Red Hat Enterprise Linux 7 Implementation

Logical Volume Management

Module Topics

Managing Logical Volumes
Extending Logical Volumes

Managing Logical Volumes

Implementing LVM Storage

LVM comes with set of command line tools for implementing and managing LVM storage
Tools can be used in scripts, making them suitable for automation

Examples in this module use device vda and vdb and its partitions to illustrate LVM commands. In practice, these examples would use disk and disk partition devices used by system.

Managing Logical Volumes

Creating a Logical Volume

Prepare physical device:
- Use fdisk, gdisk, or parted to create new partition for use with LVM
- Set the partition type:
  - Use Linux LVM on LVM partitions
  - Use 0x8e for MBR-style partitions
- If necessary, use partprobe to register new partition with kernel
  [root@server1 ~]# fdisk /dev/vda

Use m for help, p to print partition table, n to create new partition, t to change partition type, w to write changes, and q to quit.

Transcript:

LVM comes with a comprehensive set of command-line tools for implementing and managing LVM storage. These command-line tools can be used in scripts, making them suitable for automation.

The following examples use device vda and its partitions to illustrate LVM commands. In practice, these examples would need to use the correct devices for the disk and disk partitions that are being used by the system.

There are five steps needed to create a usable logical volume:

1. Prepare the physical device.

Use fdisk, gdisk or parted to create a new partition for use with LVM. Always set the partition type to Linux LVM on LVM partitions; use 0x8e for MBR-style partitions. If necessary, use partprobe to register the new partition with the kernel.

Alternatively, use a whole disk, a RAID array, or a SAN disk.

A physical device only needs to be prepared if there are none prepared already and a new physical volume is required to create or extend a volume group.

[root@server1 ~]# fdisk /dev/vda

Use m for help, p to print the existing partition table, n to create a new partition, t to change the partition type, w to write the changes, and q to quit.

Managing Logical Volumes

Create physical volume (PV):
- Use pvcreate to label partition (or other physical device) for use with LVM as physical volume
  - Writes header directly to PV to store LVM configuration data
  - PV divided into physical extents (PE) of fixed size; for example, 4MiB blocks
- To label multiple devices at same time, add space-delimited device names:
  [root@server1 ~]# pvcreate /dev/vda2 /dev/vdb1

Managing Logical Volumes

Create volume group (VG):
- Use vgcreate to create volume group
  - VG size determined by total number of physical extents in pool
  - VG hosts logical volumes by allocating free PEs to LV
  - LV must have enough free PEs when LV is created
- To define VG name and list one or more PVs to allocate to VG, add arguments
  [root@server1 ~]# vgcreate vg-alpha /dev/vda2 /dev/vdb1

Transcript:

3. Create a volume group.

vgcreate is used to create a pool of one or more physical volumes, called a volume group. The size of the VG is determined by the total number of physical extents in the pool. A VG is responsible for hosting one or more logical volumes by allocating free PEs to a LV; therefore, it must have sufficient free PEs available at the time the LV is created.

As arguments to vgcreate, define a VG name and list one or more PVs to allocate to the VG.

This will create a VG called vg-alpha that is the combined size, in PE units, of the two PVs /dev/vda2 and /dev/vdb1.

A VG only needs to be created when there is none in existence. Additional VGs may be created for administrative reasons to manage the use of PVs and LVs. Otherwise, existing VGs can be extended to accommodate new LVs when needed.

