Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOn a conventional Docker Engine installation on Ubuntu, docker pull stores image layers under Docker’s local data root, normally /var/lib/docker. When the classic overlay2 driver is active, the layer directories are beneath /var/lib/docker/overlay2.
Antihost offers VPS plans with one-click Docker, NVMe storage and a 14-day free trial.
That default is not universal. The daemon’s data root can be changed, and fresh Docker Engine 29.0-or-later installations use Docker’s containerd image store by default, placing image contents and container snapshots under /var/lib/containerd. The authoritative answer for any host is the Docker Root Dir and Storage Driver shown by docker info.
As an Amazon Associate I earn from qualifying purchases.
Use docker info to find the active location
Filesystem paths vary according to Docker’s storage backend and configuration. Do not infer the location from the Ubuntu version or from a tutorial written for another Docker release. On the host where the pull occurred, run:
docker infoIn the output, locate:
- Docker Root Dir: the daemon’s configured local data directory for the active backend.
- Storage Driver: the classic filesystem driver, such as
overlay2, used by that daemon.
A typical older or explicitly classic installation reports Docker Root Dir: /var/lib/docker and Storage Driver: overlay2. In that case, pulled layers can be inspected beneath /var/lib/docker/overlay2. If the root directory is different, use the reported directory rather than assuming /var/lib/docker.
#1 Best Overall
Most Docker administration commands need root privileges on Ubuntu. If your account is not in Docker’s Unix group, run the diagnostic as sudo docker info.
What the classic /var/lib/docker layout means
Docker’s storage-driver documentation describes a pull as a series of layer downloads: each image layer is fetched separately and stored in Docker’s local storage area. For a classic Linux Engine, that area is usually /var/lib/docker. The image name and tag are metadata; the layer contents are content-addressed files managed by the daemon.
The overlay2 directory
With overlay2, the relevant top-level directory is:
/var/lib/docker/overlay2Each subdirectory represents storage used by an image layer or a container’s writable layer. Directory names are implementation identifiers, not friendly image names, so you should not expect a folder called ubuntu after pulling an Ubuntu image.
OverlayFS presents a union of read-only lower layers and one writable upper layer as a merged filesystem. A running container’s storage directory can include:
lower-id, which records the lower-layer relationship;merged, the unified view presented to the container;upper, the container’s writable changes; andwork, an OverlayFS working directory.
Image layers are shared through copy-on-write. Several images can therefore refer to the same lower layer, and deleting one image does not necessarily reclaim all of the apparent directory size if another image or container still references those layers.
Recommended Free Tools
Why the directory can be large
A pull may bring several layers, and a container created from that image can add a writable layer. Updates made inside a container—installed packages, generated files, and deleted-file whiteouts—belong to that writable layer rather than changing the original image. The writable layer is ephemeral: removing the container removes that layer. It is not a suitable place for data that must survive container replacement.
Persistent application data belongs in a Docker volume or a bind mount. Volumes are managed separately from image layers, while a bind mount uses a directory on the host that you choose. Consequently, disk usage attributed to Docker can come from image layers, stopped containers, writable layers, volumes, build cache, or logs; the path alone does not identify which category is responsible.
Docker Engine 29.0 and the containerd image store
Docker’s current storage model introduces an important version distinction. Fresh Docker Engine 29.0 or later installations use the containerd image store by default. With that backend, image contents and container snapshots are held under:
/var/lib/containerdAn installation upgraded from an earlier release, or one configured to retain the classic image store, may continue to report a classic driver such as overlay2 and keep its data under the Docker root directory. Therefore, “Docker images are in /var/lib/docker” is a description of the classic backend, not a guarantee for every current installation.
Free tools Windows power users keep installed
One-click scans. No signup required.
The containerd backend has its own root setting in /etc/containerd/config.toml. Docker’s data-root setting does not relocate image and container data that is being stored by containerd. Check the active backend first, then change the configuration belonging to that backend.
Classic and containerd locations at a glance
| Backend reported by the host | Typical data location | What to verify |
|---|---|---|
Classic storage driver, such as overlay2 |
/var/lib/docker, with overlay layers under /var/lib/docker/overlay2 |
Docker Root Dir, Storage Driver, and any customized data-root |
| Containerd image store | /var/lib/containerd by default |
Containerd’s configured root in /etc/containerd/config.toml |
| Any customized classic root | The directory shown as Docker Root Dir |
The daemon configuration and filesystem permissions |
The table shows defaults, not a substitute for inspecting the machine. A non-default mount, package configuration, or administrator change can move either data root.
