LVM creation, extension, and snapshots · Linux

LVM operations are routine on production hosts. Knowing the right commands and order of operations prevents data loss and unnecessary downtime.

Creating an LVM stack

Initialize a disk as a Physical Volume (PV): pvcreate /dev/sdXN
Combine one or more PVs into a Volume Group (VG): vgcreate vg_data /dev/sdXN
Allocate a Logical Volume (LV): lvcreate -L 100G -n data vg_data
Create a filesystem on the LV: mkfs.ext4 /dev/vg_data/data
Mount the LV: mkdir /data; mount /dev/vg_data/data /data
Add to fstab: UUID=... /data ext4 defaults,nofail 0 2
Verify: df -h /data; lvs

Data-loss riskcreate LV stack— pvcreate initializes the PV (writes LVM metadata to the disk). vgcreate creates the VG. lvcreate allocates the LV. mkfs creates the filesystem. The order matters: PV first, then VG, then LV, then filesystem.

$ pvcreate /dev/sdc1; vgcreate vg_data /dev/sdc1; lvcreate -L 100G -n data vg_data; mkfs.ext4 /dev/vg_data/data

Physical volume '/dev/sdc1' successfully created.
Volume group 'vg_data' successfully created
Logical volume 'data' successfully created
mke2fs 1.47.0 (5-Feb-2023)
Filesystem UUID: 1a2b-3c4d-...

Extending an LV

Configuration changeextend LV— lvextend grows the LV. resize2fs grows the ext4 filesystem to match. For XFS, use xfs_growfs. Both are online - the filesystem stays mounted throughout. Order matters: extend the LV first, then the filesystem.

$ lvextend -L +50G /dev/vg_data/data; resize2fs /dev/vg_data/data

Size of logical volume vg_data/data changed from 100.00g to 150.00g
resize2fs 1.47.0 (5-Feb-2023)
Resizing the filesystem on /dev/vg_data/data to 39321600 (4k) blocks.

LVM snapshots

Configuration changeLVM snapshot— LVM snapshot creation is instant. The 5G is the initial reservation; the snapshot uses more space as the original LV changes. Use -s for snapshot, -n for the name. The snapshot is mounted as /dev/vg_data/data-snap.

$ lvcreate -L 5G -s -n data-snap /dev/vg_data/data

Logical volume 'data-snap' successfully created

Configuration changeuse and remove snapshot— Mount the snapshot to verify a change before committing it, or to copy data to another location. When done, unmount and lvremove; the snapshot disappears after the lvremove. The nouuid option matters: a snapshot is a block-identical copy, so it carries the origin's filesystem UUID, and XFS refuses to mount a second filesystem with a UUID that is already in use. Without nouuid the mount fails outright while the origin is mounted - which is always, since that is the point of a snapshot. ext4 has no such restriction and ignores the option's absence.

$ mount -o ro,nouuid /dev/vg_data/data-snap /mnt/snap; ls /mnt/snap; umount /mnt/snap; lvremove /dev/vg_data/data-snap

Snapshots dir
file1
file2
Do you really want to remove active logical volume data-snap? [y/n]: y
Logical volume 'data-snap' successfully removed

Monitoring LVM

Read-only / Safemonitoring— pvs/vgs/lvs show the storage stack. The -o+ options add extra columns. Check that the layers reconcile: 99.00g + 500.00g of PV equals the 599.00g VG, and the three LVs (5 + 90 + 295 = 390.00g) plus 209.00g free account for all of it. Note the column widths: a VG attribute is six characters (wz--n-) and an LV attribute is ten (-wi-ao----); a six-character string in an LV row means you are reading the wrong output. PFree and VFree are critical for capacity planning - if VFree is low, the VG cannot grow any LV without adding a new PV.

$ pvs -o+pv_size,pv_free,pv_used; vgs -o+vg_size,vg_free,vg_used; lvs -o+lv_size,lv_used

  PV         VG   Fmt  Attr PSize   PFree   Used
/dev/sda2  vg0  lvm2 a--   99.00g   4.00g   95.00g
/dev/sdb1  vg0  lvm2 a--  500.00g 205.00g  295.00g
--- VG   #PV #LV #SN Attr   VSize   VFree
vg0    2   3   0 wz--n- 599.00g 209.00g
--- LV   VG  Attr       LSize   Used
root vg0 -wi-ao----   5.00g
var  vg0 -wi-ao----  90.00g
data vg0 -wi-ao---- 295.00g