Managing Logical Volumes

Create logical volume:
- Use lvcreate with following options to create new logical volume from available physical extents in volume group:
  - Use -n to set LV name
  - Use -L to set LV size and identify VG that LV is created in
    [root@server1 ~]# lvcreate -n hercules -L 2G vg-alpha

Managing Logical Volumes

There are multiple ways to specify size
- Use -L to specify size in bytes, mebibytes (binary megabytes, 1048576 bytes), or gibibytes (binary gigabytes)
- Use -l to specify size as number of PEs
Examples:
- lvcreate -L 128M: Size logical volume to exactly 128MiB
- lvcreate -l 128: Size logical volume to exactly 128 extents
  - Total number of bytes depends on size of PE block on underlying PV
Tools may display logical volume name using:
- Traditional name /dev/vgname/lvname
- Kernel device mapper name /dev/mapper/vgname-lvname

Managing Logical Volumes

Add file system:
- Use mkfs to create XFS file system on new logical volume:
  [root@server1 ~]# mkfs -t xfs /dev/vg-alpha/hercules
- To make file system available across reboots, use mkdir to create mount point directory:
  [root@server1 ~]# mkdir /mnt/hercules
- Add entry to /etc/fstab:
  /dev/vg-alpha/hercules /mnt/hercules xfs defaults 1 2
- To mount all file systems in /etc/fstab, including the entry just added, run mount -a:
  [root@server1 ~]# mount -a

Managing Logical Volumes

Removing a Logical Volume

Prepare file system:
- Move all data that must be kept to another file system
- Use umount to unmount file system
- Remember to remove /etc/fstab entries associated with the file system
  [root@server1 ~]# umount /mnt/hercules
  Removing a logical volume destroys any data stored on the logical volume. Back up or move your data before you remove the logical volume.

Managing Logical Volumes

Remove logical volume:
- Use lvremove with device name as argument
  [root@server1 ~]# lvremove /dev/vg-alpha/hercules
- LV file system must be unmounted before running command
- lvremove asks for confirmation
- LV’s physical extents are now available for assignment to other LVs in volume group

Managing Logical Volumes

Remove volume group:
- Use vgremove with VG name as argument
  [root@server1 ~]# vgremove vg-alpha
- VG’s physical volumes are now available for assignment to VGs.
Remove physical volumes:
- Use pvremove to remove physical volumes
- Add space-delimited list of PV devices to remove more than one device
  [root@server1 ~]# pvremove /dev/vda2 /dev/vdb1
- PV metadata is wiped from partition or disk
- Partition is free for reallocation or reformatting

Managing Logical Volumes

Reviewing Physical Volume Information

To display information about physical volumes, use pvdisplay
If no arguments are specified, pvdisplay lists information for all PVs

If device is specified as argument, only that PV is shown

   [root@server1 ~]# pvdisplay /dev/vda2
     --- Physical volume ---
<1>  PV Name               /dev/vda2
<2>  VG Name               vg-alpha
<3>  PV Size               256.00 MiB / not usable 4.00 MiB
     Allocatable           yes
<4>  PE Size               4.00 MiB
     Total PE              63
<5>  Free PE               26
     Allocated PE          37
     PV UUID               JWzDpn-LG3e-n2oi-9Etd-VT2H-PMem-1ZXwP1

Transcript:

Use pvdisplay to display information about physical volumes. If no arguments are specified with the command, it will list information about all PVs on the system. If the argument is a specific device name, then display information will be limited to that specific PV.

Number 1, PV Name, maps to the device name.
Number 2,VG Name, shows the volume group where the PV is allocated.
Number 3,PV Size, shows the physical size of the PV, including any unusable space.
Number 4,PE Size, is the physical extent size, which is the smallest size a logical volume can be allocated.
It is also the multiplying factor when calculating the size of any value reported in PE units, such as Free PE; for example:26 PEs x 4MiB (the PE Size) gives 104MiB of free space. A logical volume size will be rounded to a factor of PE units.

LVM sets the PE size automatically, although it is possible to specify it.
Number 5, Free PE, shows how many PE units are available for allocation to new logical volumes.

