-
How LVM works
LVM (Logical Volume Manager) logical volume management is a logical layer added between disk partition and file system to shield the layout of lower disk partition for file system, provide an abstract volume and establish file system on the volume. The administrator can use LVM to dynamically adjust the size of the file system without repartitioning the disk, and the file system managed by LVM can span the disk. When the server adds a new disk, the administrator does not need to move the original file to the new disk, but can directly expand the file system across the disk through LVM.
It is to encapsulate the underlying physical hard disk and present it to the upper application in the form of logical volume. In LVM, it encapsulates the underlying hard disk. When we operate the underlying physical hard disk, it is no longer for partition operation, but for underlying disk management operation through something called logical volume.
1.1 terms commonly used in LVM
The physical media: LVM storage media can be disk partition, whole disk, RAID array or SAN disk. The device must be initialized as LVM physical volume to be used with LVM.
Physical volume PV (physical volume): physical volume is the basic storage logical block of LVM, but compared with basic physical storage media (such as partition, disk, etc.), it contains management parameters related to LVM. To create a physical volume, it can use hard disk partition or hard disk itself;
Volume Group: an LVM Volume Group consists of one or more physical volumes
logical volume: LV is built on VG, and file system can be built on LV
PE (physical extensions): the smallest storage unit that can be allocated in the PV physical volume. The PE size can be specified, and the default is 4MB
LE (logical extension): the smallest storage unit that can be allocated in LV logical volume. In the same volume group, the size of LE and PE are the same, and one-to-one correspondence
Summary of minimum storage units:
Name minimum storage unit
Hard disk sector (512 bytes)
File system block (1K or 4K) ා mkfs.ext4 -b 2048 /dev/sdb1, maximum support to 4096
raid chunk (512K) #mdadm -C -v /dev/md5 -l 5 -n 3 -c 512 -x 1 /dev/sde{1,2,3,5}
LVM PE (4M) # vgcreate -s 4M vg1 /dev/sdb{1,2}
The main elements of LVM are as follows:
Summary: multiple disks / partitions / raid - "multiple physical volumes PV -" synthetic volume group VG - "logical volume from VG LV -" format LV mount and use
1.2 LVM benefits
Use volume groups to make multiple hard disk spaces look like a large hard disk
With logical volumes, you can partition sdb1 sdb2 sdc1 sdd2 sdf across multiple hard disk spaces
When using a logical volume, it can dynamically resize it when space is low
When adjusting the size of the logical volume, you do not need to consider the location of the logical volume on the hard disk or worry about the lack of available continuous space
LV and VG can be created, deleted and resized online. The file system on LVM also needs to be resized.
Allows you to create a snapshot that can be used to save a backup of the file system.
RAID+LVM: LVM is the volume management method of software, while RAID is the disk management method. For important data, RAID is used to protect the physical disk from business interruption due to failure. LVM is used to realize the benign management of volumes and make better use of disk resources.
2 basic steps to create LVM
1) Physical disks are formatted as PVS (spaces are divided into PE's) and PVS contain PE's
2) Different PV is added to the same VG, (PE of different PV enters the PE pool of VG) ා VG includes PV
3) Create LV logical volume in VG, create based on PE, (PE composing LV may come from different physical disks) ා LV create based on PE
4) LV can be directly used for mounting after formatting
5) In fact, the expansion and reduction of LV is to increase or decrease the number of PE that constitutes the LV, and its process will not lose the original data
2.1 commands commonly used by LVM
function
PV management command
VG management command
LV management command
Scan scan
pvscan
vgscan
lvscan
Create create
pvcreate
vgcreate
lvcreate
Display display
pvdisplay
vgdisplay
lvdisplay
Remove remove
pvremove
vgremove
lvremove
extend extension
vgextend
lvextend
reduce reduction
vgreduce
lvreduce
Some view commands will be used for the following operations:
View volume name
Simple corresponding volume information viewing
Scan all related corresponding volumes
View detailed corresponding volume information
Physical volume
pvs
pvscan
pvdisplay
Volume group
vgs
vgscan
vgdisplay
Logical volume
lvs
lvscan
lvdisplay
2.2 create and use LVM logical volumes
Create PV
Add one sdb disk
[root@xuegod63 ~]# fdisk /dev/sdb #Create 4 primary partitions, 1G each
[root@xuegod63 ~]# ls /dev/sdb*
/dev/sdb /dev/sdb1 /dev/sdb2 /dev/sdb3 /dev/sdb4
//Set partition type code: fdisk / dev / SDB = = > t = = > select partition number = = = > 8e = = = > W
//Note: now the system is very intelligent. You can directly use the default 83 Linux partition or create a pv partition.
