Failing workers with cgroup already being mounted

Currently all tasks fail on 5.7.2 native worker on fedora:

mount: /sys/fs/cgroup/memory: cgroup already mounted on /sys/fs/cgroup.

not sure how to go from here, 5.2.x worked just fine and I am not entirely sure how to get to the bottom of this.

Sample:
https://ci.spearow.io/teams/spearow/pipelines/juice-containers/jobs/create-oci-fedora-base/builds/8

Fedora 31 uses cgroups v2 by default and you have to switch back to v1 to allow containers to work

# dnf install -y grubby
# grubby --update-kernel=ALL --args="systemd.unified_cgroup_hierarchy=0"
# reboot

I also encountered problems getting networking to function properly though, while running a concourse worker natively on Fedora 31. I wasn’t able to get it fixed and instead I just ran the worker under docker.

1 Like

Thanks! I totally forgot about that!
Edit: Created a ticket for that https://github.com/concourse/concourse/issues/5080

For the network issues, take a look at
https://discuss.concourse-ci.org/t/firewalld-deleting-concourse-created-iptables-chains/1845/3