Linux File System

After the system partition is completed, the partition will be formatted with the file system (the file system is a management software that is stored to a location on the disk partition, the file system is not the entire partition ) but the data on the filesystem is on this partition, so the file system is a management software.

The file system divides the disk into two pieces: the metadata store (metadata)(inode bitmap, block bitmap, inode entry, and so on), data store ( datastore Also divided into multiple logical storage units called disk blocks (logical concept))

Data in any file system is divided into data and meta data (metadata). Data refers to the actual data in a normal file, and metadata refers to system data that describes the characteristics of a file, such as access rights, file owners, and distribution information for file data blocks (Inode ...). Wait a minute. But the metadata ★ does not include the file name ★ File name is stored on the disk block .

The following are the common file system types:

1.ext4

650) this.width=650; "border=" 0 "src=" http://s2.51cto.com/wyfs02/M01/7A/18/wKioL1aiB2eD3tx4AACplPTxmYg796.jpg " Style= "Border:none;"/>

File-systems-ext4

If you've ever installed Linux, you may have seen the word "EXT4" during installation. There is a good reason to use it: It is the file system that is currently selected for each available Linux distribution. There are, of course, other options, but it is undeniable that EXT4 (Extended 4) is the file system that almost all Linux users will choose.

What can it do?

EXT4 has all the benefits you expect in the previous file system (EXT2/EXT3), along with some improvements. There's a lot more to explore, and here's a list of the best things EXT4 brings you:

journaled File System

Log check

Multi-block file allocation

Backwards compatible EXT2 && EXT3

Continuous free space pre-allocation

Improved file system checksum (compared to previous versions)

Of course, supporting larger files at the same time

Target User

EXT4 is for those who are looking for a super-reliable base environment or those who just need to work. This filesystem does not take a snapshot of your system, and it doesn't even have the best SSD support, but if you're not too picky, you'll think it's good.

2.BtrFS

650) this.width=650; "border=" 0 "src=" http://s3.51cto.com/wyfs02/M01/7A/18/wKiom1aiByjC5OTCAACLbZxzPfw306.jpg " Style= "Border:none;"/>

File-systems-btrfs

The B-tree (b-tree) file system (also known as the Butterfs, butter file system) is a file system developed by Oracle for Linux. It is a completely new file system and is in the main development phase. The Linux community thinks it sometimes has some instability in its use. BtrFS's core principles revolve around the principle of copy-on-write (Copy-on-write). copy -on-write basically means that each bit of the data has a separate copy before the write data is complete. When the data is written, it is equivalent to making a copy of it.

What it can do

In addition to supporting write-time replication, BtrFS is capable of many other transactions-in fact, it can keep a list of features. Here are some of the most noteworthy features: support for read-only snapshots, file clones, sub-volumes, transparent compression, offline file system checksums, conversion from Ext3 & 4 to BtrFS, online defragmentation, and support for RAID 0, RAID 1, RAID 5, RAID 6, and RAID 10.

Target User

BtrFS's developers have promised that the filesystem is a new generation of alternatives to the current file system. Very correct, although it is currently in development. It has many killer features for advanced users as well as for basic users (including better performance on SSDs). This file system is intended for users who want to get more from the file system, and those who want to try to do something with the copy-on-write mechanism.

3.XFS

650) this.width=650; "border=" 0 "src=" http://s4.51cto.com/wyfs02/M02/7A/18/wKiom1aiBymDf1CdAACgNDjU5TI288.jpg " Style= "Border:none;"/>

File-systems-xfs

Created and developed by Silicon Graphics, XFS is a high-end file system positioned for speed and performance. The focus on performance makes XFS particularly well-behaved in terms of parallel IO. The XFS file system is capable of processing large amounts of data, in fact some XFS users are approaching 300TB.

What it can do

XFS is a well-tested data storage file system that was created for high-performance operations. The features include:

Striped Allocation for RAID arrays

journaled File System

Variable block size

Direct I/O

Specify rate (guaranteed-rate) I/O

Snapshot

Online defragmentation

Adjust file system size online

Target User

XFS is intended for users who want a rock-solid file system scenario. It began in the 1993 and it became better and more as time changed. If you have a home server and you're upset about how to deploy a storage environment, consider XFS. It has many features (such as snapshots) that can help your file storage system. However, it is not limited to the server side. If you are a relatively advanced user or you are interested in many of the features that BtrFS promises, try XFS. It implements a lot of features similar to BtrFS, and there is no stability problem.

4.reiser4

650) this.width=650; "border=" 0 "src=" http://s4.51cto.com/wyfs02/M02/7A/18/wKioL1aiB2fwiY1yAACNkQ6H3Qc392.jpg " Style= "Border:none;"/>

File-system-riser4

Reiser4 is ReiserFS's successor, created and developed by Namesys company. Its birth dates back to the Linspire project and DARPA. What makes it different is that there are many transactional patterns. And not just a single way of writing data; instead, there are many ways (to write).

What it can do

Reiser4 has the unique ability to use a variety of different transactional patterns. It is able to use write-time copy mode (like BtrFS), write-Anywhere (write-anywhere), log, and mixed transaction mode. It has made a number of improvements based on ReiserFS, including better log-based file system logs, better support for smaller files, and faster directory processing. Reiser4 offers a number of feature features. There are more features that can be explored, but simply speaking, it has not only made a great improvement over ReiserFS, but has also added many features.

Target User

Resier4 is ideal for users who want to apply a file system to a variety of scenarios. You might want to use a copy-on-write mechanism on one machine, use an arbitrary-location-write mechanism on another machine, and use mixed transactions on another machine, and you don't want to use many different types of file systems to accomplish this task. Reiser4 is the perfect solution for this situation.

Supplement: Common File system types and scenarios

EXT2 lovingly, CDN related, such as the Blue Flood cache service, some CDN website acceleration Service

EXT3 centos5 system Default File system

EXT4 CENTOS6 system Default File system

ZFS Solaris The default file system

XFS Cenos7 The default file system

ReiserFS Suselinux default file system, processing a large number of small files such as 100k below, recommend this system

IBM small machine using a JFS file system

SunOS use a UFS file system.

File System Selection:

The number of servers is relatively small, no special requirements for casual selection.

CentOS The default file system for the series is excellent.

A large server, it is recommended to choose a compromise file system.

This article from "Write a blog Money" blog, please be sure to keep this source http://dadonggg.blog.51cto.com/12672150/1949387

Linux File System