[root@xuegod63 ~]# pvcreate /dev/sdb{1,2,3,4} #Create pv
Physical volume "/dev/sdb1" successfully created.
Physical volume "/dev/sdb2" successfully created.
Physical volume "/dev/sdb3" successfully created.
Physical volume "/dev/sdb4" successfully created.
[root@xuegod63 ~]# pvdisplay /dev/sdb1 #View physical volume information
"/dev/sdb1" is a new physical volume of "1.00 GiB"
--- NEW Physical volume ---
PV Name /dev/sdb1
VG Name
PV Size 1.00 GiB
Allocatable NO
PE Size 0
Total PE 0
Free PE 0
Allocated PE 0
PV UUID SHKFwf-WsLr-kkox-wlee-dAXc-5eL0-hyhaTVTo create a vg volume group:
Syntax: vgcreate vg Name pv You can have more than one name pv [root@xuegod63 ~]# vgcreate vg01 /dev/sdb1 Volume group "vg01" successfully created [root@xuegod63 ~]# vgs VG #PV #LV #SN Attr VSize VFree vg01 1 0 0 wz--n- 1020.00m 1020.00m [root@xuegod63 ~]# vgdisplay vg01 --- Volume group --- VG Name vg01 System ID Format lvm2 Metadata Areas 1 Metadata Sequence No 1 VG Access read/write VG Status resizable MAX LV 0 Cur LV 0 Open LV 0 Max PV 0 Cur PV 1 Act PV 1 VG Size 1020.00 MiB PE Size 4.00 MiB Total PE 255 Alloc PE / Size 0 / 0
Create LV
lvcreate -n Specify the name of the new logical volume -L Appoint lv Size SIZE(M,G) (-l: Small l Appoint LE Quantity) vgname [root@xuegod63 ~]# lvcreate -n lv01 -L 16M vg01 Logical volume "lv01" created. [root@xuegod63 ~]# lvcreate -n lv02 -l 4 vg01 Logical volume "lv02" created. [root@xuegod63 ~]# lvs LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert lv01 vg01 -wi-a----- 16.00m lv02 vg01 -wi-a----- 16.00m [root@xuegod63 ~]# pvdisplay /dev/sdb1 --- Physical volume --- PV Name /dev/sdb1 VG Name vg01 PV Size 1.00 GiB / not usable 4.00 MiB Allocatable yes PE Size 4.00 MiB Total PE 255 Free PE 247 Allocated PE 8 # Allocated ['æ l ə ke ɪ t ɪ d] allocation, 8 PE S have been used [root@xuegod63 ~]# vgdisplay vg01 Alloc PE / Size 8 / 32.00 MiB #8 PE's have been used, 32MB Free PE / Size 247 / 988.00 MiB
2.3 file system format and mount
[root@xuegod63 ~]# mkdir /lv01 [root@xuegod63 ~]# ls /dev/vg01/ #View logical volumes lv01 lv02 [root@xuegod63 ~]# ll /dev/vg01/lv01 #In fact, lv01 is the soft link of dm-0 lrwxrwxrwx 1 root root 7 5 Month 1819:02 /dev/vg01/lv01 -> ../dm-0 [root@xuegod63 ~]# mkfs.ext4 /dev/vg01/lv01 [root@xuegod63 ~]# mount /dev/vg01/lv01 /lv01 [root@xuegod63 ~]# df -Th /lv01 //Filesystem type capacity used% free used mount point /dev/mapper/vg01-lv01 ext4 15M 268K 14M 2% /lv01 [root@xuegod63 ~]#echo "/dev/vg01/lv01 /lv01 ext4 defaults 0 0" >> /etc/fstab
2.3 for specifying PE size
Appoint PE Parameters for size: -s ,If the stored data are all large files, then PE As large as possible, fast reading speed [root@xuegod63 ~]# vgcreate -s 16M vg02 /dev/sdb2 Volume group "vg02" successfully created PE And the maximum size is 512 M [root@xuegod63 ~]# vgdisplay vg02 --- Volume group --- VG Name vg02 System ID Format lvm2 Metadata Areas 1 Metadata Sequence No 1 VG Access read/write VG Status resizable MAX LV 0 Cur LV 0 Open LV 0 Max PV 0 Cur PV 1 Act PV 1 VG Size 1008.00 MiB PE Size 16.00 MiB #It's already 16MB
2.4 LV capacity expansion
