How to List and Remove Docker Containers and Images

Every docker run leaves a container behind unless you remove it, and every pull and build adds image data. After a few weeks of development or CI, you have dozens of stopped containers, old images and gigabytes of build cache, and sooner or later a name is already in use error or a full disk.

This post shows how to list containers and images, fix name conflicts, remove what you no longer need from the safest command to the most aggressive, find where Docker keeps its data, and get out of a no space left on device situation.

TLDR

Task	Command
Running containers	docker ps
All containers, stopped ones too	docker ps -a
Remove one container	docker rm web (-f if it runs)
Remove all stopped containers	docker container prune
Remove one image	docker image rm nginx:1.30
Remove images no container uses	docker image prune -a
Remove the build cache	docker builder prune
See what uses the disk	docker system df

None of these commands remove named volumes. Your data in volumes stays until you run docker volume rm or docker volume prune --all.

Prerequisites

• Docker Engine or Docker Desktop. The recorded output below comes from Docker Engine 28.0.4. Engine 29 shows a different default layout for docker image ls.
• Permission to run docker.

How Do I List Docker Containers?

docker ps shows running containers. Add -a to include stopped and created ones:

docker ps        # running only
docker ps -a     # everything, with STATUS such as "Exited (0) 2 hours ago"

docker container ls is the same command with a longer name. The options you need most:

docker ps -a --filter status=exited                       # only stopped containers
docker ps -a --filter name=web                            # names that contain "web"
docker ps -a --filter ancestor=nginx:1.30                 # containers of one image
docker ps -q                                              # only the IDs, for scripts
docker ps -s                                              # with the size of each writable layer
docker ps --format 'table {{.Names}}\t{{.Image}}\t{{.Status}}\t{{.Ports}}'

In a Compose project, docker compose ps lists only that project's containers, and docker compose ps -a includes stopped ones.

How Do I Fix "The Container Name Is Already in Use" Error?

docker: Error response from daemon: Conflict. The container name "/web" is already in use by container "3aff97197020...". You have to remove (or rename) that container to be able to reuse that name.

Container names are unique, and a stopped container still owns its name. docker ps does not show it, so run docker ps -a --filter name=web, then pick one fix:

docker start web                 # you wanted the old container: start it again
docker rm web                    # you do not need it: remove it (add -f if it runs)
docker rename web web-old        # keep it, but free the name for a new container

To avoid the error in scripts, remove the old container first (docker rm -f web 2>/dev/null; docker run --name web ...) or let Docker remove containers on exit with docker run --rm. Compose avoids it by reusing or recreating its own containers, as long as you do not set container_name: to a name that another container already uses.

How Do I Remove Old Docker Containers?

docker rm web                    # one stopped container
docker rm -f web                 # stop it first if it runs
docker container prune           # every stopped container, after a prompt
docker container prune -f --filter "until=24h"   # stopped ones created over 24 hours ago, no prompt

docker rm deletes the container's writable layer. Named volumes stay, and so do anonymous volumes unless you add -v (docker rm -v web). For one-off containers, docker run --rm removes the container as soon as it exits, so nothing piles up.

How Do I Remove Docker Images?

Remove one image by name or ID with docker image rm (the short form is docker rmi):

docker image rm nginx:1.30
docker rmi 53d8f012a1a7

Removing a name with docker image rm only removes the image when no other tag points to it. If a container uses the image, even a stopped one, Docker refuses:

docker ps -a --filter ancestor=myapp:1.0 -q | xargs -r docker rm   # remove the containers first
docker image rm myapp:1.0

For bulk cleanup, prune is safer than a loop over docker rmi:

docker image ls --filter dangling=true      # untagged leftovers of rebuilds
docker image prune                          # remove those dangling images
docker image prune -a                       # every image that no container uses
docker image prune -a --filter "until=168h" # the same, but only images older than 7 days

docker image prune -a never removes an image that a container uses. The images come back with the next docker pull or docker run, so the cost of an aggressive prune is download time, not data. Images that you built locally and never pushed are the exception: push or docker save them first.

Delete All Local Docker Images

To start from zero, remove the containers first, then every image:

docker ps -aq | xargs -r docker rm -f   # every container, running or stopped
docker image prune -a -f                 # every image that no container uses, which is now all of them

The one-liner you will find everywhere, docker rmi $(docker images -q), fails in three ways at once. Here one container runs nginx:1.30, a stopped container still references hello-world, and alpine:3.22 also has the tag myapp:dev:

docker images -q prints an ID once per tag, so an image with two tags is listed twice and refused twice. A running container blocks its image even with -f; a stopped one blocks it unless you force. docker image prune -a skips images that any container uses, stopped containers included, which is why nginx and hello-world survived the first prune. When nothing is left, $(docker images -q) is empty and docker rmi gets no argument; xargs -r runs nothing in that case. The output above is from Docker Engine 28.0.4.

docker system prune -a removes stopped containers, unused networks, unused images and the build cache in one command, but it leaves running containers and their images alone. Add --volumes only if you mean to delete data as well.

