Detailed description of RAID level and implementation instance operations

Detailed description of RAID level and implementation instance operations

I. Brief History of RAID

1. RAID Origin

Berkeley: Berkeley proposed: A case for Redundent Arrays of Inexpensive Disks is RAID (cheap Redundant Array)

Today: Redundant Arrays of Independent Disks Independent Redundant Array

2. Advantages and Performance

1) Improve IO capability through parallel disk read/write:

2) improve durability through the Disk redundancy (fault tolerance) mechanism:

3. Level: multiple disks are organized in different ways.

RAID0, RAID1, RAID2, RAID3, RAID4, RAID5, RAID6, RAID7, RAID10, RAID01, RAID50, JBOD

4. Implementation Method

(1) Software Simulation Implementation: Software RAID, combined with the multi devices tool in the kernel

Generally, it is not used in the production environment.

(2) Hardware Implementation: hardware-level RAID is implemented by configuring multiple hard disks in bios.

External disk array: Provides adaptation capability through expansion card

Internal RAID: Integrated RAID Controller for the motherboard

1) the serial port attached storage can be extended through the adapter interface of the SAS interface, which is represented by the logical unit number of the logical unit number.

Narrowband: 8 interfaces, 7 targets

Broadband: 16 interfaces and 15 targets

2) The RAID Controller must be used by a system driver. You can set it in the BIOS and connect it to the RAID disk array through an adapter.

3) The RAID Controller has a CPU and memory to accelerate. In addition, it sets the power mode to enable emergency writing during power failure.

Ii. RAID level and working mode

Common levels: RAID-0, RAID-1, RAID-5, RAID-10, RAID-50, JBOD

1. RAID-0: 0, strip volume, strip;

Performance: improved reading and writing performance; no fault tolerance capability

Space usage: N * min (S1, S2 ,...),

Minimum number of disks required: at least two disks

Strip: Creates Strip Based on the smallest disk capacity.

2. RAID-1: 1, mirror volume, mirror;

Performance: read performance is improved, write performance is slightly reduced, and redundancy is available.

Space usage: 1 * min (S1, S2,...), with the same storage content

Space Utilization: 1/2

Minimum number of hard disks required: at least two disks

3. RAID-4: A single hard disk is used as a centralized verification disk.

The independent access method is used to change a small data block with a strip from RAID3 to a larger data block.

The verification is fast, but the write speed is very low, and the controller design is more complex.

Stores the verification code in one hard disk by exclusive or operation.

A hard disk may be damaged, but the performance bottleneck may be caused by centralized verification.

4. RAID-5: cyclic verification storage mechanism

Performance: Read and Write performance is improved, with Fault Tolerance: 1 Disk

Spatial use: (N-1) * min (S1, S2,...) Short Board Effect

Space Utilization (n-1)/n

Minimum number of disks: at least three disks, which can be downgraded

5. RAID-6: cyclic verification storage mechanism. One hard disk is dedicated to the verification disk, and the verification code is stored twice.

Performance: improved Read and Write Performance; Fault Tolerance: 2 Disks

Space usage: (N-2) * min (S1, S2 ,...)

Minimum number of disks: Minimum four disks

6. RAID-7: Commercial Version

The new RAID standard, an independent "storage computer", is equipped with an intelligent real-time operating system and a software tool for storage management. It can run independently of the host without occupying host CPU resources.

7. RAID-10: first make the image and then make the Strip

Performance: improved reading and writing with Redundancy

Space Utilization: 1/2

Minimum number of disks: at least 4 disks required

8. RAID-01: First strip and then mirror

Performance: improved reading and writing with Redundancy

Space Utilization: 1/2

Minimum number of disks: at least 4 disks required

9. RAID-50:

Performance: improved reading and writing; redundant

Space Utilization: (n-2)/n

Minimum number of disks: at least 6 disks required

10. JBOD: Just a Bunch Of Disks combines the space Of multiple Disks into a large continuous space.

Performance: no improvement; no redundancy

Space usage: sum of disk sizes (S1, S2 ,...)

Space Utilization: 100%

Minimum number of disks: at least two disks are required.

Iii. RAID kernel implementation and observation commands

Implementation of Software RAID on CentOS 6

1. md (multi devices) tool, mdadm: Modeled Tool

Command syntax: mdadm [mode] <raiddevice> [options] <component-devices>

Supported RAID levels: LINEAR mode, RAID0, RAID1, RAID4, RAID5, RAID6, RAID10;

Mode: Create:-C; assemble:-A; Monitor:-F; manage:-f,-r,-

<Raiddevice>:/dev/md #

<Component-devices>: any block Device

(1)-C: Creation Mode

-N #: Use # Block devices to create the RAID;

-L #: indicates the level of the RAID to be created;

-A {yes | no}: automatically creates the device file of the target RAID device;

-C CHUNK_SIZE: indicates the block size. The default value is 512 KB.

-X #: specify the number of idle disks;

(2) Management Mode:

-F: indicates that the specified disk is damaged;

-A: Add a disk.

-R: Remove Disk

(3)-D: displays raid details. mdadm-D/dev/md #

(4) observe the md status: cat/proc/mdstat

(5) Stop the md device: mdadm-S/dev/md #

2. watch: Dynamic Observation command running process

Format: watch-n # 'command'

-N #: refresh interval, in seconds;

How to build a RAID 10 array on Linux

Debian soft RAID Installation notes-use mdadm to install RAID1

Common RAID technology introduction and demo (Multi-chart)

The most common disk array in Linux-RAID 5

RAID0 + 1 and RAID5 Performance Test Results

Getting started with Linux: disk array (RAID)

This article permanently updates the link address: