RAID disk array

One 、RAID Introduce

RAID, Full name Redundant Arrays of Inexpensive Disk , Independent redundant disk array .

Disk arrays are It consists of many independent physical hard disks , Combined into a large capacity hard disk group , So as to provide higher storage performance than a single hard disk and provide data backup technology .

RAID Can be realized including Improve I/O（ Reading and writing ） Ability 、 Improve durability 、 Improve data redundancy And so on .

DAID How to implement ：

• External connection ： Provide adaptability through expansion cards
• Built in ： motherboard integrated DAID controller , It is usually installed before the operating system BIOS Set in the
• Software style ： Through the software of the operating system , For example, Qunhui's NAS

The different ways in which the entire column of disks is composed are called RAID Level , Most commonly used RAID There are four kinds of ：
RAID 0、RAID 1、RAID 5、RAID 10（10 Not ten , It is 1+0, Do it first RAID 1, Do it again RAID 0）

however , This is not RAID All levels of , There are also RAID 2、DAID 3、RAID 4、RAID 6、RAID 50 wait , It's just that it's not often used in production environments for performance reasons , So as to be eliminated or carried away by a stroke .

Two 、 Several commonly used RAID

1.RAID 0

RAID 0 , Also called striped storage . It's in all RAID Highest storage performance in class , The principle is to store a continuous piece of data on different hard disks （ The process is carried out simultaneously ）. If you want to achieve RAID 0, At least two hard disks are required .

If I have a continuous piece of data 123456789, Three hard disks are used at the same time a,b,c Did RAID 0, Then its storage mode is 1 save a,2 save b,3 save c,4 save a,5 save b,6 save c,7 save a,8 save b,9 save c, And so on , And the stored processes are simultaneous , thus it can be seen , With the same data ,RAID 0 The storage speed of increases with the increase of hard disk **（ Single hard disk N times ,N by RAID 0 Number of hard disks in ）**.

however ,RAID 0 Just to improve the performance , and There is no guarantee for the reliability of the data , A disk failure will affect all data , therefore RAID 0 It can't be used in the situation with high data security requirements .

It can be seen from the above figure , When a hard disk hangs up , Some data will be lost , This will affect the integrity of the data .

2.RAID 1

RAID 1 , Also known as mirrored storage . It is realized through disk data mirroring data redundancy , stay Pair Data backed up by each other is generated on independent disks of . Generally speaking , Is that all （ The number of hard disks is even , namely 2n）RAID 1 All data is written to the hard disk in , Its capacity is only equal to the capacity of a hard disk , Because the other piece is just a data mirror . When raw data is busy , Data can be read directly from the mirror copy .

RAID 1 Of Writing speed is usually slow , Because the data has to be written into two hard disks and compared , But it's Data reading is faster than a single hard disk , Because the data will be read out from the faster one of the two hard disks .

RAID1 It is in the hard disk The unit cost is the highest Of , but It provides high data security and availability , When a hard disk fails , The system can Automatically switch to the mirrored hard disk for reading / Write , also There is no need to reorganize invalid data .

It can be seen from the above figure , When a hard disk hangs up , The system will read the complete data information from the backup hard disk , Does not affect the integrity of the data .

3.RAID 5

RAID 5 It's a storage capability 、 A storage solution that combines data security and storage cost , Its At least three hard disks are required to form an array .

RAID 5 With and RAID 0 Similar data reading speed , It's just There is one more parity information , therefore The speed of writing data is slightly slower than that of a single disk .

Since multiple data correspond to one parity information ,RAID 5 The disk space utilization ratio of RAID 1 high , by n-1/n（n Is the number of hard disks ）

RAID 5 Of High reliability , It is allowed to damage a disc , When a disk is damaged ,RAID 5 Meeting Using the remaining data and the corresponding parity information to recover the damaged data , Thus, the integrity of image data will not . Of course , If two disks break at the same time , Or when one disk is damaged and the problem is not solved, another disk is broken , Then only GG 了 .

As shown in the figure above , If disk 0 damage , The other three hard disks can recover data according to the verification information , So it does not affect the integrity of the data

4.RAID10

RAID10, refer to RAID 1 + 0 , Do it first RAID 1 , Do it again RAID 0 . it yes RAID 1 and RAID 0 Products of combination , Divide data in consecutive bits or bytes and read in parallel / Write multiple disks at the same time , For each disk as a disk image for redundancy .

RAID 10 meanwhile Have RAID 0 Fast read and write speed and RAID 1 High reliability of data , however CPU The occupancy rate is also higher , and Disk utilization is low （n/2）.

Because of the use of RAID 0 High reading and writing efficiency and RAID 1 High data protection 、 Resilience , send RAID 10 It has become a high cost-effective grade , At present, almost all RAID All control cards support this level .

3、 ... and 、Linux System construction software RAID

stay Linux In the system , To build RAID disk array , Need to use mdadm Tools .

Format ： mdadm [ Pattern ] <raiddevice> [ Options ] <component-device>

Options	Pattern
-C	Create mode
-A	Assembly mode
-F	Monitoring mode
-f,-r,-a	management model

management model	effect
-f	Mark the specified disk as damaged
-r	Remove disk
-a	Add disks

Options	effect
-n #	Use # Devices to create this RAID
-l #	Specify the created RAID The level of
-x #	Specify the size of the free disk
-c	Specify block size , Unit is k
-D	Show RAID Details of

below , I will demonstrate to you in detail Linux Build in RAID 5 The process of ：

1. First , We now add the virtual machine management interface 4 Block hard disk . add to 4 The purpose of the block is to 3 Block making RAID 5 , The rest is a hot spare , namely RAID 5 If one disc is damaged , The hot spare can be directly replaced .
   
   

2. yum install -y mdadm
   

3. mdadm -Cv md0 -l 5 -n 3 /dev/sd[b-d] -x 1 /dev/sde
   

4. mdadm -D /dev/md/md0
   

5. format
   

6. mount
   

7. simulation sdd damage