Aller au contenu

DockerThinPoolFilledUpFE#

This alert is triggered when the available space on the node has been less than 10% of the minimal space required by the docker thin pool for at least 30 minutes.

As a result, Docker will not be able to pull new docker images and this can lead to a pod not being scheduled on the node.

Troubleshooting hints#

  • Check if you can make space by removing dangling and unused images for the node

    kubectl node-shell <NODE> -- docker image prune -a

  • Check that the previous step has resolved the issue.

    # "Data Space Available" should be higher than "Thin Pool Minimum Free Space"
kubectl node-shell <NODE> -- docker info

  • If it has not, find a way to make some space on the Docker thinpool filesystem or increase its size using this tutorial ("Resize a direct-lvm thin pool" section)