First, determine whether there is available expansion space, because the space is created from VG, and LV cannot expand across VG [root@xuegod63 ~]# vgs VG #PV #LV #SN Attr VSize VFree vg01 1 2 0 wz--n- 1020.00m 988.00m vg02 1 0 0 wz--n- 1008.00m 1008.00m The command used is as follows: Expand logical volume [root@xuegod63 ~]# lvextend -L +30m /dev/vg01/lv01 Note: when specifying the size, it is different to expand 30m and to 30m Capacity expansion 30m = = = = = - L + 30m Expand capacity to 30m = ===== > L 30m [root@xuegod63 ~]# lvextend -L +30m /dev/vg01/lv01 Rounding size to boundary between physical extents: 32.00 MiB. Size of logical volume vg01/lv01 changed from 16.00 MiB (4 extents) to 48.00 MiB (12 extents). Logical volume vg01/lv01 successfully resized. [root@xuegod63 ~]# lvs LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert Lv01 vg01 - wi Ao ---- 48.00m × LV has been successfully expanded lv02 vg01 -wi-a----- 16.00m [root@xuegod63 ~]# df -Th /lv01 Filesystem type capacity used% free used mount point /dev/mapper/vg01-lv01 ext4 15M 268K 14M 2% /lv01
Note: it can be seen that although LV has been expanded, the file system size is still the same. Now we start to expand the file system
ext4 file system expansion command syntax: resize2fs logical volume name
xfs file system expansion command syntax: xfs ﹣ growfs mount point
The difference between resize2fs and xfs'growfs is that the parameters passed are different. Xfs'growfs is the mount point used; resize2fs is the logical volume name, and the resize2fs command cannot be used for XFS type file systems
[root@xuegod63 ~]# resize2fs /dev/vg01/lv01 resize2fs 1.42.9 (28-Dec-2013) Filesystem at /dev/vg01/lv01 is mounted on /lv01; on-line resizing required old_desc_blocks = 1, new_desc_blocks = 1 The filesystem on /dev/vg01/lv01 is now 49152 blocks long. [root@xuegod63 ~]# df -Th /lv01 //Filesystem type capacity used% free used mount point /dev/mapper/vg01-lv01 ext4 46M (Expansion success) 522 K 43M 2% /lv01 [root@xuegod63 ~]# lvextend -L 80M -r /dev/vg01/lv01 #Direct expansion to 80M space, one step in place, no need to expand the file system [root@xuegod63 ~]# df -T /lv01/ //File system type 1K - block used% used% mount point /dev/mapper/vg01-lv01 ext4 78303 776 73761 2% /lv01 [root@xuegod63 ~]# df -Th /lv01/ //Filesystem type capacity used% free used mount point /dev/mapper/vg01-lv01 ext4 77M 776K 73M 2% /lv01
2.5 VG capacity expansion
[root@xuegod63 ~]# vgs VG #PV #LV #SN Attr VSize VFree vg01 1 2 0 wz--n- 1020.00m 924.00m vg02 1 0 0 wz--n- 1008.00m 1008.00m vg Expansion scenario: vg There is not enough space in the volume group. You need to add a new hard disk [root@xuegod63 ~]# pvcreate /dev/sdb3 # Create pv [root@xuegod63 ~]# vgextend vg01 /dev/sdb3 #Capacity expansion Volume group "vg01" successfully extended [root@xuegod63 ~]# vgs VG #PV #LV #SN Attr VSize VFree vg01 2 2 0 wz--n- 1.99g <1.90g vg02 1 0 0 wz--n- 1008.00m 1008.00m
2.6 LVM reduction
LVM can be increased dynamically, can it be reduced dynamically?
Answer: LVM can be increased or reduced dynamically, but XFS does not support dynamic reduction, so we cannot achieve dynamic reduction based on XFS. btrfs file system supports online reduction.
