Application and implementation of RAID technology.
RAID, short for Redundant Arrays of Independent Disks, is a cheap redundant disk array in Chinese. The basic idea is to organically combine several smaller, relatively inexpensive disk drives to make them perform better than an expensive large disk. Increasing the transmission rate and providing fault-tolerant function are the greatest advantages. RAID was originally used in the high-end server market, but with the rapid development of computer technology, RAID technology has penetrated into all areas of computer. Nowadays, RAID control chips can also be seen everywhere in the home computer motherboard. Due to the limited space, this article gives a brief introduction to several common RAID levels.
(1) RAID 0
RAID 0 is a RAID level with no data redundancy, low cost, high read-write performance and high storage space utilization. It is suitable for special applications such as Video, Audio storage and temporary file dump, which require very strict speed requirements. However, due to the lack of data redundancy, its security is greatly reduced, and any disk damage that constitutes the array will bring catastrophic data loss.
In RAID0, all the damaged data of a disk are lost and there is no redundancy function.
(2) RAID 1
RAID 1 uses Disk Mirroring technology, which is a complete mirror of "two pieces" of disk data. It has good security, simple technology and convenient management. But it can not be extended, only two hard disks, available capacity is only a single disk capacity, data space waste is large, so it can be said to be the most uneconomical RAID type, but the data security is extremely demanding, is the preferred RAID type.
Damaged data of a hard disk in RAID1 will not be lost
(3) RAID 10
RAID 10, that is RAID 1 + 0, some RAID controllers can directly select RAID 10 type to complete the creation of RAID 10, while some RAID controllers need to make several RAID 1 disks first, and then make these RAID 1 into RAID 0 to achieve RAID 10 in two steps. There is no difference between the two ways in principle.
In the figure above, four disks are RAID 10, and each RAID 1 allows one disk to go offline.
As shown in the figure above, 1/2 of the hard disks in different groups are allowed to be offline while the RAID group is still working properly. But in practical work, when a hard disk in RAID 10 group appears now "different group" or "the same group", if it only mechanically understands the fault-tolerant number of RAID 10, and fails to repair in time, once the second hard disk in the same group is offline, then RAID0 made between groups is no fault-tolerant mechanism, the consequences will be catastrophic.
(4) RAID 5
RAID 5 is a widely used RAID technology. It divides individual disks into strips first, and then performs parity check (XOR) on the same strips. The "checking data" are evenly distributed on each disk. RAID 5 arrays constructed with N-block disks (requiring N or more than 3) can have the capacity of (N-1) block disks, and the utilization of storage space is very high. RAID 5 allows any hard disk in the group to fail offline. It has the advantages of better data security, higher read and write speed, space utilization and high economy.
RAID5 allows a hard disk to go offline without losing data
As shown in the figure above, after HDD3 hard disk failure is offline, the whole RAID group can run normally, but it must be explained that the operation of replacing and repairing the failed hard disk should be carried out immediately at this time. Do not think that the RAID group can still work and delay the maintenance time. Because once a hard disk is offline in the RAID 5 group, the RAID group can work normally at this time, but it is actually "working with illness", some RAID cards will display the RAID as "Degraded" (downgraded). At this time, every time the data is read and written, a large number of data back-calculation is needed to extract the valid data. Therefore, the normal hard disk in RAID group will have a high I/O reading and writing pressure, so the high-load I/O operation will easily cause the logical or physical damage to the hard disk. If the second hard disk in the same RAID5 group is offline again, the RAID5 group will be damaged. There is a risk of data loss.
(5) RAID 6
The full name of RAID 6 is "Independent Data Disks with Two Independent Distributed parity schemes" (independent data disks with two independent distributed verification schemes). This RAID level is developed on the basis of RAID 5, so it works in the same way as RAID 5. The difference is that RAID 5 writes the check codes into one drive, while RAID 6 writes the check codes into two drives, which enhances the fault tolerance of the disk. At the same time, RAID 6 arrays allow two disks to fail, but the phase is different. The number of arrayed disks should be at least 4.
(6) RAID 50 and RAID 60
RAID 50, or RAID 5 + 0, is similar to the RAID 10 mentioned above. The same principle is to make several RAID 5 groups of disks first, and then these RAID 5 groups together to make RAID 0. RAID 60, that is, RAID 6 and RAID 0, requires at least eight disks.
(7) RAID 1E
RAID 1E (Enhanced) is an enhanced version of RAID 1. The working principle of RAID 1E is basically the same as RAID 1, but RAID 1E has stronger data recovery ability. But because RAID 1E writes a data at least twice, the processor load of RAID 1E will be very high, resulting in the decline of disk read and write ability. RAID 1E requires at least three hard disks to implement, supporting 3 to 16 disks, allowing one hard disk in the group to fail offline. if you want to raid data recovery, please use the data recovery software.
Several types of RAID introduced above are not all, and some types of RAID are not introduced. With the development of RAID technology, new types of RAID will be introduced in the future. The implementation of RAID needs corresponding RAID controller to be able to use, so the requirement of computer motherboard is very high. Generally, the use and popularization of RAID technology is limited in family use. In addition, as far as the role of personal computers in family use is concerned, the use of RIAD technology is not very necessary. At present, RAID technology is mainly used in enterprises, institutions, scientific research, education and other high-end server areas.
There are two ways to realize RAID: one is HostRAID implemented by on-board controller; the other is Hard RAID implemented by extension card. The hardness here is for on-board RAID. Because there are independent RAID control chips on extension card to complete all the operations of RAID creation, reading and writing, it is not like Host implemented by program in on-board hard disk driver chip. RAID, so commonly known as "hard RAID". At present, the mainstream RAID controllers are LSI and Intel. Adaptec RAID controllers are seldom available in the market.