LVM

Introduction

LVM is a Logical Volume Manager. It helps creating storage spaces not restricted by physical storage drives. This flexibility is achieved by a mapping between logical storage units and physical storage units.

File systems are setup on Logical Volumes (LV). Space from a Volume Group (VG) is attributed to a Logical Volume (LV), while Logical Volumes (LV) combine Physical Volumes (PV) together.

LVM can be setup either:

1. Directly on bare storage drives or,
2. on top of a RAID array.

Prepare storage

Using bare drives

Using bare drives provides no data redundancy. The main advantage of this solution is that hard-drives of different sizes can be combined. Using this setup, hard-drives of increased capacity can be added as needed to an existing LVM to increase its storage size.

In this example, two hard-drives identified as /dev/sda and /dev/sdb are used:

Create a partition using 100% of space on each hard-drive.

parted -s /dev/sda mklabel gpt mkpart primary 1MiB 100% set 1 lvm on print
parted -s /dev/sdb mklabel gpt mkpart primary 1MiB 100% set 1 lvm on print

Using RAID

In contrast to bare drives, RAID provides data redundancy. To setup a RAID array, please follow these instructions.

In this example, two hard-drives identified as /dev/sda and /dev/sdb are represented (please note that some RAID levels, like RAID5, require at least 3 drives):

Create new LVM

1. Create the Volume Group VolGroup1 using the two partitions just created. The Physical Volumes are created automatically for each partition using 100% of the space.

   • Using bare devices (see above):

     vgcreate VolGroup1 /dev/sda1 /dev/sdb1
     

   • Using RAID array (see above):

     vgcreate VolGroup1 /dev/md0
     

2. Create Logical Volume (LV) VolData using space in VolGroup1.

   lvcreate -l 100%FREE VolGroup1 -n VolData
   

   This will use 100% of the available space in VolGroup1. To use a limited amount of space, replace -l 100%FREE by -L 500G or -L 10T. For example:

   lvcreate -L 500G VolGroup1 -n VolData
   

   TipMultiple Logical Volumes (LV) can be created until VolGroup1 is full.

3. Format the volume VolData (here using XFS).

   • Using bare devices (see above):

     mkfs.xfs -L data /dev/VolGroup1/VolData
     

   • Using RAID array (see above).

     Formating on top of a RAID array requires proper alignment of the filesystem with the RAID. See commands in the RAID help.

4. Edit /etc/fstab

   • The configuration file /etc/fstab contains the necessary information to automate the process of mounting partitions. Learn more about this file on the Arch Wiki. Add an entry in /etc/fstab the LVM named data:

     LABEL=data   /data   xfs   defaults   0   2
     

   This standard entry can be added for any other LVM volume created. The 1st (LABEL) and 2nd (mount point) columns need to be adapted to the volume to be mounted.

Modify LVM

Add physical space

Using bare drives

In case a new hard-drive (or partition) needs to be added to an existing Volume Group (VG), VG can be extended.

1. Partition the new hard-drive, here /dev/sdc, to create /dev/sdc1 using parted as seen above.
2. Add the new space to VolGroup1 (the Physical Volume is created automatically).

   vgextend VolGroup1 /dev/sdc1
   

3. Eventually, extend the Logical Volume (LV) VolData to use this added space (see below).

Extend volume size

If space is available in a Volume Group (VG), Logical Volume (LV) can be extended and associated filesystem as well. To add space to the Logical Volume VolData:

lvresize -l +100%FREE --resizefs VolGroup1/VolData

To use a limited amount of space, replace -l 100%FREE by -L +500G or -L +10T. For example:

lvresize -L +500G --resizefs VolGroup1/VolData

The option --resizefs will automatically resize the filesystem on VolData.