Issues Implementing Kubespray on NixOS

[jayrajmeh0]

What happened?

I encountered issues while attempting to implement Kubespray on NixOS. Specifically, the problems are related to apt, which is not available on NixOS. I need to determine if Kubespray supports NixOS or if I need to switch to another Linux distribution. If Kubespray is supported on NixOS, I would like to know the steps I need to follow to resolve the issues.

What did you expect to happen?

1. Successful implementation of Kubespray on NixOS without any package management issues
2. Clear steps or documentation on how to adapt Kubespray for use with NixOS, if possible.

How can we reproduce it (as minimally and precisely as possible)?

1. Attempt to Implement Kubespray:
   - Tried to set up Kubespray on a NixOS system following the standard installation instructions.
2. Encounter Issues:
   - Faced issues related to apt, which is not available in NixOS as it uses a different package management system.

OS

NixOS as the operating system

Version of Ansible

2.16.8

Version of Python

3.10

Version of Kubespray (commit)

v2.25.0

Network plugin used

calico

Full inventory with variables

ansible-playbook -i inventory/mycluster/hosts.yaml --become --become-user=root cluster.yml

Command used to invoke ansible

ansible-playbook -i inventory/mycluster/hosts.yaml --become --become-user=root cluster.yml

Output of ansible run

Anything else we need to know

I would appreciate guidance on the following:
1. Does Kubespray support NixOS? If so, what are the specific steps to successfully implement it on NixOS?
2. If Kubespray does not support NixOS, what alternative Linux distributions are recommended for seamless Kubespray implementation?
Any additional documentation or best practices for using Kubespray with NixOS or suggested alternatives.

[yankay]

HI @jayrajmeh0

The kubespray is not tested on NixOS.
We can use it in Ubuntu or RedHat-based Linux.

BTW, Welcome to help support it.

[jayrajmeh0]

@yankay How can I create an ISO image of a worker node for use during auto-scaling of nodes?