Creating LVM storage requires building structures in a logical workflow.

Figure 10.1: Logical Volume Manager workflow

Partitioning is optional when already present. Use the parted command to create a partition on the physical device. Set the physical device to the Linux LVM partition type. Use the udevadm settle command to register the new partition with the kernel.

[root@host ~]# parted /dev/vdb mklabel gpt mkpart primary 1MiB 769MiB
...output omitted...
[root@host ~]# parted /dev/vdb mkpart primary 770MiB 1026MiB
[root@host ~]# parted /dev/vdb set 1 lvm on
[root@host ~]# parted /dev/vdb set 2 lvm on
[root@host ~]# udevadm settle

Use the pvcreate command to label the physical partition as an LVM physical volume. Label multiple devices simultaneously by using space-delimited device names as arguments to the pvcreate command. This example labels the /dev/vdb1 and /dev/vdb2 devices as PVs that are ready for creating volume groups.

[root@host ~]# pvcreate /dev/vdb1 /dev/vdb2
  Physical volume "/dev/vdb1" successfully created.
  Physical volume "/dev/vdb2" successfully created.
  Creating devices file /etc/lvm/devices/system.devices

The vgcreate command builds one or more physical volumes into a volume group. The first argument is a volume group name, followed by one or more physical volumes to allocate to this VG. This example creates the vg01 VG by using the /dev/vdb1 and /dev/vdb2 PVs.

[root@host ~]# vgcreate vg01 /dev/vdb1 /dev/vdb2
  Volume group "vg01" successfully created

The lvcreate command creates a logical volume from the available PEs in a volume group. Use the lvcreate command to set the LV name and size, and the VG name to contain this logical volume. This example creates the lv01 LV with 300 MiB in the vg01 VG.

[root@host ~]# lvcreate -n lv01 -L 300M vg01
  Logical volume "lv01" created.

This command might fail if the volume group does not have enough free physical extents. The LV size rounds up to the next PE size value when the size does not exactly match.

The lvcreate command -L option requires sizes in bytes, mebibytes (binary megabytes, 1048576 bytes), and gibibytes (binary gigabytes), or similar. The lowercase -l requires sizes that are specified as a number of physical extents. The following commands are two choices for creating the same LV with the same size:

RHEL 9 uses an LVM Virtual Data Optimizer (VDO) implementation for managing VDO volumes. The previous Python-based VDO management tools are still available but are no longer needed.

The VDO provides inline block-level deduplication, compression, and thin provisioning for storage. Configure a VDO volume to use up to 256 TB of physical storage. Manage VDO as a type of LVM logical volume (LVs), similar to LVM thinly provisioned volumes. An LVM VDO is composed of two logical volumes:

LVM VDO presents the deduplicated storage as a regular logical volume (LV). The VDO volume can be formatted with standard file systems, or shared as a block device, or used to build other storage layers, the same as any normal logical volume.

To use VDO deduplication and compression, install the vdo and kmod-kvdo packages.

[root@host ~]# dnf install vdo kmod-kvdo

Verify that the selected LVM volume group has enough free storage capacity. Use the lvcreate command with the --type vdo parameter to create a VDO LV.

[root@host ~]# lvcreate --type vdo --name vdo-lv01 --size 5G vg01
    Logical blocks defaulted to 523108 blocks.
    The VDO volume can address 2 GB in 1 data slab.
    It can grow to address at most 16 TB of physical storage in 8192 slabs.
    If a larger maximum size might be needed, use bigger slabs.
  Logical volume "vdo-lv01" created.

Specify the logical volume by using either the /dev/vgname/lvname traditional name, or the /dev/mapper/vgname-lvname kernel device mapper name.

Use the mkfs command to create a file system on the new logical volume.

[root@host ~]# mkfs -t xfs /dev/vg01/vdo-lv01
...output omitted...

Create a mount point by using the mkdir command.

[root@host ~]# mkdir /mnt/data

To make the file system available persistently, add an entry to the /etc/fstab file.

