it has been seen that RAID 5, when implemented, can suffer from bad performance when overloaded with many writes. Sequences of read- modify – write must be flowed for the parity and data block with the latter being updated on every write.
Hard Drive Data Recovery issues
It has been seen that when there is a failure of system while there are writes which are active; the stripes parity doesn’t remain continuous with the data available. If this is not found out and taken care of at the right time before the failing of a block, loss of data may occur. This happens as wrong parity is used to make up for that unavailable (missing) block in that particular stripe. This fragile nature is sometimes known as the ‘write hole’. In order to reduce the occurrence of this to happen or to prevent it, cache which has been given a battery backup or a same kind of technique is often used. RAID 6 too has faced this issue.
The capacity of a single drive is used by parity in a single array, the number of hard disks that can be contained are limited by the storage manager’s capacity. If the numbers of disks in an array increase, the chances of losing data due to multiple drive failure also increase with it. The ratio between the total drives and the drives that can fail before data is lost is reduced and this becomes the main reason for it.
Finally, it can be said that RAID drives faciliate easy hard drive data recovery by design and this makes them a very suitable choice for storing important high-volume data.