Corrupted filesystem on remote docker machine

My builds started failing last Thursday and after looking into it looks like whenever we work on a Docker container using the python:3.9 image the filesystem is corrupted.

Strangely this only seems to happen on the python:3.9 tag (specifically image id 1e76b28bfd4e) but works with other tags (eg. python:3.9-alpine is fine). And that same image id works fine in other environments, leading me to think there’s something wrong with using some python images specifically on CircleCI remove docker environments.

Has anyone seen this issue before? Is there anything I can do to fix it on my end?


What version of remote Docker are you using? The default?

AFAIK we’re not setting a version anywhere, so I assume we’re using the default? Here’s the relevant step from our config file:

- setup_remote_docker:
    docker_layer_caching: true

Resolution: upgrading to docker 20.10.7 solved the issue.
Related docs: Running Docker Commands - CircleCI

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.