Missing selinux files on CentOS causes failure of kubeone apply

[michalvanco]

What happened:
When using CentOS (7.7) to provision kubeone cluster, apply steps fails on missing two files on the system (/etc/sysconfig/selinux, /etc/selinux/config):

...
[136.144.48.169] + sudo setenforce 0
[136.144.48.169] setenforce: SELinux is disabled
[136.144.48.169] + true
[136.144.48.169] + sudo sed -i s/SELINUX=enforcing/SELINUX=permissive/g /etc/sysconfig/selinux
[136.144.48.169] sed: can't read /etc/sysconfig/selinux: No such file or directory
ERRO[08:44:19 UTC] failed to install kubeadm: Process exited with status 2  node=136.144.48.169
WARN[08:44:19 UTC] Task failed, error was: at least one of the tasks has encountered an error
...

[136.144.48.169] + sudo sed -i s/SELINUX=enforcing/SELINUX=permissive/g /etc/sysconfig/selinux
[136.144.48.169] + sudo sed -i s/SELINUX=enforcing/SELINUX=permissive/g /etc/selinux/config
[136.144.48.169] sed: can't read /etc/selinux/config: No such file or directory
ERRO[08:49:49 UTC] failed to install kubeadm: Process exited with status 2  node=136.144.48.169
WARN[08:49:49 UTC] Task failed, error was: at least one of the tasks has encountered an error

What is the expected behavior:
kubeone apply works smoothly without touching the instances at all, if files does not exist, it may safely continue in this case because selinux is disabled

How to reproduce the issue:
Get the instances with CentOS 7.7, use the static controlPlane setup and run kubeone apply -m kubeone.yaml

apiVersion: kubeone.io/v1beta1
kind: KubeOneCluster
name: kkp-cluster
versions:
  kubernetes: '1.21.4'
cloudProvider:
  none: {}
machineController:
  deploy: false
apiEndpoint:
  host: '136.144.48.169'
  port: 6443
controlPlane:
  hosts:
    - publicAddress: '136.144.48.169'
      privateAddress: '136.144.48.169'
      sshUsername: root
      sshAgentSocket: 'env:SSH_AUTH_SOCK'

Anything else we need to know?
Can be workaround by creating empty files on the system:

touch /etc/sysconfig/selinux
touch /etc/selinux/config

Information about the environment:
KubeOne version (kubeone version): 1.3.0-alpha.0
Operating system: CentOS

lsb_release -a
LSB Version:	:core-4.1-amd64:core-4.1-noarch
Distributor ID:	CentOS
Description:	CentOS Linux release 7.7.1908 (Core)
Release:	7.7.1908
Codename:	Core

Provider you're deploying cluster on: Equinix bare-metal
Operating system you're deploying on: CentOS

[kron4eg]

Is this some special type of CentOS?

[michalvanco]

Not sure @kron4eg, it's bare metal instance with CentOS 7. We can probably try it out with regular instance on other provider.