Introduction

LVM, or Logical Volume Management, is a storage device management technology that gives users the power to pool and abstract the physical layout of component storage devices for easier and flexible administration. Utilizing the device mapper Linux kernel framework, the current iteration, LVM2, can be used to gather existing storage devices into groups and allocate logical units from the combined space as needed.

In this guide, we will cover how to manage your storage devices with LVM. We will show you how to display information about volumes and potential targets, how to create and destroy volumes of various types, and how to modify existing volumes through resizing or transformation. We will be using an Ubuntu 16.04 server to demonstrate these operations.

Prerequisites

In order to follow along, you should have access to an Ubuntu 16.04 server. You will need to have a non-root user with sudo privileges configured for administrative tasks. You can follow our Ubuntu 16.04 initial server setup guide to create the necessary account.

To get familiar with LVM components and concepts and to test out a basic LVM configuration, follow our introduction to LVM guide prior to starting on this tutorial.

When you are ready, log into your server with your sudo user.

Display Information about Physical Volumes, Volume Groups, and Logical Volumes

It is important to be able to get information about the various LVM components in your system easily. Fortunately, the LVM tool suite provides an abundant amount of tools for displaying information about every layer in the LVM stack.

Displaying Information about All LVM Compatible Block Storage Devices

To display all of the available block storage devices that LVM can potentially manage, use the lvmdiskscan command:

sudo lvmdiskscan

Output
  /dev/ram0  [      64.00 MiB] 
  /dev/sda   [     200.00 GiB] 
  /dev/ram1  [      64.00 MiB] 

  . . .

  /dev/ram15 [      64.00 MiB] 
  /dev/sdb   [     100.00 GiB] 
  2 disks
  17 partitions
  0 LVM physical volume whole disks
  0 LVM physical volumes

Ignoring the /dev/ram* devices (part of the Linux ram disk implementation), we can see the devices that can potentially be used as physical volumes for LVM.

This will likely be your first step when adding new storage devices to use with LVM.

Displaying Information about Physical Volumes

A header is written to storage devices to mark them as free to use as LVM components. Devices with these headers are called physical volumes.

You can display all of the physical devices on your system by using lvmdiskscan with the -l option, which will only return physical volumes:

sudo lvmdiskscan -l

Output
  WARNING: only considering LVM devices
  /dev/sda                   [     200.00 GiB] LVM physical volume
  /dev/sdb                   [     100.00 GiB] LVM physical volume
  2 LVM physical volume whole disks
  0 LVM physical volumes

The pvscan command is fairly similar to the above, in that it searches all available devices for LVM physical volumes. The output format is a bit different and it include a small amount of additional information:

sudo pvscan

Output
  PV /dev/sda   VG LVMVolGroup     lvm2 [200.00 GiB / 0    free]
  PV /dev/sdb   VG LVMVolGroup     lvm2 [100.00 GiB / 10.00 GiB free]
  Total: 2 [299.99 GiB] / in use: 2 [299.99 GiB] / in no VG: 0 [0   ]

If you need more detail, the pvs and pvdisplay commands are better options.

The pvs command is highly configurable and can display information in many different formats. Because its output can be tightly controlled, it is frequently used in when scripting or automation is needed. Its basic output provides a useful at-a-glance summary similar to the earlier commands:

sudo pvs

Output
  PV         VG          Fmt  Attr PSize   PFree 
  /dev/sda   LVMVolGroup lvm2 a--  200.00g     0 
  /dev/sdb   LVMVolGroup lvm2 a--  100.00g 10.00g

For more verbose, human-readable output, the pvdisplay command is usually a better option:

sudo pvdisplay

Output
  --- Physical volume ---
  PV Name               /dev/sda
  VG Name               LVMVolGroup
  PV Size               200.00 GiB / not usable 4.00 MiB
  Allocatable           yes (but full)
  PE Size               4.00 MiB
  Total PE              51199
  Free PE               0
  Allocated PE          51199
  PV UUID               kRUOyU-0ib4-ujPh-kAJP-eeQv-ztRL-4EkaDQ

  --- Physical volume ---
  PV Name               /dev/sdb
  VG Name               LVMVolGroup
  PV Size               100.00 GiB / not usable 4.00 MiB
  Allocatable           yes 
  PE Size               4.00 MiB
  Total PE              25599
  Free PE               2560
  Allocated PE          23039
  PV UUID               udcuRJ-jCDC-26nD-ro9u-QQNd-D6VL-GEIlD7

As you can see the pvdisplay command is often the easiest command for getting detailed information about physical volumes.