[root@xuegod63 ~]# lvreduce -L -20m /dev/vg01/lv01 WARNING: Reducing active and open logical volume to 60.00 MiB. THIS MAY DESTROY YOUR DATA (filesystem etc.) Do you really want to reduce vg01/lv01? [y/n]: y Size of logical volume vg01/lv01 changed from 80.00 MiB (20 extents) to 60.00 MiB (15 extents). Logical volume vg01/lv01 successfully resized. #Reduce success //But the file system did not shrink successfully: [root@xuegod63 ~]# df -h /lv01/ //File system capacity used% free used% mount point /dev/mapper/vg01-lv01 77M 776K 73M 2% /lv01 #Found that the space on the file system has not changed [root@xuegod63 ~]# lvextend -L 10M -r /dev/vg01/lv01 #These two commands can't be executed successfully [root@xuegod63 ~]# resize2fs /dev/vg01/lv01 #These two commands can't be executed successfully
Note: to reduce VG, make sure that your physical volume is used, because it cannot reduce a PV in use
[root@xuegod63 ~]# vgs VG #PV #LV #SN Attr VSize VFree vg01 2 2 0 wz--n- 1.99g <1.92g vg02 1 0 0 wz--n- 1008.00m 1008.00m [root@xuegod63 ~]# pvs PV VG Fmt Attr PSize PFree /dev/sdb1 vg01 lvm2 a-- 1020.00m 944.00m /dev/sdb2 vg02 lvm2 a-- 1008.00m 1008.00m /dev/sdb3 vg01 lvm2 a-- 1020.00m 1020.00m /dev/sdb4 lvm2 --- 1.00g 1.00g [root@xuegod63 ~]# cp -r /boot/grub /lv01/ #Copy some test data [root@xuegod63 ~]# vgreduce vg01 /dev/sdb1 #Failed to move sdb1 out because sdb1 is in use Physical volume "/dev/sdb1" still in use
Extension: if sdb1 is a disk array, and this disk array has been used for too long, what should we do if we have to remove it?
Mobile data: [root@xuegod63 ~]# pvmove /dev/sdb1 /dev/sdb3 #Move the data on sdb1 to the newly added sdb3 pv /dev/sdb1: Moved: 23.53% /dev/sdb1: Moved: 76.47% /dev/sdb1: Moved: 100.00% [root@xuegod63 ~]# vgreduce vg01 /dev/sdb1 #Move data out after moving Removed "/dev/sdb1" from volume group "vg01" [root@xuegod63 ~]# pvs PV VG Fmt Attr PSize PFree /dev/sdb1 lvm2 --- 1.00g 1.00g /dev/sdb2 vg02 lvm2 a-- 1008.00m 1008.00m /dev/sdb3 vg01 lvm2 a-- 1020.00m 952.00m #sdb3 is the only one in vg01
2.7 LVM deletion
Create LVM process:
pvcreate create PV - > vgcreate create volume group - > lvcreate create logical volume - > mkfs.xfs LV format - > mount
Delete LVM process:
umount unmount - > lvremove LV move out all logical volumes in volume group - > vgremove VG move out volume group - > pvremove move out pv
[root@xuegod63 ~]# umount /lv01 [root@xuegod63 ~]# lvremove /dev/vg01/lv01 Do you really want to remove active logical volume vg01/lv01? [y/n]: y Logical volume "lv01" successfully removed [root@xuegod63 ~]# lvs LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert lv02 vg01 -wi-a----- 16.00m #lv01 is no longer visible [root@xuegod63 ~]# vgremove vg01 #Move directly out of a volume group Do you really want to remove volume group "vg01" containing 1 logical volumes? [y/n]: y Do you really want to remove active logical volume vg01/lv02? [y/n]: y #If there is lv in the volume group, when you move it out, you will be prompted if you want to move it out. Let's move it out directly here Logical volume "lv02" successfully removed Volume group "vg01" successfully removed [root@xuegod63 ~]# vgs VG #PV #LV #SN Attr VSize VFree vg02 1 0 0 wz--n- 1008.00m 1008.00m #No vg01 //Remove PV sdb1 [root@xuegod63 ~]# pvs PV VG Fmt Attr PSize PFree /dev/sdb1 lvm2 --- 1.00g 1.00g /dev/sdb2 vg02 lvm2 a-- 1008.00m 1008.00m /dev/sdb3 lvm2 --- 1.00g 1.00g /dev/sdb4 lvm2 --- 1.00g 1.00g [root@xuegod63 ~]# pvremove /dev/sdb1 #Moved out Labels on physical volume "/dev/sdb1" successfully wiped. [root@xuegod63 ~]# pvs PV VG Fmt Attr PSize PFree /dev/sdb2 vg02 lvm2 a-- 1008.00m 1008.00m /dev/sdb3 lvm2 --- 1.00g 1.00g /dev/sdb4 lvm2 --- 1.00g 1.00g