💾 LVM Basics

Before a disk can join an LVM volume group it must be initialized as a Physical Volume:

# First — identify available disks
lsblk
fdisk -l | grep "^Disk /dev"

# Initialize a whole disk as a PV (most common — no partitioning needed)
sudo pvcreate /dev/sdb
sudo pvcreate /dev/sdc
sudo pvcreate /dev/sdd

# Initialize multiple PVs at once
sudo pvcreate /dev/sd{b,c,d}

# Initialize a partition (rather than a whole disk)
sudo pvcreate /dev/sdb1

# Verify
sudo pvs
sudo pvdisplay /dev/sdb

A Volume Group pools one or more PVs into a single storage resource:

# Create a VG named vg_data from three PVs
sudo vgcreate vg_data /dev/sdb /dev/sdc /dev/sdd

# Create with a specific physical extent size (default is 4MB)
sudo vgcreate -s 16M vg_data /dev/sdb /dev/sdc

# Verify
sudo vgs
sudo vgdisplay vg_data

Logical Volumes are carved from the VG pool — each becomes a block device you can format and mount:

# Create an LV of a specific size
sudo lvcreate -L 100G -n lv_home vg_data
sudo lvcreate -L 200G -n lv_var  vg_data
sudo lvcreate -L 50G  -n lv_opt  vg_data

# Create an LV using a percentage of free space
sudo lvcreate -l 50%FREE -n lv_scratch vg_data

# Create an LV using ALL remaining free space
sudo lvcreate -l 100%FREE -n lv_data vg_data

# Verify
sudo lvs
sudo lvdisplay vg_data/lv_home

An LV is just a block device — format it and mount it like any partition:

# Format with ext4
sudo mkfs.ext4 /dev/vg_data/lv_home
sudo mkfs.ext4 /dev/vg_data/lv_var

# Format with XFS (default on RHEL 8/9)
sudo mkfs.xfs /dev/vg_data/lv_opt

# Create mount points and mount
sudo mkdir -p /home /var/opt
sudo mount /dev/vg_data/lv_home /home
sudo mount /dev/vg_data/lv_var  /var
sudo mount /dev/vg_data/lv_opt  /opt

# Add to /etc/fstab for persistent mounts
echo "/dev/vg_data/lv_home  /home  ext4  defaults  0 2" | sudo tee -a /etc/fstab
echo "/dev/vg_data/lv_var   /var   ext4  defaults  0 2" | sudo tee -a /etc/fstab
echo "/dev/vg_data/lv_opt   /opt   xfs   defaults  0 2" | sudo tee -a /etc/fstab

# Verify mounts
df -h | grep vg_data
lsblk -f

The short display commands give quick status — use them constantly:

# Summary views — most useful day-to-day
sudo pvs    # all physical volumes
sudo vgs    # all volume groups
sudo lvs    # all logical volumes

# Verbose detail
sudo pvdisplay
sudo vgdisplay
sudo lvdisplay

# Specific object
sudo pvdisplay /dev/sdb
sudo vgdisplay vg_data
sudo lvdisplay /dev/vg_data/lv_home

# Extended output with segments
sudo pvs -v
sudo lvs -o +devices

# Full picture — lsblk shows the tree
lsblk
lsblk -f   # includes filesystem type and UUID

The full workflow — new disks to mounted filesystems in one sequence:

#!/bin/bash
# New LVM setup — two 500GB disks → one VG → two LVs

DISK1=/dev/sdb
DISK2=/dev/sdc
VG=vg_apps
LV1=lv_appdata
LV2=lv_applogs

# Step 1: Initialize PVs
sudo pvcreate $DISK1 $DISK2

# Step 2: Create VG
sudo vgcreate $VG $DISK1 $DISK2

# Step 3: Create LVs
sudo lvcreate -L 800G -n $LV1 $VG
sudo lvcreate -l 100%FREE -n $LV2 $VG

# Step 4: Format
sudo mkfs.xfs /dev/$VG/$LV1
sudo mkfs.xfs /dev/$VG/$LV2

# Step 5: Mount
sudo mkdir -p /opt/appdata /var/log/apps
sudo mount /dev/$VG/$LV1 /opt/appdata
sudo mount /dev/$VG/$LV2 /var/log/apps

# Step 6: fstab
echo "/dev/$VG/$LV1  /opt/appdata    xfs  defaults  0 2" | sudo tee -a /etc/fstab
echo "/dev/$VG/$LV2  /var/log/apps   xfs  defaults  0 2" | sudo tee -a /etc/fstab

# Step 7: Verify
sudo pvs && sudo vgs && sudo lvs
df -h /opt/appdata /var/log/apps

The most common real-world scenario — a VM gets a new virtual disk and you need to add it to an existing volume group:

# After adding disk in hypervisor — rescan for new device
sudo echo "- - -" > /sys/class/scsi_host/host0/scan
lsblk   # confirm new disk appears (e.g., /dev/sde)

# Initialize as PV
sudo pvcreate /dev/sde

# Add to existing VG
sudo vgextend vg_data /dev/sde

# Verify VG now has more free space
sudo vgs

# Extend an existing LV to use the new space
sudo lvextend -L +200G /dev/vg_data/lv_home

# Grow the filesystem (ext4)
sudo resize2fs /dev/vg_data/lv_home

# Grow the filesystem (XFS — must be mounted)
sudo xfs_growfs /home

# Verify
df -h /home

Removal must be done in reverse order — LV first, then VG, then PV:

# Step 1: Unmount the filesystem
sudo umount /opt/appdata

# Step 2: Remove from fstab
sudo vi /etc/fstab   # remove the relevant line

# Step 3: Remove the LV
sudo lvremove /dev/vg_apps/lv_appdata
# Confirm with 'y' when prompted

# Step 4: Remove the VG (only if removing all LVs)
sudo vgremove vg_apps

# Step 5: Remove the PV label
sudo pvremove /dev/sdb

# Verify everything is gone
sudo pvs
sudo vgs
sudo lvs