Quiz Entry - updated: 2026.07.14
What is VDO in LVM, and what three things does it do?
VDO (Virtual Data Optimizer) is an LVM layer that transparently saves space using inline deduplication, compression, and thin provisioning as data is written.
VDO sits in the storage stack and shrinks data in-line — while it's being written, not in a later batch job — so applications never see the optimization happening. Its three techniques stack:
| Technique | What it does |
|---|---|
| Deduplication | Spots blocks identical to ones already stored and keeps just one copy, pointing the duplicates at it |
| Compression | Squeezes each unique block smaller, then packs several compressed blocks together into physical blocks |
| Thin provisioning | Hands out logical space on demand and discards all-zero blocks, so you only consume physical space for real data |
dnf install vdo kmod-kvdo # the tool + kernel module
lvcreate --type vdo --name vdo-lv01 --size 5G vg01 # a VDO-backed LV
Why "redundant data" is the magic word: dedup and compression only pay off when blocks actually repeat. The big wins are therefore:
- Backup storage — successive backups share most of their blocks.
- VM and container images — dozens of guests run near-identical OS files, so most blocks are duplicates.
On already-compressed or encrypted data (videos, archives) VDO saves almost nothing, because there's no redundancy left to find.