/dev/vg01/vdo-lv01 /mnt/data xfs defaults 0 0

Mount the LV by using the mount command.

[root@host ~]# mount /mnt/data/

The pvdisplay command displays information about the PVs. Use the command without arguments to list information of all PVs on the system. Providing the name of the PV as the argument to the command shows the information that is specific to the PV.

[root@host ~]# pvdisplay /dev/vdb1
  --- Physical volume ---
  PV Name               /dev/vdb1                         
  VG Name               vg01                              
  PV Size               731.98 MiB / not usable 3.98 MiB  
  Allocatable           yes
  PE Size               4.00 MiB                          
  Total PE              182
  Free PE               107                               
  Allocated PE          75
  PV UUID               zP0gD9-NxTV-Qtoi-yfQD-TGpL-0Yj0-wExh2N

The vgdisplay command shows the information about volume groups. To list information about all VGs, use the command without arguments. Provide the name of the VG as an argument to list information that is specific to the VG.

[root@host ~]# vgdisplay vg01
  --- Volume group ---
  VG Name               vg01                
  System ID
  Format                lvm2
  Metadata Areas        2
  Metadata Sequence No  2
  VG Access             read/write
  VG Status             resizable
  MAX LV                0
  Cur LV                1
  Open LV               1
  Max PV                0
  Cur PV                2
  Act PV                2
  VG Size               1012.00 MiB         
  PE Size               4.00 MiB
  Total PE              253                 
  Alloc PE / Size       75 / 300.00 MiB
  Free  PE / Size       178 / 712.00 MiB    
  VG UUID               jK5M1M-Yvlk-kxU2-bxmS-dNjQ-Bs3L-DRlJNc

The lvdisplay command displays information about logical volumes. Use the command without arguments to list information of all LVs. Providing the name of the LV as an argument displays the information that is specific to the LV.

[root@host ~]# lvdisplay /dev/vg01/lv01
  --- Logical volume ---
  LV Path                /dev/vg01/lv01         
  LV Name                lv01
  VG Name                vg01                   
  LV UUID                FVmNel-u25R-dt3p-C5L6-VP2w-QRNP-scqrbq
  LV Write Access        read/write
  LV Creation host, time servera.lab.example.com, 2022-04-07 10:45:34 -0400
  LV Status              available
  # open                 1
  LV Size                300.00 MiB             
  Current LE             75                     
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     8192
  Block device           253:0

You might need to add more disk space to extend a VG. You can add physical volumes to a VG to extend its available size.

Prepare the physical device and create the physical volume when not present.

[root@host ~]# parted /dev/vdb mkpart primary 1072MiB 1648MiB
...output omitted...
[root@host ~]# parted /dev/vdb set 3 lvm on
...output omitted...
[root@host ~]# udevadm settle
[root@host ~]# pvcreate /dev/vdb3
  Physical volume "/dev/vdb3" successfully created.

The vgextend command adds the new PV to the VG. Provide the VG and PV names as arguments to the vgextend command.

[root@host ~]# vgextend vg01 /dev/vdb3
  Volume group "vg01" successfully extended

This command extends the vg01 VG by the /dev/vdb3 PV size.

A benefit of using logical volumes is increasing their size without experiencing any downtime. Add free physical extents to the LV from the VG, to extend the LV capacity and to expand the file system. Use the vgdisplay command to confirm that the volume group has enough free space for the LV extension.

Use the lvextend command to extend the LV.

[root@host ~]# lvextend -L +500M /dev/vg01/lv01
  Size of logical volume vg01/lv01 changed from 300.00 MiB (75 extents) to 800.00 MiB (200 extents).
  Logical volume vg01/lv01 successfully resized.

This command increases the size of the lv01 logical volume by 500 MiB. The (+) plus sign in front of the size means adding this value to the existing size; otherwise, without the plus sign, the value defines the final size of the LV.

The lvextend command -l option expects the number of PE as the argument. The lvextend command -L option expects sizes in bytes, mebibytes, gibibytes, and similar.