Where Does Docker Store Images and Containers?

docker info -f '{{.DockerRootDir}}' prints Docker's data directory. The defaults:

Platform	Where the data is
Linux, Docker Engine	/var/lib/docker for containers, volumes and logs; /var/lib/containerd for images with the containerd image store
Docker Desktop on macOS	Inside a Linux VM, in one disk image file under ~/Library/Containers/com.docker.docker/Data/vms/0/data/
Docker Desktop on Windows	Inside a WSL 2 virtual disk (a .vhdx file) under %LOCALAPPDATA%\Docker\wsl
Windows containers	C:\ProgramData\Docker

Docker Engine 29 uses the containerd image store for new installs. Image layers are then kept by containerd under /var/lib/containerd, not in /var/lib/docker/overlay2. docker info shows which store you have: the Storage Driver line reads overlayfs with driver-type: io.containerd.snapshotter.v1 for the containerd store, and overlay2 for the classic one. Hosts that were upgraded from older versions keep the classic store until you switch.

Do not delete files in these directories by hand. Docker tracks every layer in its own database, and removing directories behind its back corrupts images and containers. Use the prune commands, or the steps below.

On Docker Desktop, the VM disk file only grows. After a big cleanup, the file can stay large on disk. Docker Desktop's Troubleshoot menu has options to clean up or reset the data, and Settings > Resources sets the disk size limit.

How Do I Fix "No Space Left on Device" in Docker?

When a build or a container fails with no space left on device, the disk that holds Docker's data directory is full. Check that disk first, then see which kind of Docker object uses the space:

# Where Docker keeps its data, and how full that disk is
docker info -f '{{.DockerRootDir}}'
df -h "$(docker info -f '{{.DockerRootDir}}')"

# Images, containers, volumes and build cache, with what can be reclaimed
docker system df

Then free space from the safest step to the most aggressive:

1. docker container prune removes stopped containers.
2. docker image prune removes dangling images, the untagged leftovers of rebuilds. docker image prune -a removes every image that no container uses; they download again when needed.
3. docker builder prune removes the build cache, which grows fast on CI machines and build servers.
4. docker system prune does steps 1 to 3 (dangling images only) plus unused networks.

Volumes are safer than the --volumes flag suggests. Since Docker Engine 23, docker system prune --volumes and docker volume prune remove only anonymous volumes. In our test on Engine 28, a prune removed the anonymous volume of a deleted container and kept a named volume called named-data. Named volumes go only with docker volume rm or docker volume prune --all, so check docker volume ls before you use either one.

If docker system df shows little to reclaim but the disk is still full, look at two more places:

• Container logs. The default json-file driver never rotates, and docker system df does not count logs. sudo du -sh /var/lib/docker/containers/*/*-json.log | sort -h | tail finds the big ones. Clear a log that already fills the disk shows how to clear one and set a limit.
• Inodes. df -i shows whether the file system ran out of inodes rather than bytes, which happens with many small layer files. Pruning helps here too.

If Docker simply needs more room, move its data directory to a bigger disk, as below.

How Do I Move Docker's Data to Another Disk?

On Linux, set data-root in /etc/docker/daemon.json and move the data while Docker is stopped:

sudo systemctl stop docker docker.socket
sudo rsync -aHAX /var/lib/docker/ /mnt/bigdisk/docker/
echo '{ "data-root": "/mnt/bigdisk/docker" }' | sudo tee /etc/docker/daemon.json   # merge by hand if the file has other settings
sudo systemctl start docker
docker info -f '{{.DockerRootDir}}'     # check the new path, then delete the old directory

With the containerd image store, images live in containerd's directory, which data-root does not move. Change root in /etc/containerd/config.toml too, stop containerd with Docker, and move /var/lib/containerd the same way. On Docker Desktop, change the disk image location in Settings > Resources > Advanced.

How Do I Keep Docker From Filling the Disk Again?

• Run one-off containers with --rm.
• Limit container logs. The default json-file driver keeps logs forever. Set "log-driver": "local" or "log-opts": {"max-size": "10m", "max-file": "3"} in daemon.json.
• Prune on a schedule on build servers and CI runners, for example a weekly cron job: docker system prune -af --filter "until=168h". On a server that runs production containers, prune only stopped containers and dangling images, so a rollback does not have to download the previous image.
• Watch docker system df and df -h in your monitoring, not only after the disk is full.

Related Resources

• Docker image vs container: why removing a container keeps the image
• How to keep data when a container exits: which commands remove volumes
• Docker Desktop's local log driver: what changes for log disk use
• Introduction to Docker: Free Disk Space
• Docker terminal simulator: practice ps, rm and prune without risk
• Docker flashcards: review the commands