To discover the logical extents that have been mapped to each volume, pass in the -m option to pvdisplay:

sudo pvdisplay -m

Output
  --- Physical volume ---
  PV Name               /dev/sda
  VG Name               LVMVolGroup
  PV Size               200.00 GiB / not usable 4.00 MiB
  Allocatable           yes 
  PE Size               4.00 MiB
  Total PE              51199
  Free PE               38395
  Allocated PE          12804
  PV UUID               kRUOyU-0ib4-ujPh-kAJP-eeQv-ztRL-4EkaDQ

  --- Physical Segments ---
  Physical extent 0 to 0:
    Logical volume  /dev/LVMVolGroup/db_rmeta_0
    Logical extents 0 to 0
  Physical extent 1 to 5120:
    Logical volume  /dev/LVMVolGroup/db_rimage_0
    Logical extents 0 to 5119

   . . .

This can be very useful when trying to determine which data is held on which physical disk for management purposes.

Displaying Information about Volume Groups

LVM also has plenty of tools to display information about volume groups.

The vgscan command can be used to scan the system for available volume groups. It also rebuilds the cache file when necessary. It is a good command to use when you are importing a volume group into a new system:

sudo vgscan

Output
  Reading all physical volumes.  This may take a while...
  Found volume group "LVMVolGroup" using metadata type lvm2

The command does not output very much information, but it should be able to find every available volume group on the system. To display more information, the vgs and vgdisplay commands are available.

Like its physical volume counterpart, the vgs command is versatile and can display a large amount of information in a variety of formats. Because its output can be manipulated easily, it is frequently used in when scripting or automation is needed. For example, some helpful output modifications are to show the physical devices and the logical volume path:

sudo vgs -o +devices,lv_path

Output
  VG          #PV #LV #SN Attr   VSize   VFree  Devices        Path                      
  LVMVolGroup   2   4   0 wz--n- 299.99g 10.00g /dev/sda(0)    /dev/LVMVolGroup/projects 
  LVMVolGroup   2   4   0 wz--n- 299.99g 10.00g /dev/sda(2560) /dev/LVMVolGroup/www      
  LVMVolGroup   2   4   0 wz--n- 299.99g 10.00g /dev/sda(3840) /dev/LVMVolGroup/db       
  LVMVolGroup   2   4   0 wz--n- 299.99g 10.00g /dev/sda(8960) /dev/LVMVolGroup/workspace
  LVMVolGroup   2   4   0 wz--n- 299.99g 10.00g /dev/sdb(0)    /dev/LVMVolGroup/workspace

For more verbose, human-readable output, the vgdisplay command is a usually the best choice. Adding the -v flag also provides information about the physical volumes the volume group is built upon, and the logical volumes that were created using the volume group:

sudo vgdisplay -v

Output
    Using volume group(s) on command line.
  --- Volume group ---
  VG Name               LVMVolGroup
  . . .

  --- Logical volume ---
  LV Path                /dev/LVMVolGroup/projects
  . . .

  --- Logical volume ---
  LV Path                /dev/LVMVolGroup/www
  . . .

  --- Logical volume ---
  LV Path                /dev/LVMVolGroup/db
  . . .

  --- Logical volume ---
  LV Path                /dev/LVMVolGroup/workspace
  . . .

  --- Physical volumes ---
  PV Name               /dev/sda     
  . . .

  PV Name               /dev/sdb     
  . . .

The vgdisplay command is useful because it can tie together information about many different elements of the LVM stack.

Displaying Information about Logical Volumes

To display information about logical volumes, LVM has a related set of tools.

As with the other LVM components, the lvscan option scans the system and outputs minimal information about the logical volumes it finds:

sudo lvscan

Output
  ACTIVE            '/dev/LVMVolGroup/projects' [10.00 GiB] inherit
  ACTIVE            '/dev/LVMVolGroup/www' [5.00 GiB] inherit
  ACTIVE            '/dev/LVMVolGroup/db' [20.00 GiB] inherit
  ACTIVE            '/dev/LVMVolGroup/workspace' [254.99 GiB] inherit