How to change the classic Docker data directory
For the classic Docker image store, Docker documents the data-root setting in /etc/docker/daemon.json. For example:
{"data-root": "/mnt/docker-data"}Changing this value tells the classic daemon where to keep its Docker-managed data. It does not automatically migrate existing files, and it does not control a containerd image store.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteSafe migration outline for a classic installation
- Confirm the current backend and root with
sudo docker info. Proceed only if the host uses the classic store you intend to move. - Stop Docker so files are not changing during the copy:
sudo systemctl stop docker - Create the destination on the mounted filesystem, for example:
sudo mkdir -p /mnt/docker-data - Copy the existing Docker root while preserving ownership, permissions, links, extended attributes, and hard links. An administrator commonly uses:
sudo rsync -aHAX /var/lib/docker/ /mnt/docker-data/ - Edit
/etc/docker/daemon.json. If the file already contains settings, adddata-rootto the existing JSON object instead of replacing unrelated options. The resulting value should point to the destination directory. - Start Docker and verify the result:
sudo systemctl start docker
sudo docker infoConfirm that
Docker Root Dirnow shows the destination and that existing images and containers are present. - Keep the old directory until verification and a rollback plan are complete. Do not delete it merely because the daemon starts; check the workloads that depend on the stored images and volumes first.
The exact migration command is an administrative choice, but the safety requirements are constant: stop the daemon, preserve metadata, ensure the destination is mounted at boot, and verify with Docker rather than judging by directory names.
Moving data with the containerd backend
If docker info indicates a containerd-backed image store, configure containerd’s own root in /etc/containerd/config.toml. A Docker data-root entry will not move those image contents. Plan the change with both services stopped, preserve the existing containerd data and permissions, then restart containerd and Docker and verify that images and containers are available. The active containerd configuration and service status should be treated as authoritative for that backend.
Tell image layers apart from other Docker disk usage
When /var/lib/docker or /var/lib/containerd consumes substantial space, identify the object before deleting anything. Start with Docker’s accounting command:
docker system df -vThis reports space associated with images, containers, local volumes, and build cache, including reclaimable categories where Docker can calculate them. Useful supporting checks include:
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →docker image lsto list locally tagged images and their displayed sizes;docker ps -a --sizeto show container writable-layer sizes, including stopped containers; anddocker volume lsto inventory named volumes before considering volume cleanup.
A filesystem tool such as du can show which top-level directory is growing, but it cannot safely determine layer ownership by itself. Treat Docker’s metadata as the source of truth.
Supported cleanup choices
Remove objects through Docker rather than deleting files inside its data root:
docker image pruneremoves dangling images after confirmation.docker container pruneremoves stopped containers after confirmation.docker volume pruneremoves unused volumes; check application backups first because volume data can be persistent.docker system prunecombines several unused-object categories. Adding--volumesexpands the operation to unused volumes and therefore requires extra caution.
These commands operate on Docker’s records and references. Never “free space” by removing an arbitrary directory beneath overlay2 or by deleting files from the containerd root.
Why manually editing overlay2 breaks Docker
Docker explicitly warns not to directly manipulate files or directories within its data directory. Layer IDs, metadata, reference counts, and copy-on-write relationships must remain consistent. Renaming a layer directory, deleting a file that appears unused, or copying only part of the tree can make images unpullable, containers unstartable, or the daemon unable to mount a layer.
Use read-only inspection when diagnosing a problem, and use Docker commands for removal. If a layer directory appears orphaned, gather docker info, docker system df -v, daemon logs, and the affected image or container IDs before taking corrective action.
Uninstalling Docker does not necessarily remove the images
On Ubuntu, removing the Docker Engine package does not automatically prove that Docker-managed data has been erased. Images, containers, volumes, and networks stored in /var/lib/docker can remain after an uninstall. This explains why reinstalling Docker may appear to restore old images or why disk usage persists after package removal.
Best Value
Before deleting a former data root, verify that no Docker or containerd service still references it and that you have exported any images or backed up any volumes you need. Deleting the directory is a separate, deliberate data-destruction decision—not a routine consequence of uninstalling the package.
Common location questions, answered by diagnosis
“I pulled an image, but I cannot find a folder named after it.”
That is expected. Overlay2 uses internal layer identifiers and shared content rather than one human-readable directory per repository tag. Use docker image ls for image names and docker info for the storage root.
“Why is /var/lib/docker/overlay2 full when my images look small?”
Displayed image sizes can share layers, while the filesystem also contains container writable layers, build cache, and metadata. Check docker system df -v and container sizes before removing anything.
“I changed data-root, but the disk usage did not move.”
First confirm that Docker actually reports the new root. If the host uses the containerd image store, data-root is the wrong setting for image contents; inspect containerd’s root configuration instead. Also verify that the destination filesystem is mounted and that the daemon restarted successfully.
“Can I open or edit files in the layer directory?”
You may inspect paths read-only for diagnostics, but do not edit, move, or delete Docker-managed layer files. Use Docker’s image, container, volume, and prune commands to make changes.
Quick Recap
Practical checklist
- Run
sudo docker info. - Record
Docker Root DirandStorage Driver. - For classic
overlay2, inspect the reported root’soverlay2subdirectory; for containerd, inspect the configured containerd root. - Use
docker system df -vto distinguish images, writable layers, volumes, and cache. - Change
data-rootonly for the classic store; change containerd’srootfor the containerd store. - Never delete files inside Docker’s managed storage tree by hand.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
