Linux Disk Management--raid principle

RAID (redundant Arrays of independent Disks) redundant array of independent disks, primarily used to store data at different levels of level at the logical level by combining multiple physical disks or disk partitions at the hardware level The result is an increase in data storage performance metrics. Hierarchies are not sorted by number size, but simply represent a type of storage. This article mainly describes the principle and characteristics of the common different levels of RAID.

RAID metrics:

1. IO performance: Read/write rate of data in raid

2. Redundancy: Ability to perform redundant backups of data

3. Disk space: Disk space after the raid, set the disk for RAID is n block size s of the disk,

4. Disk Requirements: Required number of disks required

RAID Key Concepts:

The size of the storage unit in the Chunk:raid, specifying the appropriate chunk size when the disk is built into raid

RAID-0

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M02/72/52/wKioL1Xg73KQExDRAAE6nBpL9Ss387.jpg "title=" Raid0.jpg "alt=" Wkiol1xg73kqexdraae6nbpl9ss387.jpg "/>

As shown in the RAID-0 principle, the main implementation is to make a large logical disk of multiple physical disks, in the data IO when the individual chunk will be written to the polling physical disk, RAID-0 mainly to achieve the expansion of the partition space.

RAID0 Performance Analysis:

1. IO performance: Improved read and write performance, because reading and writing data while reading and writing data to multiple disk space, will increase the read and write rate

2. Redundancy capability: No redundant backup of data is implemented,

3. Disk space: NS

4. Disk Requirements: Minimum 2 disks required

5. Fault tolerance: No disk damage allowed

RAID-1

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M02/72/55/wKiom1Xg7VfA3a9mAAEynot_MHk465.jpg "title=" Raid1.jpg "alt=" Wkiom1xg7vfa3a9maaeynot_mhk465.jpg "/>

As shown in the RAID-1 principle, the main implementation is to make a number of physical disks two of the same size of the main spare disk group, the data IO when the individual chunk will be written to two disk groups, RAID-1 mainly to achieve redundant storage of data.

RAID1 Performance Analysis:

1. IO performance: Read the data because there are multiple data backup, the reading rate will increase; Write rate will drop when writing data because multiple chunk data blocks are written at the same time

2. Redundancy capability: Store multiple data backups

3. Disk space: NS/2

4. Disk requirements: Requires a minimum of 2 disks and a multiple of 2

5. Fault tolerance: up to 1 pieces of disk damage allowed

RAID-5

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M00/72/52/wKioL1Xg73KTv9ovAAE6ubmMsao261.jpg "title=" Raid5.jpg "alt=" Wkiol1xg73ktv9ovaae6ubmmsao261.jpg "/>

As shown in the RAID-5 principle, the main implementation of the data IO through the N-1 block disk to read and write data, the nth disk to read and write data of the check block (through the cyclic redundancy check calculation), so that through the cyclic redundancy check the data backup, the disk that holds the check block is polled on the N block disk

RAID5 Performance Analysis:

1. IO performance: Improved read and write performance, because reading and writing data while reading and writing data to multiple disk space, will increase the read and write rate

2. Redundancy capability: ability to provide redundancy

3. Disk space: (n-1) S/n

4. Disk Requirements: Minimum 3 disks required

5. Fault tolerance: up to 1 pieces of disk damage allowed

RAID01

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M00/72/52/wKioL1Xg9mjA-wpqAAK5bXO64fk406.jpg "title=" Raid01.jpg "alt=" Wkiol1xg9mja-wpqaak5bxo64fk406.jpg "/>

As shown in the RAID01 principle, it is primarily through RAID-0 to expand the disk space, and then through the RAID-1 to the disk group redundant backup

RAID01 Performance Analysis:

1. IO performance: Improved read and write performance, because reading and writing data while reading and writing data to multiple disk space, will increase the read and write rate

2. Redundancy capability: ability to provide redundancy

3. Disk space: 1/2

4. Disk Requirements: Minimum 4 disks required

5. Fault tolerance: Do not allow different RAID-0 groups to damage the same location of the disk

RAID10

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M00/72/55/wKiom1Xg7Veg_0v0AALCYTkRobU844.jpg "title=" Raid10.jpg "alt=" Wkiom1xg7veg_0v0aalcytkrobu844.jpg "/>

As shown in the RAID10 principle, it is primarily a redundant backup of the data through RAID-1, and the expansion of the disk group space by RAID-0

RAID10 Performance Analysis:

1. IO performance: Improved read and write performance, because reading and writing data while reading and writing data to multiple disk space, will increase the read and write rate

2. Redundancy capability: ability to provide redundancy

3. Disk space: 1/2

4. Disk Requirements: Minimum 4 disks required

5. Fault tolerance: The same RAID-1 group is not allowed to damage 2 disks, allowing different RAID-1 groups to damage multiple disks

In practical use, RAID10 is more practical than RAID01.

These are the principles and characteristics of a common RAID disk array, which should be used to select the appropriate category based on the specific usage environment.

This article is from the "Pavel" blog, make sure to keep this source http://pavel86.blog.51cto.com/8349178/1689655

Linux Disk Management--raid principle