Remove unused images

[miguelramos]

Question:
How can I remove unused images with a cron job?

Current i have this kind of script running:

docker container rm $(docker ps -qa --no-trunc --filter "status=exited") || true >> /var/log/docker/containers.log 2>&1
docker rmi $(docker image ls -aq --filter "dangling=true") || true >> /var/log/docker/images_dangling.log 2>&1

But when i check images list in portainer browser, the unused images are there. So how can i remove periodically unused images on portainer to free space?

[tronyx]

Portainer cannot do that. Put your above commands into a script and create a cronjob for the script.

[miguelramos]

Portainer cannot do that. Put your above commands into a script and create a cronjob for the script.

I don't know if you understood but my script example above is a cron job, and unused images are not clean when it runs, I can only remove images if I go to portainer and remove it manually.

[tronyx]

You can use docker system prune as well, even in a script/cronjob, to cleanup unused images.

[ghost]

@christronyxyocum Portainer does in fact have cron job functionality.

1. Enable host management in settings and browse to the host jobs view
   

2. Click on add a schedule
   

3. Configure and save the schedule & the job will run as desired.
   

I just tested this example on Portainer 1.23 to remove un-used images and it worked as intended.
Note that docker image prune -af will remove ALL un-used images

[ghost]

Here is a command that you can perhaps modify to your liking, this will delete images older than 24 hours
docker image prune --all --filter "until=24h" --force official docker documentation: https://docs.docker.com/config/pruning/#prune-images

Note that --force is required for execution without the confirmation prompt

[ghost]

I'll close this as I believe i have answered the question. If needed I can re-open to discuss further

[gianlucapisati]

I'm having problems with this configuration, I've added the host jobs as suggested, using ubuntu:latest as base image and adding
docker image prune --all --force but the images seems not to be removed.

If I manually launch the container I receive this error:
Failure OCI runtime create failed: container_linux.go:348: starting container process caused "exec: \"sh\": executable file not found in $PATH": unknown

I've noticed that the CMD of the container is sh /tmp/script.sh. It seems strange to me!

By the way, it's not clear to me if this command is executed just on the host node or everywhere

[kungl]

Never use latest tag, I have the same problem with the ubuntu:latest image, but if you use ubuntu:18.04 it'll work. You can set endpoints in "Target endpoints" section (above the Create sections button) .

[iambryancs]

@kungl you saved me! Thank you. I think when someone creates a new Schedule under Host Jobs, the Image's value under Job configuration should not use latest as the default tag.

[cinderblock]

If you're using Watchtower to automatically update containers, and that is generating a lot of unused images, you might want to enable WATCHTOWER_CLEANUP.