🔧 LVM Management

Growing an LV is the most common LVM operation — always online, always safe:

# Extend by an additional amount
sudo lvextend -L +50G /dev/vg_data/lv_home

# Extend to a specific total size
sudo lvextend -L 300G /dev/vg_data/lv_home

# Extend using all remaining free space in the VG
sudo lvextend -l +100%FREE /dev/vg_data/lv_home

# Extend AND resize the filesystem in one command (-r flag)
sudo lvextend -L +50G -r /dev/vg_data/lv_home

# Verify
sudo lvs /dev/vg_data/lv_home
df -h /home

If you extended without -r, resize the filesystem manually. The method depends on the filesystem type:

### ext4 — resize2fs (works mounted or unmounted) ###

# Extend the LV first
sudo lvextend -L +100G /dev/vg_data/lv_home

# Resize ext4 to fill the LV (no size needed — fills automatically)
sudo resize2fs /dev/vg_data/lv_home

# Verify
df -h /home


### XFS — xfs_growfs (must be mounted) ###

# Extend the LV first
sudo lvextend -L +100G /dev/vg_data/lv_var

# Grow XFS filesystem — use the mount point, not the device
sudo xfs_growfs /var

# Verify
df -h /var

Shrinking an LV is more dangerous than extending — the filesystem must be shrunk first, then the LV. Only ext4 supports shrinking; XFS cannot be reduced:

### ext4 ONLY — XFS cannot be shrunk ###

# Step 1: Unmount the filesystem
sudo umount /home

# Step 2: Check filesystem integrity first
sudo e2fsck -f /dev/vg_data/lv_home

# Step 3: Shrink the filesystem FIRST (before the LV)
sudo resize2fs /dev/vg_data/lv_home 80G

# Step 4: Shrink the LV to match
sudo lvreduce -L 80G /dev/vg_data/lv_home

# Step 5: Remount
sudo mount /dev/vg_data/lv_home /home

# Step 6: Verify
df -h /home

When a VG runs low on free space, add a new physical disk to it:

# Step 1: Confirm the VG is running low
sudo vgs
# VFree is small — time to add storage

# Step 2: Initialize the new disk as a PV
sudo pvcreate /dev/sde

# Step 3: Add it to the existing VG
sudo vgextend vg_data /dev/sde

# Step 4: Verify VG now has more free space
sudo vgs
sudo pvs   # shows new PV is now part of vg_data

# Step 5: Now extend LVs as needed
sudo lvextend -L +200G -r /dev/vg_data/lv_home

pvmove migrates data off a physical volume onto other PVs in the same VG — online, while the filesystem is mounted and in use. Essential for replacing a failing disk without downtime:

# Move ALL data off /dev/sdb to other PVs in the VG
sudo pvmove /dev/sdb

# Move data from one specific PV to another specific PV
sudo pvmove /dev/sdb /dev/sde

# Move data from a specific LV on a PV
sudo pvmove -n lv_home /dev/sdb

# Run in background (large volumes take time)
sudo pvmove -b /dev/sdb

# Monitor progress of a background pvmove
sudo lvs -a | grep pvmove
sudo watch -n5 'sudo pvs'

# After pvmove completes — remove the PV from the VG
sudo vgreduce vg_data /dev/sdb

# Remove the PV label (disk is now free)
sudo pvremove /dev/sdb

Renaming LVM objects — useful when server roles change or naming conventions need to be corrected:

# Rename a Logical Volume
sudo lvrename vg_data lv_home lv_users

# Or specify full paths
sudo lvrename /dev/vg_data/lv_home /dev/vg_data/lv_users

# Rename a Volume Group
sudo vgrename vg_data vg_production

# After renaming VG — update /etc/fstab entries
sudo vi /etc/fstab
# Change /dev/vg_data/... to /dev/vg_production/...

# Verify new names
sudo vgs
sudo lvs

Regular health checks to catch problems before they become outages:

# Quick overall status
sudo pvs && sudo vgs && sudo lvs

# Check for any partial or failed LVs
sudo lvs -a | grep -v "-wi-ao"

# Detailed attribute flags — look for anything not 'a' (active)
sudo lvs -o name,vgname,attr,size,pool_lv

# Check PV for missing or duplicate UUIDs
sudo pvscan

# Check VG consistency
sudo vgck vg_data

# Scan all block devices for LVM metadata
sudo lvscan

# Backup LVM metadata (do this regularly)
sudo vgcfgbackup vg_data
# Saved to /etc/lvm/backup/vg_data

# List metadata backups
ls -la /etc/lvm/backup/
ls -la /etc/lvm/archive/

When a PV disk fails and the VG is in a degraded state:

# VG shows as partial — a PV is missing
sudo vgs
# Attr shows 'p' (partial) instead of normal

# Activate VG in partial mode (access surviving LVs)
sudo vgchange -a y --partial vg_data

# Identify which PV is missing
sudo pvs
# Missing PV shows as [unknown]

# If replacement disk is available — add it
sudo pvcreate /dev/sdf
sudo vgextend vg_data /dev/sdf

# Remove the missing PV record from the VG
sudo vgreduce --removemissing vg_data

# If VG metadata backup exists — restore
sudo vgcfgrestore -f /etc/lvm/archive/vg_data_00001.vg vg_data

# Check filesystem integrity after recovery
sudo e2fsck -f /dev/vg_data/lv_home