为什么 Docker 映像占用了我的磁盘空间，而这些空间并没有被 Docker 使用

It's a kernel problem with devicemapper, which affects the RedHat family of OS (RedHat, Fedora, CentOS, and Amazon Linux). Deleted containers don't free up mapped disk space. This means that on the affected OSs you'll slowly run out of space as you start and restart containers.

The Docker project is aware of this, and the kernel is supposedly fixed in upstream (https://github.com/docker/docker/issues/3182).

A work-around of sorts is to give Docker its own volume to write to ("When Docker eats up you disk space"). This doesn't actually stop it from eating space, just from taking down other parts of your system after it does.

My solution was to uninstall docker, then delete all its files, then reinstall:

sudo yum remove docker
sudo rm -rf /var/lib/docker
sudo yum install docker

This got my space back, but it's not much different than just launching a replacement instance. I have not found a nicer solution.

Move the /var/lib/docker directory.

Assuming the /data directory has enough room, if not, substitute for one that does,

sudo systemctl stop docker


sudo mv /var/lib/docker /data




sudo ln -s /data/docker /var/lib/docker


sudo systemctl start docker

This way, you don't have to reconfigure docker.

Yes, Docker use /var/lib/docker folder to store the layers. There are ways to reclaim the space and move the storage to some other directory.

You can mount a bigger disk space and move the content of /var/lib/docker to the new mount location and make sym link.

There is detail explanation on how to do above task.

http://www.scmtechblog.net/2016/06/clean-up-docker-images-from-local-to.html

You can remove the intermediate layers too.

https://github.com/vishalvsh1/docker-image-cleanup

As mentioned in issue #18867 - Delete data in a container devicemapper can not free used space from Github.com

Try running the below command:

# docker ps -qa | xargs docker inspect --format='\{\{ .State.Pid }}' | xargs -IZ fstrim /proc/Z/root/

It uses the fstrim tool to trim the devicemapper thinly-provisioned disk.

Deleting my entire /var/lib/docker is not ok for me. These are a safer ways:

Solution 1:

The following commands from the issue clear up space for me and it's a lot safer than deleting /var/lib/docker or for Windows check your disk image location here.

Before:

docker info

Example output:

Metadata file:
Data Space Used: 53.38 GB
Data Space Total: 53.39 GB
Data Space Available: 8.389 MB
Metadata Space Used: 6.234 MB
Metadata Space Total: 54.53 MB
Metadata Space Available: 48.29 MB

Command in newer versions of Docker e.g. 17.x +

docker system prune -a

It will show you a warning that it will remove all the stopped containers,networks, images and build cache. Generally it's safe to remove this. (Next time you run a container it may pull from the Docker registry)

Example output:

Total reclaimed space: 1.243GB

You can then run docker info again to see what has been cleaned up

docker info

Solution 2:

Along with this, make sure your programs inside the docker container are not writing many/huge files to the file system.

Check your running docker process's space usage size

docker ps -s #may take minutes to return

or for all containers, even exited

docker ps -as #may take minutes to return

You can then delete the offending container/s

docker rm <CONTAINER ID>

Find the possible culprit which may be using gigs of space

docker exec -it <CONTAINER ID> "/bin/sh"
du -h

In my case the program was writing gigs of temp files.

(Nathaniel Waisbrot mentioned in the accepted answer this issue and I got some info from the issue)

OR

Commands in older versions of Docker e.g. 1.13.x (run as root not sudo):

# Delete 'exited' containers
docker rm -v $(docker ps -a -q -f status=exited)


# Delete 'dangling' images (If there are no images you will get a docker: "rmi" requires a minimum of 1 argument)
docker rmi $(docker images -f "dangling=true" -q)


# Delete 'dangling' volumes (If there are no images you will get a docker: "volume rm" requires a minimum of 1 argument)
docker volume rm $(docker volume ls -qf dangling=true)

After :

> docker info
Metadata file:
Data Space Used: 1.43 GB
Data Space Total: 53.39 GB
Data Space Available: 51.96 GB
Metadata Space Used: 577.5 kB
Metadata Space Total: 54.53 MB
Metadata Space Available: 53.95 MB

I had a similar problem and I think this happens when you don't have enough space in the disk for all your docker images. I had 6GB reserved for docker images which it turned out not to be enough in my case. Anyway, I had removed every image and container and still disk looked full. Most of the space was being used by /var/lib/docker/devicemapper and /var/lib/docker/tmp.

This command didn't work for me:

# docker ps -qa | xargs docker inspect --format='\{\{ .State.Pid }}' | xargs -IZ fstrim /proc/Z/root/

First, I stopped docker service:

sudo service docker stop

Then I deleted /var/lib/docker:

Then I did what somebody suggested here in https://github.com/docker/docker/issues/18867#issuecomment-232301073

• Remove existing instance of docker metadata rm -rf /var/lib/docker

  sudo rm -rf /var/lib/docker

• Pass following options to docker daemon: -s devicemapper --storage-opt dm.fs=xfs --storage-opt dm.mountopt=discard

• Start docker daemon.

For last two steps, I run:

sudo dockerd -s devicemapper --storage-opt dm.fs=xfs --storage-opt dm.mountopt=discard

maybe you can try docker system prune to remove all the images that are not important

Docker prune by default does not remove volumes,

you can try something like

docker volume prune -f

Had the same problem. In my scenario my vbox was running out of storage space. After an investigation found out that my docker local volumes were eating up 30gb. Ubuntu 16.04 host.

To find out yours.

docker system df


TYPE                TOTAL               ACTIVE              SIZE                RECLAIMABLE
Images              3                   0                   1.361GB             1.361GB (100%)
Containers          0                   0                   0B                  0B
Local Volumes       7                   0                   9.413GB             9.413GB (100%)
Build Cache                                                 0B                  0B






docker system prune --volumes




WARNING! This will remove:
- all stopped containers
- all networks not used by at least one container
- all volumes not used by at least one container
- all dangling images
- all build cache
Are you sure you want to continue? [y/N]

This frees up disk space of not used local volumes. At my scenario freed 20 GB of storage space. Make sure that containers which you want to keep are running before doing this if you want to keep them since this removes all stopped containers.

For the ones running into this issue in MacOS, the solution that has worked for me was to look for Docker.raw file which is the one Docker uses to reserve the logical space in the host and then deleting. If you have Docker Desktop, you can go:

Preferences -> Resources -> Advanced and then look under the Disk image location tab.

Navigate to that folder in terminal and just delete the Docker.raw file ($ rm -rf Docker.raw)

Important Note: Only do this is you don't need any of your existing images or volumes.

In case you're running K3S(miniikube) over docker run the following command:

minikube ssh -- docker system prune

Minikube would use Docker Local Volumes. So you have to ssh within minikube to prune your system. PS: If you want to clean up your volumes:

minikube ssh -- docker system prune -a --volumes -f