For more complete information, the lvs command is flexible, powerful, and easy to use in in scripts:

sudo lvs

Output
  LV        VG          Attr       LSize   Pool Origin Data%  Meta%  Move Log Cpy%Sync Convert
  db        LVMVolGroup -wi-ao----  20.00g                                                    
  projects  LVMVolGroup -wi-ao----  10.00g                                                    
  workspace LVMVolGroup -wi-ao---- 254.99g                                                    
  www       LVMVolGroup -wi-ao----   5.00g

To find out about the number of stripes and the logical volume type, use the --segments option:

sudo lvs --segments

Output
  LV           VG          Attr       #Str Type    SSize 
  db           LVMVolGroup rwi-a-r---    2 raid1   20.00g
  mirrored_vol LVMVolGroup rwi-a-r---    3 raid1   10.00g
  test         LVMVolGroup rwi-a-r---    3 raid5   10.00g
  test2        LVMVolGroup -wi-a-----    2 striped 10.00g
  test3        LVMVolGroup rwi-a-r---    2 raid1   10.00g

The most human-readable output is produced by the lvdisplay command.

When the -m flag is added, the tool will also display information about how the logical volume is broken down and distributed:

sudo lvdisplay -m

Output
  --- Logical volume ---
  LV Path                /dev/LVMVolGroup/projects
  LV Name                projects
  VG Name                LVMVolGroup
  LV UUID                IN4GZm-ePJU-zAAn-DRO3-1f2w-qSN8-ahisNK
  LV Write Access        read/write
  LV Creation host, time lvmtest, 2016-09-09 21:00:03 +0000
  LV Status              available
  # open                 1
  LV Size                10.00 GiB
  Current LE             2560
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     256
  Block device           252:0

  --- Segments ---
  Logical extents 0 to 2559:
    Type        linear
    Physical volume /dev/sda
    Physical extents    0 to 2559


  . . .

As you can see from the output towards the bottom, the /dev/LVMVolGroup/projects logical volume is contained entirely within the /dev/sda physical volume in this example. This information is useful if you need to remove that underlying device and wish to move the data off to specific locations.

Create or Extend LVM Components

This section will discuss how to create and expand physical volumes, volume groups, and logical volumes.

Creating Physical Volumes From Raw Storage Devices

In order to use storage devices with LVM, they must first be marked as a physical volume. This specifies that LVM can use the device within a volume group.

First, use the lvmdiskscan command to find all block devices that LVM can see and use:

sudo lvmdiskscan

Output
  /dev/ram0  [      64.00 MiB] 
  /dev/sda   [     200.00 GiB] 
  /dev/ram1  [      64.00 MiB] 

  . . .

  /dev/ram15 [      64.00 MiB] 
  /dev/sdb   [     100.00 GiB] 
  2 disks
  17 partitions
  0 LVM physical volume whole disks
  0 LVM physical volumes

Here, ignoring the /dev/ram* devices, we can see the devices that are suitable to be turned in physical volumes for LVM.

Warning: Make sure that you double-check that the devices you intend to use with LVM do not have any important data already written to them. Using these devices within LVM will overwrite the current contents. If you already have important data on your server, make backups before proceeding.

To mark the storage devices as LVM physical volumes, use pvcreate. You can pass in multiple devices at once:

sudo pvcreate /dev/sda /dev/sdb

This should write an LVM header on all of the target devices to mark them as LVM physical volumes.

Creating a New Volume Group from Physical Volumes

To create a new volume group from LVM physical volumes, use the vgcreate command. You will have to provide a volume group name, followed by at least one LVM physical volume:

sudo vgcreate volume_group_name /dev/sda

This example will create your volume group with a single initial physical volume. You can pass in more than one physical volume at creation if you'd like:

sudo vgcreate volume_group_name /dev/sda /dev/sdb /dev/sdc

Usually you will only need a single volume group per server. All LVM-managed storage can be added to that pool and then logical volumes can be allocated from that.

One reason you may wish to have more than one volume group is if you feel you need to use different extent sizes for different volumes. Usually you will not have to set the extent size (the default size of 4M is adequate for most uses), but if you need to, you can do so upon volume group creation by passing the -s option:

suod vgcreate -s 8M volume_group_name /dev/sda

This will create a new volume group with an 8M extent size.

Adding a Physical Volume to an Existing Volume Group

To expand a volume group by adding additional physical volumes, use the vgextend command. This command takes a volume group followed by the physical volumes to add. You can pass in multiple devices at once if you'd like:

sudo vgextend volume_group_name /dev/sdb

The physical volume will be added to the volume group, expanding the available capacity of the storage pool.

