Docker service cannot be started sometimes, because if /var/run/docker.sock does not exist on the host system/node the docker creates/binds directory with the same name, but the dockerd fails that the socket is a directory

Ansible log:

TASK [docker : Setup docker dns settings] ***************************************************************************************************************************************************************************************************

task path: /ansible/roles/docker/tasks/main.yml:19

ok: [kubernetes-node-1] => {"changed": false, "file_attr_diff": {}, "msg": ""}

META: ran handlers

TASK [docker : Ensure docker is started] ****************************************************************************************************************************************************************************************************

task path: /ansible/roles/docker/tasks/main.yml:30

fatal: [kubernetes-node-1]: FAILED! => Unknown macro: {"changed"}

Systemd log:

– Unit docker.service has begun starting up.
dockerd[23116]: can't create unix socket /var/run/docker.sock: is a directory
systemd[1]: docker.service: main process exited, code=exited, status=1/FAILURE
systemd[1]: Failed to start Docker Application Container Engine.
– Subject: Unit docker.service has failed

Workaround: rm -rf /var/run/docker.sock && systemctl start docker.service