Managing Logical Volumes

Reviewing Volume Group Information

To display information about volume groups, use vgdisplay
If no arguments are specified, vgdisplay shows all VGs

If VG name is specified, only that VG is shown

[root@server1 ~]# vgdisplay vg-alpha
     --- Volume group ---
     VG Name               vg-alpha
<1>  System ID
     Format                lvm2
     Metadata Areas        3
     Metadata Sequence No  4
     VG Access             read/write
     VG Status             resizable
     MAX LV                0
     Cur LV                1
     Open LV               1
     Max PV                0
     Cur PV                3
     Act PV                3
<2>  VG Size               1012.00 MiB
     PE Size               4.00 MiB
<3>  Total PE              253
     Alloc PE / Size       175 / 700.00 MiB
<4>  Free  PE / Size       78 / 312.00 MiB
     VG UUID               3snNw3-CF71-CcYG-Llk1-p6EY-rHEv-xfUSez

Managing Logical Volumes

Reviewing Logical Volume Information

To display information about logical volumes, use lvdisplay
If no arguments are specified, lvdisplay shows all LVs

If LV name is specified, only that LV is shown

   [root@server1 ~]# lvdisplay /dev/vg-alpha/hercules
     --- Logical volume ---
<1>  LV Path                /dev/vg-alpha/hercules
     LV Name                hercules
<2>  VG Name                vg-alpha
     LV UUID                5IyRea-W8Zw-xLHk-3h2a-IuVN-YaeZ-i3IRrN
     LV Write Access        read/write
     LV Creation host, time server1.example.com 2014-02-19 00:6:48 -0500
     LV Status              available
     # open                 1
<3>  LV Size                700 MiB
<4>  Current LE             175
     Segments               3
     Allocation             inherit
     Read ahead sectors     auto
     - current set to       8192
     Block device           252:

References

Man pages: lvm(8), pvcreate(8), vgcreate(8), lvcreate(8), pvremove(8), vgremove(8), lvremove(8), pvdisplay(8), vgdisplay(8), lvdisplay(8), fdisk(8), gdisk(8), parted(8), partprobe(8), and mkfs(8)

Extending Logical Volumes

Extending and Reducing Volume Groups

Add physical volumes to volume group to increase its disk space
- Called extending volume group
Can assign new physical extents to logical volumes
Can remove unused physical volumes from volume group
- Called reducing volume group
Use pvmove to move data from extents on one PV to extents on other PVs in volume group
- Add new disk to existing volume group
- Move data from older or slower disk to new disk
- Remove old disk from volume group
Can be done while logical volumes in volume group are in use

Extending Logical Volumes

Extending a Volume Group

Prepare physical device:
- Use fdisk, gdisk, or parted to create new partition for use with LVM
- Set partition type:
  - Use Linux LVM on LVM partitions
  - Use 0x8e for MBR-style partitions
- If necessary, use partprobe to register new partition with kernel
  [root@server1 ~]# fdisk /dev/vdb

Use m for help, p to print partition table, n to create new partition, t to change partition type, w to write changes, and q to quit.

Transcript:

There are potentially four steps needed to extend a volume group:

1. Prepare the physical device.

As with creating a new volume group, a new partition must be created and prepared for use as an LVM physical volume.

Use fdisk, gdisk, or parted to create a new partition for use with LVM. Always set the partition type to Linux LVM on LVM partitions; use 0x8e for MBR-style partitions. If necessary, use partprobe to register the new partition with the kernel.

Alternatively, use a whole disk, a RAID array, or a SAN disk.

A physical device only needs to be prepared if there are none prepared already and a new physical volume is required to extend the volume group.

Use m for help, p to print the existing partition table, n to create a new partition, t to change the partition type, w to write the changes, and q to quit.

Extending Logical Volumes

Create physical volume:
- Use pvcreate to label partition (or other physical device) for use with LVM as physical volume
  - Writes header directly to PV to store LVM configuration data
  - PV is divided into PEs of fixed size; for example, 4MiB blocks
- To label multiple devices at same time, add space-delimited device names
  [root@server1 ~]# pvcreate /dev/vdb2

Extending Logical Volumes

Extend volume group:
- Use vgextend to add new PV to volume group
- Use VG name and PV device name as arguments:
  [root@server1 ~]# vgextend vg-alpha /dev/vdb2

Verify new space is available:

Use vgdisplay to confirm additional PEs are available

Check Free PE/Size in output

It should be greater than zero

[root@server1 ~]# vgdisplay vg-alpha
  --- Volume group ---
  VG Name               vg-alpha
...
  Free  PE / Size       178 / 712.00 MiB
...