Creating a Logical Volume by Specifying Size

To a logical volume from a volume group storage pool, use the lvcreate command. Specify the size of the logical volume with the -L option, specify a name with the -n option, and pass in the volume group to allocate the space from.

For instance, to create a 10G logical volume named test from the LVMVolGroup volume group, type:

sudo lvcreate -L 10G -n test LVMVolGroup

Provided that the volume group has enough free space to accommodate the volume capacity, the new logical volume will be created.

Creating a Logical Volume From All Remaining Free Space

If you wish to create a volume using the remaining free space within a volume group, use the vgcreate command with the -n option to name and pass in the volume group as before. Instead of passing in a size, use the -l 100%FREE option, which uses the remaining extents within the volume group to form the logical volume:

sudo lvcreate -l 100%FREE -n test2 LVMVolGroup

This should use up the remaining space in the logical volume.

Creating Logical volumes with Advanced Options

Logical volumes can be created with some advanced options as well. Some options that you may wish to consider are:

• --type: This specifies the type of logical volume, which determines how the logical volume is allocated. Some of the available types will not be available if there are not enough underlying physical volumes to correctly create the chosen topography. Some of the most common types are:
  • linear: The default type. The underlying physical devices used (if more than one) will simply be appended to each other, one after the other.
  • striped: Similar to RAID 0, the striped topology divides data into chunks and spread in a round-robin fashion across the underlying physical volumes. This can lead to performance improvements, but might lead to greater data vulnerability. This requires the -i option described below and a minimum of two physical volumes.
  • raid1: Creates a mirrored RAID 1 volume. By default, the mirror will have two copies, but more can be specified by the -m option described below. Requires a minimum of two physical volumes.
  • raid5: Creates a RAID 5 volume. Requires a minimum of three physical volumes.
  • raid6: Creates a RAID 6 volume. Requires a minimum of four physical volumes.
• -m: Specifies the number of additional copies of data to keep. A value of "1" specifies that one additional copy is maintained, for a total of two sets of data.
• -i: Specifies the number of stripes that should be maintained. This is required for the striped type, and can modify the default behavior of some of the other RAID options.
• -s: Specifies that the action should create a snapshot from an existing logical volume instead of a new independent logical volume.

We'll provide a few examples of these options to demonstrate how they'd typically be used.

To create a striped volume, you must specify at least two stripes. This topology and stripe count requires a minimum of two physical volumes with available capacity:

sudo lvcreate --type striped -i 2 -L 10G -n striped_vol LVMVolGroup

To create a mirrored volume, use the raid1 type. If you want more than two sets of data, use the -m option. This example uses -m 2 to create a total of three sets of data (LVM counts this as one original data set with two mirrors). You will need at least three physical volumes for this to succeed:

sudo lvcreate --type raid1 -m 2 -L 20G -n mirrored_vol LVMVolGroup

To create a snapshot of a volume, you must provide the original logical volume to snapshot instead of the volume group. Snapshots do not take up much space initially, but grow in size as changes are made to the logical volume it is tracking. The size used during this procedure is the maximum size that the snapshot can be (snapshots that grow past this size are broken and cannot be used; snapshots approaching their capacity can be extended however):

sudo lvcreate -s -L 10G -n snap_test LVMVolGroup/test

sudo lvconvert --merge LVMVolGroup/snap_test

As you can see, there are a number of options that can dramatically alter the way that your logical volumes function.

Growing the Size of a Logical Volume

One of the main advantages of LVM is the flexibility it provides in provisioning logical volumes. You can easily adjust the number or size of volumes on the fly without stopping the system.

To grow the size of an existing logical volume, use the lvresize command. Use the -L flag to specify a new size. You can also use relative sizes by adding a "+" size. In that case, LVM will increase the size of the logical volume by the amount specified. To automatically resize the filesystem being used on the logical volume as well, pass in the --resizefs flag.

To correctly provide the name of the logical volume to expand, you'll need to give the volume group, followed by a slash, followed by the logical volume:

sudo lvresize -L +5G --resizefs LVMVolGroup/test

In this example, the logical volume and the filesystem of the test logical volume on the LVMVolGroup volume group will both be increased by 5G.

If you wish to handle the filesystem expansion manually, you can take out the --resizefs option and use the filesystem's native expansion utility afterwards. For example, for an Ext4 filesystem, you could type:

sudo lvresize -L +5G LVMVolGroup/test
sudo resize2fs /dev/LVMVolGroup/test

This would leave you with the same result.

