RAID disk array is a simple explanation, it is a combination of multiple hard disks through RAID Controller (sub Hardware, Software) into a virtual single large-capacity hard disk to use, which features multiple hard disks at the same time to speed up the read speed and provide fault Tolerant Fault Tolerant, so RAID is used as the main access to data storage not Backup Solution in the usual way. There is a basic concept in RAID disk arrays called EDAP (Extended Data Availability and Protection), which emphasizes expandability and fault-tolerant mechanisms, and it is also the main focus of various vendors such as Mylex, IBM, HP, Compaq, Adaptec, Infortrend, and other companies. Adaptec, Infortrend, etc., including the following actions can be handled without downtime: RAID disk arrays support automatic detection of hard disk failure. RAID disk array support to rebuild the data of bad hard disk tracks. RAID disk array supports Hot Spare without downtime. RAID disk array supports Hot Swap without downtime. RAID disk array supports expanding the hard disk capacity and so on. RAID Level NRAID: Hard Disk Continuous Usage NRAID means no RAID function. It uses the total capacity of the hard disks to form logical disks (no stripe read/write). In other words, the logical disk capacity it generates is the sum of the physical disk capacity. In addition, NRAID does not provide data redundancy. JBOD: JBOD means that the controller treats each hard disk in the machine as a separate hard disk, so that each hard disk is used as a single independent logical disk. In addition, JBOD does not provide data redundancy. RAID 0: RAID 0 - Disk Stripping without parity (commonly used) is also known as data chunking, where the data is divided into a number of equal sized chunks and written to different hard disks on the array, also known as "Stripping". This technique is also known as "Stripping" (i.e., striping the data), which distributes the data on multiple disks and operates the disks in parallel when reading and writing. Theoretically, its capacity and data transfer rate is a single hard disk N times. N is the total number of hard disks constituting RAID0. N is the total number of hard disks in a RAID0. Of course, if the array controller has multiple hard disk channels, RAID0 operation of hard disks on multiple channels will result in higher I/O performance. Therefore, it is commonly used in the field of graphics, video, etc. RAID0 I/O transfer rate is higher, but the mean time to failure (MTTF) is only one Nth of that of a single disk, so the reliability of RAID0 is the worst. RAID1: RAID 1 - Disk Mirroring (more commonly used) Also known as mirroring. That is, each working disk has a mirror disk, each time you write data must be written to the mirror disk at the same time, read data only from the working disk read out, once the working disk failure immediately transferred to the mirror disk, read data from the mirror disk. When replacing the faulty disk, the data can be reconstructed to restore the correct data on the working disk, this array has high reliability, but its effective capacity is reduced to less than half of the total capacity, so RAID1 is commonly used in applications with very strict fault tolerance requirements, such as finance, finance and other fields. RAID (0+1): Combines RAID 0 and RAID 1 - striping reads and writes while using mirroring. RAID (0+1) allows for multiple hard disk failures because it uses the hard disks exclusively for data redundancy. RAID (0+1) is automatically implemented if there are more than two hard disks in RAID 1. RAID2: Also known as bit-crossover, it uses Hamming code for disk error checksums and bit-crossover access, and is used for reading and writing large data, but the redundancy information overhead is too high (more than one parity disk), and has been eliminated. RAID3: RAID 3 - Parallel Disk Array is a single disk for fault-tolerant parallel transmission. That is, using Stripping technology to divide the data into chunks, these chunks are heterodyne checksums, checksum data is written to the last hard disk. It is characterized by a disk for the checksum disk, data in the form of bits or bytes stored in each disk (scattered records in the group of the same sector on each hard disk). When a hard disk fails, the write operation continues to operate on both the data disk and the checksum disk, except for the failed disk. Read operations are performed by reconstructing the data that should be on the failed disk through heterodyne calculations on the remaining data and parity disks.The advantages of RAID3 are parallel I/O transfers and single-disk fault tolerance, which provides a high degree of reliability. Disadvantages: each read and write to involve the entire group, each time only one I/O. RAID4: similar to RAID3, the difference is: RAID3 is by bit or byte cross-access, while RAID4 is by block (sector) access, you can individually operate on a disk, do not have to be like RAID3, even if each small I / O operation involves the entire group, only involves the group of two hard disk (a data disk, a check disk) can be, thus improving the small amount of data I / O speed. Disadvantage: For randomly dispersed small data I/O, the fixed checksum disk becomes an I/O bottleneck, for example: transaction processing. For example, if you do two small write operations, one on stripe1 of drive2 and one on stripe2 of drive3, they both have to write to the parity disk, so there is a contention for the parity disk. RAID5: RAID 5 - Striping with floating parity drive (most commonly used) is a rotating parity independent access to the array, it is different from RAID3, RAID4, there is no fixed parity disk, but according to some rules of the parity information evenly distributed in the array belongs to the hard disk, so in each hard disk, both data information and parity information. Instead, the parity information is distributed evenly across the array's hard disks according to some rule, so that on each disk, there is both data and parity information. This change solves the problem of contention for parity disks, allowing multiple write operations to be performed concurrently within the same group. Therefore, RAID5 is suitable for large data volume operations, but also for a variety of transaction processing, it is a fast, high-capacity and fault-tolerant distribution of disk arrays. When there are N array disks, the user space is N-1 disk capacity. In RAID3 and RAID5, after a hard disk fails, the RAID group changes from ONLINE to DEGRADED mode, but I/O reading and writing are not affected until the failed disk is recovered. However, if a second disk fails in the DEGRADED state, the data of the entire RAID group will be lost.