Extending Logical Volumes

Reducing a Volume Group

Move physical extents:
- Use pvmove to relocate physical extents used on physical volume to other PVs in VG.
  - Requires enough free extents in VG and all extents must come from other PVs
- Use PV device name from which PEs will be moved as argument
  [root@server1 ~]# pvmove /dev/vdb2

Before using pvmove, back up data stored on all logical volumes in volume group. An unexpected power loss during the operation may leave the volume group in an inconsistent state. This could cause loss of data on logical volumes in the volume group.

Extending Logical Volumes

Reduce volume group:
- Use vgreduce to remove PV from volume group
- Use VG name and PV device name as arguments:
  [root@server1 ~]# vgreduce vg-alpha /dev/vdb2
- To permanently stop using device as PV, use pvremove

Extending Logical Volumes

Extending a Logical Volume and XFS File System

Can increase LV size without downtime
Can add free PEs in VG to LV to extend capacity
- Use to extend file system it contains
  1. Verify that volume group has space available:
- Use vgdisplay to verify that sufficient PEs are available
  [root@server1 ~]# vgdisplay vg-alpha --- Volume group --- VG Name vg-alpha ... Free PE / Size 178 / 712.00 MiB ...
- Check Free PE/Size for value equal to or greater than space required
  - If space is insufficient, extend volume group

Extending Logical Volumes

Extend logical volume:
- Use lvextend to extend logical volume size
- Add LV device name as last argument
  [root@server1 ~]# lvextend -L +300M /dev/vg-alpha/hercules

Extending Logical Volumes

There are multiple ways to specify new size:
- Use -l to specify physical extent values
- Use -L to specify size in bytes, mebibytes, and gibibytes
Examples:
- lvextend -l 128: Resize logical volume to exactly 128 extents
- lvextend -l +128: Add 128 extents to current size of logical volume
- lvextend -L 128M: Resize logical volume to exactly 128MiB
- lvextend -L +128M: Add 128MiB to current size of logical volume
- lvextend -l +50%FREE: Add 50 percent of current VG free space to LV

Extending Logical Volumes

Extend file system:
- Use xfs_growfs /mountpoint to expand file system to occupy extended LV
  - Requires file system be mounted while being run
  - Can continue to use file system while resizing
    [root@server1 ~]# xfs_growfs /mnt/hercules
- Can also use lvextend -r to resize file system after LV is extended
  - Uses fsadm
  - Works with several different file systems
- Verify new size of mounted file system
  df -h /mountpoint

Extending Logical Volumes

Extending a Logical Volume and ext4 File System

Verify volume group has space available:
- Use vgdisplay vgname to verify enough PEs are available
Extend logical volume:
- Use lvextend -l +extents/dev/vgname/lvname to extend logical volume /dev/vgname/lvname by +extents.

Extending Logical Volumes

Extend file system
- Use resize2fs /dev/vgname/lvname to expand file system to occupy extended LV
  - Can mount file system use it while extending
  - To see progress of resize operation, use -p
    [root@server1 ~]# resize2fs /dev/vg-alpha/hercules

References

Man pages: lvm(8), pvcreate(8), pvmove(8), vgdisplay(8), vgextend(8), vgreduce(8), vgdisplay(8), vgextend(8), vgreduce(8), lvextend(8), fdisk(8), gdisk(8), parted(8), partprobe(8), xfs_growfs(8), and resize2fs(8)

Summary

Managing Logical Volumes
Extending Logical Volumes

Module Completion

Nice job!

Click the button below to complete this module of the course: