Btrfs (Btree FS) file system

Btrfs (Btree FS) file system:
is a copy to write file system developed by Oracle in 2007 to enable large-scale storage of missing pools, snapshots, checksums and cross-device access issues in the file system.

Btrfs Features:
1.btrfs can consist of multiple underlying physical volumes, RAID enabled, online Add, remove, modify
2. Copy-on-write update mechanism (CoW): Copy, update, and replace pointers instead of "in-place" updates;
3. Extensibility: The overall performance after expansion does not decrease with the increase of system capacity
4. Data consistency: Btrfs uses the cow mechanism to support the checksum, ensuring the integrity of the data and determined that the unexpected hardware failure caused the loss
5. Sub-Volume: sub_volume#
6, Snapshot and clone: Snapshot support snapshot; a copy of a sub-volume, equivalent to a snapshot in VMware
7. Transparent compression: Automatically compress the data, non-user interaction, automatic decompression when viewing


Btrfs File System creation:

Mkfs.btrfs Common parameters:
-l ' LABEL ' #设定文件系统卷标
-D <type>: Raid0, RAID1, RAID5, Raid6, RAID10, single# indicates the type of data storage
-M <profile>: RAID0, RAID1, RAID5, Raid6, RAID10, single, DUP #语言数据存储机制
-O <feature> #格式化时启动的特性
-O List-all: Lists all supported feature;


First we add a few hard drives to do the Btree file system:

[[email protected] ~]# fdisk -l disk  /dev/sda:21.5 GB, 21474836480  bytes, 41943040  sector units =  Sector  of 1 * 512 = 512 bytes sector size (logical/Physical): 512   bytes  / 512  bytes i/o  size (min/Best):512  bytes  / 512  BYTE disk label type: DOS disk identifier:0x000f09eb    device  Boot      Start          End      Blocks    id  system/dev/sda1   *        2048      1026047      512000   83   Linux/dev/sda2         1026048     41943039    20458496   8E  LINUX LVM Disk  /dev/sdb:21.5  GB, 21474836480  bytes,41943040  Sector UNITS =&NBSP, Sector  of 1 * 512 = 512 bytes sector size (logical/Physical):512  bytes  / 512  byte i/o  size (min/Best):512  bytes  / 512  bytes Disk  /dev/sdc:21.5 GB, 21474836480  BYTE,41943040  sector units =  sector  of 1 * 512 = 512 bytes sector size (logical/physical) :512  bytes  / 512  bytes i/o  size (min/Best):512  bytes  / 512  bytes

Here we use SDB and SDC two disk directly formatted as Btrfs file system:

[[email protected] ~]# mkfs.btrfs-l mybtrfs/dev/sdb/dev/sdcbtrfs v3.16.2 See http://btrfs.wiki.kernel.org for more info Rmation. Turning on Incompat feature ' extref ': increased Hardlink limit per file to 65536adding DEVICE/DEV/SDC ID 2FS created Labe L Mybtrfs on/dev/sdb# Volume label created on SDB for "Mybtrfs" nodesize 16384 leafsize 16384 sectorsize 4096 size 40.00gib# space Total 40G

Is it easy to create a Btrfs file system?


Now that the creation is complete, let's take a look at the Btrfs specific information:

Btrfs FileSystem Show #btrfs文件系统的查看命令

[Email protected] ~]# btrfs filesystem showlabel: ' mybtrfs ' UUID:1C74225E-A36B-46B2-8BDE-BEF50D0F89FB total devices 2 FS bytes used 112.00KiB devid 1 size 20.00GiB used 2.03GiB path/dev/sdb devid 2 size 20.00GiB used 2.01GiB Path/dev/sdcbtrfs v3.16.2

We can also use Blkid [/dev/device] to view:

[[email protected] ~]# blkid/dev/sdb/dev/sdc# View two disk information/dev/sdb:label= "Mybtrfs" uuid= " 1C74225E-A36B-46B2-8BDE-BEF50D0F89FB "uuid_sub=" 87045ae0-ae5f-4e56-811c-1b246532047c "TYPE=" Btrfs "/dev/sdc:label = "Mybtrfs" uuid= "1C74225E-A36B-46B2-8BDE-BEF50D0F89FB" uuid_sub= "6c491f79-b13e-44c3-94e0-ad7e895b7613" TYPE= " Btrfs

#包括卷标名LABLE, Uuid,type type information, and we noticed that the UUID of the two disks is the same because the two disks belong to the same btrfs system, and the sub volumes are different, Btrfs supports multiple sub-volumes

Then we can mount the Btrfs file system.