The xfs_growfs command helps to expand the file system to occupy the extended LV. The target file system must be mounted before you use the xfs_growfs command. You can continue to use the file system when resizing.

[root@host ~]# xfs_growfs /mnt/data/
...output omitted...
data blocks changed from 76800 to 204800

The steps for extending LV by using the ext4 file system are the same as for LV by using the XFS file system, except for the step to resize the file system.

The resize2fs command expands the file system to occupy the new extended LV. You can continue to use the file system when resizing.

[root@host ~]# resize2fs /dev/vg01/lv01
resize2fs 1.46.5 (30-Dec-2021)
Resizing the filesystem on /dev/vg01/lv01 to 256000 (4k) blocks.
The filesystem on /dev/vg01/lv01 is now 256000 (4k) blocks long.

The xfs_growfs command expands an XFS file system. This command typically takes the mount point as an argument, or otherwise can indirectly use a device name by mounting the device and then specifying the mount point. In contrast, the resize2fs command resizes ext2, ext3, or ext4 file systems and takes the device name as an argument. The xfs_growfs command supports only online resizing, whereas the resize2fs command supports both online and offline resizing. Although you can resize an ext4 file system up or down, you can resize an XFS file system only up.

You can extend the LVs that are used as swap space; however, the process differs from expanding the ext4 or XFS file system. Logical volumes that are used as swap space must be offline to extend them.

Use the swapoff command to deactivate the swap space on the LV.

[root@host ~]# swapoff -v /dev/vg01/swap
swapoff /dev/vg01/swap

Use the lvextend command to extend the LV.

[root@host ~]# lvextend -L +300M /dev/vg01/swap
  Size of logical volume vg01/swap changed from 500.00 MiB (125 extents) to 800.00 MiB (200 extents).
  Logical volume vg01/swap successfully resized.

Use the mkswap command to format the LV as a swap space.

[root@host ~]# mkswap /dev/vg01/swap
mkswap: /dev/vg01/swap: warning: wiping old swap signature.
Setting up swapspace version 1, size = 800 MiB (838856704 bytes)
no label, UUID=25b4d602-6180-4b1c-974e-7f40634ad660

Use the swapon command to activate the swap space on the LV.

[root@host ~]# swapon /dev/vg01/swap

Reducing a VG involves removing unused PVs from the VG. The pvmove command moves data from extents on one PV to extents on another PV with enough free extents in the same VG. You can continue to use the LV from the same VG when reducing. Use the pvmove command -A option to automatically back up the metadata of the VG after a change. This option uses the vgcfgbackup command to back up metadata automatically.

[root@host ~]# pvmove -A y /dev/vdb3

Use the vgreduce command to remove a PV from a VG.

[root@host ~]# vgreduce vg01 /dev/vdb3
  Removed "/dev/vdb3" from volume group "vg01"

Move to another file system all data that must be kept. Use the umount command to unmount the file system, and then remove any /etc/fstab entries that are associated with this file system.

[root@host ~]# umount /mnt/data

Use the lvremove DEVICE-NAME command to remove a logical volume that is no longer needed. Unmount the LV file system before running this command. The command prompts for confirmation before removing the LV.

[root@host ~]# lvremove /dev/vg01/lv01
Do you really want to remove active logical volume vg01/lv01? [y/n]: y
  Logical volume "lv01" successfully removed.

The LV's physical extents are freed and available to assign to existing or new LVs in the volume group.

Use the vgremove VG-NAME command to remove a volume group that is no longer needed.

[root@host ~]# vgremove vg01
  Volume group "vg01" successfully removed

The VG's physical volumes are freed and available to assign to existing or new VGs on the system.

Use the pvremove command to remove physical volumes that are no longer needed. Use a space-delimited list of PV devices to remove more than one device at a time. This command deletes the PV metadata from the partition (or disk). The partition is now available for reallocation or reformatting.

[root@host ~]# pvremove /dev/vdb1 /dev/vdb2
  Labels on physical volume "/dev/vdb1" successfully wiped.
  Labels on physical volume "/dev/vdb2" successfully wiped.