Shulou(Shulou.com)11/24 Report--

Have you ever experienced a computer hard disk failure that led to the loss of all data?

After learning that the data could not be recovered, I wanted to cry without tears. When I thought of the life images and learning materials saved over the years, the hard accumulated experimental data were gone, and it was so painful that I could not breathe.

The loss of personal data has been unacceptable, in case the storage system of the enterprise server collapses, such as the financial server of the bank, the training server of the Internet enterprise, the impact is even more incalculable.

In order to avoid this kind of situation, and to improve the read and write rate of massive data, RAID (Redundant Arrays of Independent Disks, independent redundant disk array) arises at the historic moment.

In order to better understand RAID, let's first understand the following concepts.

01. Hard disk, disk, disk array? Silly can not distinguish between hard disk, as the main storage device of modern computer, large capacity, fast read and write, mainly divided into new solid state hard disk, traditional mechanical hard disk and hybrid hard disk.

Since floppy disks have hard disks, of course there are also floppy disks (that is, floppy disks). Floppy disks were removable storage media used by early personal computers. At that time, only 3.5in floppy disks with 1.4MB capacity were commonly used, and their reading and writing speed was slow and had long since faded out of the market.

Soft and hard disks are collectively called disks, and "magnetic" means the use of magnetic recording technology to store data.

Disk array RAID is a redundant array composed of independent disks, referred to as disk array.

General computers equipped with 1 to 2 hard drives can meet the basic needs, while enterprise servers require extremely high data storage capacity and read and write speed, ordinary hard drives are far from being able to meet. So engineers put multiple independent hard drives into a super-large capacity group to provide higher storage performance, I / O performance and reliability than a single disk.

02, some questions about the first acquaintance of RAID, the first acquaintance of RAID, there may be some questions about the concept, the editor will briefly answer the question first.

It turns out that RAID is a large hard disk superimposed with multiple layers of different Buff?

Strictly speaking, RAID is a multi-hard disk operation mode and a high-performance storage technology. For ease of understanding, RAID can be regarded as a large hard disk with a bonus effect, which is composed of several ordinary hard drives in different forms.

So what's the difference between RAID and simply adding a hard disk?

Increasing the hard disk only increases the storage space, still can only read and write one disk at a time, once the hard disk is damaged, it will lead to data loss. In RAID, data is stored on multiple hard disks, and the bandwidth and speed of reading and writing data are improved by parallel processing. And multiple hard drives work at the same time to increase the average time between failures, even if a hard disk failure, timely replacement will not affect the normal work of the entire array, but also can restore the data in the failed hard disk.

All right! In other words, RAID can greatly improve the data throughput of the hard disk group, and provide fault tolerance to ensure data security.

This wave of summary is very good! Here's how RAID combines hard drives to achieve these functions.

03, RAID-level RAID mainly uses striping, mirroring, and data verification to combine hard drives. Firstly, four kinds of standard RAID are introduced.

RAID 0

Also known as striping or striping, the continuous data (A1 ~ A8) are written to multiple hard disks separately, which can read and write concurrently, make full use of bus bandwidth, and have the highest storage performance. Once a single disk failure will threaten data security, it is only suitable for scenarios that do not seek security but only speed.

Is there a safer plan?

RAID 1

Also known as mirroring, a section of data (A1 ~ A4) is written to the work disk and the mirror disk respectively, which is equivalent to automatic real-time backup, and the data security is the highest. However, the implementation cost is high and the utilization of the hard disk is low, so it is only suitable for scenarios where security is fully guaranteed, such as finance and monitoring servers.

Is there a solution that combines performance and security?

RAID 5

The continuous data (A1 ~ A3, B1 ~ B3, C1 ~ C3, D1 ~ D3) and the corresponding parity information (Ap, Bp, Cp, Dp, Ep) are written into different hard disks respectively. In the case of a single disk failure, the normal disk data and parity information can be used for XOR recovery. RAID 5 takes into account the performance, cost and reliability of storage, and is common in all kinds of private NAS servers for civilian use.

What if both hard drives are damaged at the same time?

RAID 6

On the basis of RAID 5, a set of parity information (Aq, Bq, Cq, Dq, Eq) is added. Two sets of independent parity information can achieve double parity and recover data even if two hard drives are damaged at the same time. RAID 6 has higher read performance and fault tolerance, but low write performance and high implementation cost.

By adding the standard RAID, the following three kinds of mixed RAID with balanced performance and reliability can be obtained.

RAID 1+RAID 0=RAID 10

RAID 10 can be understood as using four hard drives (Disk 0 and Disk 1) disk 2 and Disk 3 to form two RAID 1, and then two RAID 1 to form a RAID 0. RAID 10 has both advantages, but the hard disk utilization is not high.

RAID 5+RAID 0=RAID 50

In RAID 50, RAID 0 raid 50 composed of multiple RAID 5 volumes can bear the risk of single disk failure in multiple volumes at the same time, which greatly improves the fault tolerance and inherits the high storage performance of RAID 0.

RAID 6+RAID 0=RAID 60

The overall structure and characteristics of RAID 60 are similar to those of RAID 50, and it also inherits the advantages of RAID 6. Even if double disk failures occur in multiple RAID 6 volumes at the same time, the security of the data will not be affected.

All levels of RAID have their own emphasis on reading and writing performance and security and reliability. The editor makes the following summary:

As can be seen from the table, RAID is characterized by larger capacity, higher I / 0 performance, and stronger data security and reliability.

How can I use RAID? At this point, I'm sure you already know the advantages of RAID, but you may still be confused about how to use RAID. In fact, there are three ways to implement RAID:

The software RAID is integrated into the server operating system, which has the advantages of simple configuration management and low cost. All tasks are performed by CPU, which leads to heavy load and poor overall performance of CPU.

The hardware RAID has a computing unit and an I / O processor, and manages the RAID configuration independently of the operating system. Such as external disk array cabinets, internal disk array cards. Although the cost is higher, but the use and configuration is more flexible, can provide better performance, give full play to the characteristics of RAID, so the application scale is larger.

The software-hardware hybrid RAID has RAID controllers and drivers, but lacks I / O processors, so it still needs drivers to perform tasks.

The external structure of the RAID card in the form of PCIe is shown in the following figure.

RAID card not only combines RAID Cache to improve data reading, writing and transmission performance, but also provides some powerful functions, such as:

■ hard disk pass-through: user instructions are directly transmitted to the hard disk.

■ hard drive energy saving: allows the hard drive to stop when it is idle.

■ data power-off protection: provide supercapacitor for power supply and power calibration.

■ hot backup: provide a hot spare to automatically replace the failed disk, and automatically use the free disk to replace the failed disk when the hot spare is not set.

■ consistency check: the data is calculated periodically and compared with the corresponding redundant data. If the data is inconsistent, the error information will be automatically repaired and saved.

Summary: with the gradual development of hard disk interface and processor technology, RAID technology is becoming more and more mature. With its advantages of large capacity, high performance, reliable and controllable, it is widely used in large and medium-sized databases, high-performance computing, Internet applications and other fields to meet the needs of enterprise users for data storage capacity, I / O performance, data security, business continuity and so on.

At present, many technologies have emerged about RAID, such as load balancing, fault detection, isolation recovery and so on. It is expected that revolutionary new technologies will emerge in the future to promote the great progress of RAID.

This article comes from the official account of Wechat: ZTE documents (ID:ztedoc)

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.