[[email protected] ~]# mkdir /mydata# Create a mount point [email  protected] ~]# mount -t btrfs /dev/sdb /mydata/#挂载btrfs文件系统 [[email  protected] ~]# mount | tail -n 5   #mount查看最后一个挂载成功mqueue  on  /dev/mqueue type mqueue  (Rw,relatime,seclabel) Debugfs on /sys/kernel/debug type  debugfs  (rw,relatime)/dev/sda1 on /boot type xfs  (Rw,relatime,seclabel, Attr2,inode64,noquota)/dev/sr0 on /media/cdrom type iso9660  (ro,relatime)/dev/sdb  on /mydata type btrfs  (Rw,relatime,seclabel,space_cache) 

Note: Mount mounts because two disks are btrfs, so just mount any block of two disks

Btrfs filesystem:
    show View information
     show--all-devices: View all devices
     [device]  View a specific device's Btrfs
     DF: Usage rate of Mounted btrfs file system
     Resize: Modify Btrfs file system size

[[email protected] ~]# btrfs filesystem show --all-deviceslabel:  ' Mybtrfs '   uuid: 1c74225e-a36b-46b2-8bde-bef50d0f89fb    total devices 2  fs bytes used 896.00kib    devid    1 size  20.00GiB used 2.03GiB path /dev/sdb    devid     2 size 20.00GiB used 2.01GiB path /dev/sdcBtrfs v3.16.2[[email  Protected] ~]# btrfs filesystem show /dev/sdc failed to open /dev /fd0: no such device or addresslabel:  ' Mybtrfs '   uuid:  1c74225e-a36b-46b2-8bde-bef50d0f89fb    total devices 2 fs bytes  Used 896.00kib    devid    1 size 20.00gib used  2.03gib path /dEv/sdb    devid    2 size 20.00gib used 2.01gib  path /dev/sdc

Btrfs v3.16.2 #因为属于同一个文件系统, so looking at one shows all


After the Btrfs file system is created, we enter the directory to create the file to test for success:

[Email protected] ~]# cd/mydata/[[email protected] mydata]# Ls[[email protected] mydata]# touch A.txt[[email protected] mydata]# cp/etc/grub2.cfg/[[email protected] mydata]# ls #可以添加文件, btrfs normal use a.txt grub2.cfg

Uninstall we can use Umount [/dev/sd#] to uninstall

We can also use Btrfs's built-in transparent compression mechanism when mounting file systems, so that users do not need to participate:

Mount-o compress={lzo|zlib} [DEVICE] [Mount_point]
Supports two compression algorithms for Lzo and zlib

After setting this parameter we will store the data in Btrfs file system automatically by Lzo or zlib compression mechanism, that is, transparent compression

[[email protected] ~]# umount/dev/sdb[[email protected] ~]# mount-o compress=lzo/dev/sdb/mydata/[[email protected] ~]# Cp/etc/rc.d/init.d/functions/mydata/[[email protected] ~]# cd/mydata/[[email protected] mydata]# ll-h functions/etc/ Rc.d/init.d/functions-rw-r--r--. 1 root root 14K January 2015/etc/rc.d/init.d/functions-rw-r--r--. 1 root root 14K September 3 15:21 functions

Because the functions file is too small, we see that the size does not change, large files can see the size of the compressed

Btrfs can also be added online to reduce file system size by:

Btrfs filesystem resize + size [Mount_point]

[[Email protected] mydata]# btrfs filesystem resize -10g /mydata/resize   '/mydata/'  of  ' -10g ' #这样原来的btrfs文件系统大小减少10G大小 [[Email protected] mydata]# btrfs  filesystem df /mydata# View Btrfs usage data, raid0: total=2.00gib, used=776.00kibdata,  single: total=8.00mib, used=0.00system, raid1: total=8.00mib, used= 16.00kibsystem, single: total=4.00mib, used=0.00metadata, raid1: total=1.00gib,  Used=112.00kibmetadata, single: total=8.00mib, used=0.00globalreserve, single: total= 16.00MIB, USED=0.00[[EMAIL PROTECTED] MYDATA]# DF -LH File System                    capacity    used    available   used%  mount point/dev/mapper/centos-root   18g  4.2g   14g    24% /devtmpfs                 481m      0  481M    0% /devtmpfs                     490M    72K  490M    1% /dev/shmtmpfs                     490m   14m   477M    3% /runtmpfs                     490m     0   490M    0% /sys/fs/cgroup/dev/sda1                 497M  125M  373M    25% /boot/dev/sr0                 7.1g   7.1g     0  100% /media/cdrom/dev/sdb                   30G  1.1M    18g    1% /mydata# is now 30G

This way, Btrfs is reduced to 30G by the original 40G, while the CP and the files created before us are unaffected, provided there is room for the data to be stored.