Remove or Downsize LVM Components

Since capacity reduction can result in data loss, the procedures to shrink the available capacity, either by reducing the size of or removing components are typically a bit more involved.

Reducing the Size of a Logical Volume

To shrink a logical volume, you should first back up your data. Because this reduces the available capacity, mistakes can lead to data loss.

When you are ready, check on how much space is currently being used:

df -h

Output
Filesystem                         Size  Used Avail Use% Mounted on
. . .
/dev/mapper/LVMVolGroup-test       4.8G  521M  4.1G  12% /mnt/test

In this example, it looks like a little over 521M of the space is currently in use. Use this to help you estimate the size that you can reduce the volume to.

Next, unmount the filesystem. Unlike expansions, filesystem shrinking should be performed when unmounted:

cd ~
sudo umount /dev/LVMVolGroup/test

After unmounting, check the filesystem to ensure that everything is in working order. Pass in the filesystem type with the -t option. We'll use -f to check even when the filesystem appears okay:

sudo fsck -t ext4 -f /dev/LVMVolGroup/test

After checking the filesystem, you can reduce the filesystem size using the filesystem's native tools. For Ext4 filesystems, this would be the resize2fs command. Pass in the final size for the filesystem:

Warning: The safest option here is to choose a final size that is a fair amount larger than your current usage. Give yourself some buffer room to avoid data loss and ensure that you have backups in place.

sudo resize2fs -p /dev/LVMVolGroup/test 3G

Once the operation is complete, resize the logical volume by passing the same size to the lvresize command with the -L flag:

sudo lvresize -L 3G LVMVolGroup/test

You will be warned about the possibility of data loss. If you are ready, type y to proceed.

After the logical volume has been reduced, check the filesystem again:

sudo fsck -t ext4 -f /dev/LVMVolGroup/test

If everything is functioning correctly, you can remount the filesystem using your usual mount command:

sudo mount /dev/LVMVolGroup/test /mnt/test

Your logical volume should now be reduced to the appropriate size.

Removing a Logical Volume

If you no longer need a logical volume, you can remove it with the lvremove command.

First, unmount the logical volume if it is currently mounted:

cd ~
sudo umount /dev/LVMVolGroup/test

Afterwards, remove the logical volume by typing:

sudo lvremove LVMVolGroup/test

You will be asked to confirm the procedure. If you are certain you want to delete the logical volume, type y.

Removing a Volume Group

To remove an entire volume group, including all of the logical volumes within it, use the vgremove command.

Before you remove a volume group, you should usually remove the logical volumes using the procedure above. At the very least, you must make sure that you unmount any logical volumes that the volume group contains:

sudo umount /dev/LVMVolGroup/www
sudo umount /dev/LVMVolGroup/projects
sudo umount /dev/LVMVolGroup/db

Afterwards, you can delete the entire volume group by passing the volume group name to the vgremove command:

sudo vgremove LVMVolGroup

You will be prompted to confirm that you wish to remove the volume group. If you have any logical volumes still present, you will be given individual confirmation prompts for those before removing.

Removing a Physical Volume

If you wish to remove a physical volume from LVM management, the procedure you will need depends on whether the device is currently being used by LVM.

If the physical volume is in use, you will have to move the physical extents located on the device to a different location. This requires the volume group to have enough other physical volumes to handle the physical extents. If you are using more complex logical volume types, you might have to have additional physical volumes even when you have plenty of free space in order to accommodate the topology.

When you have enough physical volumes in the volume group to handle the physical extents, move them off of the physical volume you wish to remove by typing:

sudo pvmove /dev/sda

This process can take awhile depending on the size of the volumes and the amount of data to transfer.

Once the extents have been relocated to peer volumes, you can remove the physical volume from the volume group by typing:

sudo vgreduce LVMVolGroup /dev/sda

This will remove the vacated physical volume from the volume group. After this is complete, you can remove the physical volume marker from the storage device by typing:

sudo pvremove /dev/sda

You should now be able to use the removed storage device for other purposes or remove it from the system entirely.

Conclusion

By now, you should have a working understanding of how to manage storage devices on Ubuntu 16.04 with LVM. You should know how to get information about the state of existing LVM components, how to use LVM to compose your storage system, and how to modify volumes to meet your needs. Feel free to test these concepts in a safe environment to get a better grasp of how they fit together.