What is RAID?

Redundant Array of Independent Dinks (RAID) is a technology of combining two or more disk drives into one unit. This disk array improves various parameters of the system– it can enlarge the disk storage capacity or increase the data transfer speed. There are several types of such combinations of disks giving various results respectively. They are called RAID levels.

The basic types are level 0 called RAID 0 or stripping and RAID 1, called mirroring. There are also types of combinations of these two simplest RAIDS into successive levels. RAID 0 is used to improve the speed of data transmission. The data is divided into parts called stripes and each part or stripe sets of data are parallelly transferred to each disk. However, there are disadvantages of this solution. Any disk failure destroys the whole array. The failure of one disk blocks access to data stripes, stripes cannot be retrieved and the whole set of data cannot be reconstructed.

The data in level 0 is not being corrected since there is not any system of error correction when saving. The more disks are connected into such array, the higher the risk of system failure. The disks are not protected against any failure – in case when one disk crashes the data from it cannot be retrieved. RAID 1 (mirroring) is the second option. Here disks are also connected into one unit, but not in order to achieve better transfer speed but to provide much better fault tolerance in comparison with RAID 0.

In mirroring, the data is sent to two or more mirror disks in such a way that one disk becomes a copy of the first.

For this reason, the failure of one disk does not mean the failure of the whole system. RAID 1 structure protects the system from disk errors or failures and it continues to operate as long as at least one drive in the mirrored set is functioning. The disk which fails can be substituted with a new one and the data is copied back to it from the mirror disk.

To find out more check wikipedia site about RAID here.

Posted in Uncategorized | 2 Comments »

2 Comments

  1. Alex says:

    We had this come up in tech support in #centos two weeks ago and spent 6 hours nunrisg a user [tasked to be a sysadmin by his PHB] through several single point of failure’ errors. I meant to write about it still need to, I guess, as there are lots of OTHER ways to hose a network and lose access to dataBest regards Russ

    • Micheline says:

      This entry is great.I decided to go with eight 200GB pelarlal ATA drives. They’re Seagates and they run nice and quiet and cool. With a 3Ware Escalade 7506-8 controller I have plenty of performance for my 1.4 terrabyte RAID-5. 160GB drives might have been cheaper but I couldn’t fit 1+TB in a mid tower with only 160GB per drive.

Leave a Reply