[[email protected] mydata]# lsa.txt functions grub2.cfg[[email protected] mydata]# cat grub2.cfg #文件照样可以查看 # # do not EDIT This file## It was automatically generated by Grub2-mkconfig using templates# from/etc/grub.d and Settings From/etc/defa ult/grub#### Begin/etc/grub.d/00_header # # #set pager=1if [-S $prefix/grubenv]; Then Load_envfiif ["${next_entry}"]; Then set default= "${next_entry}" set next_entry= save_env next_entry set Boot_once=trueelse

Since we can reduce the size of the space, let's try to increase its size.

[[Email protected] mydata]# btrfs filesystem resize +10g /mydata/resize   '/mydata/'  of  ' +10g ' #这里我们增加10G大小 [[email protected] mydata]# [[email  protected] mydata]# btrfs filesystem showlabel:  ' Mybtrfs '   uuid:  1c74225e-a36b-46b2-8bde-bef50d0f89fb    total devices 2 fs bytes  Used 904.00kib    devid    1 size 20.00gib used  2.03GiB path /dev/sdb    devid    2 size  20.00gib used 2.01gib path /dev/sdcbtrfs v3.16.2[[email protected] mydata]#  LL -LH Total dosage  20k-rw-r--r--.  1 root root    0 9 month     3 15:00 a.txt-rw-r--r--.  1 root root  14k 9 Month     3 15:21 functions-rw-R--r--.  1 root root 4.0k 9 Month    3 15:00 grub2.cfg[[email  protected] mydata]# df -lh /mydata/File System          capacity    used    available   used%  mount point/dev/sdb          40g  1.1m   38g    1% /mydata# this again for the original 40G

[Email protected] mydata]# btrfs filesystem resize Max/mydata/resize '/mydata/' of ' Max ' #指定扩展为最大容量

Btrfs supports increasing online to reduce the size of existing capacity, but what if the loaded disk capacity is not enough??

The Btrfs file system also supports the direct addition of disk devices to the file system:

Btrfs device Add [Divice] [Mount_point]

Here we have a reserved disk. SDD, let's look at the effect:

[[email protected] mydata]# btrfs device add /dev/sdd /mydata/#直接加载 [email  PROTECTED] MYDATA]# DF -LH File System                    capacity    used    available   used%  mount point/dev/mapper/ Centos-root   18g  4.2g   14g   24% /devtmpfs                  481m      0  481M    0% /devtmpfs                     490m    72K  490M    1% /dev/shmtmpfs                     490M    14M  477M    3% /runtmpfs                     490M     0  490M     0% /sys/fs/cgroup/dev/sda1                 497m  125m  373m   25% /boot/dev/sr0                  7.1g   7.1G     0  100% /media/cdrom/dev/sdb                   60g  1.1m    56g    1% /mydata# we see that the load is increased from the original 40G size to 60G size

The Btrfs file system also supports data file equalization operations: (usually rarely done)
Btrfs balance start [Mount_point]
Pause: Paused
Cancel: Canceled
Resume: Continue
Status: State

[[email protected] mydata]# Btrfs balance Start/mydata/done, had to relocate 6 out of 6 chunks# migrated 6 Chunks[[email Protect Ed] mydata]# btrfs balance status/mydata/no balance found on '/mydata/'

#状态信息由于文件很小, execution has ended and viewing can only be seen during the migration process, and large files are available.

To create a sub-volume:
Btrfs subvolume Create [mount_point]/[files]

[[email protected] mydata]# btrfs subvolume Create use/mydata/logs# creating sub-volume Creation Subvolume '/mydata/logs ' [[email protected] mydata]# btrfs subvolume list/mydata/#列出子卷信息ID 260 Gen-top Level 5 path logs

Mount a sub-volume:

[Email protected] ~]# umount/mydata/#需要先卸载挂载点 [[email protected] ~]# mount-o subvol=logs/dev/sdb/mnt/#挂载子卷 [[Email Prot Ected] ~]# ls/mnt/#logs没有文件 [[email protected] ~]# cp/var/log/messages/mnt/#cp文件到子卷可以 [[email protected] ~]# ls /mnt/messages

[[email protected] ~]# btrfs subvolume show /mnt/#查看子卷详细信息/mnt     Name:             logs     uuid:              cb76840f-56df-db4e-8077-2faf03a8351a    parent uuid:          -    Creation time:          2015-09-03 16:37:54    Object ID:          260    Generation  (Gen):      55     Gen at creation:     52     parent:         5    top level:          5    flags:             -     snapshot (s):

Delete a sub-volume using: Btrfs subvolume delete/mydata/logs


There is also a snapshot of the creation of the sub-volume, no specific action is written.

[Email protected] ~]# Btrfs subvolume snapshot/mydata/logs/mydata/logs_snapshot

Our Btrfs file system is introduced here, such as where there are problems also hope that the great God pointed out, thank you!

This article is from the "Open Source Prosperity" blog, please be sure to keep this source http://zhaohongfei.blog.51cto.com/6258685/1691126

Btrfs (Btree FS) file system