Test playbook using molecule (#67)

* Test cluster using molecule

* Fix detection of first control node

* Include --flannel-iface and --node-ip as k3s arguments

* Store logs of k3s-init.service as GitHub job artifacts
This commit is contained in:
Simon Leiner 2022-09-03 17:36:28 +02:00 committed by GitHub
parent 3c36dc8bfd
commit a6b2a95b7e
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
25 changed files with 407 additions and 243 deletions

View File

@ -1,3 +1,16 @@
---
exclude_paths:
# default paths
- '.cache/'
- '.github/'
- 'test/fixtures/formatting-before/'
- 'test/fixtures/formatting-prettier/'
# The "converge" and "reset" playbooks use import_playbook in
# conjunction with the "env" lookup plugin, which lets the
# syntax check of ansible-lint fail.
- '**/converge.yml'
- '**/reset.yml'
skip_list:
- 'fqcn-builtins'

View File

@ -7,24 +7,21 @@ on:
- master
jobs:
vagrant:
name: Vagrant
molecule:
name: Molecule
runs-on: macos-12
strategy:
matrix:
scenario:
- default
env:
HOMEBREW_NO_INSTALL_CLEANUP: 1
VAGRANT_CWD: ${{ github.workspace }}/vagrant
PYTHON_VERSION: "3.10"
steps:
- name: Check out the codebase
uses: actions/checkout@2541b1294d2704b0964813337f33b291d3f8596b # 3.0.2
- name: Install Ansible
run: brew install ansible
- name: Install role dependencies
run: ansible-galaxy install -r collections/requirements.yml
- name: Configure VirtualBox
run: >-
sudo mkdir -p /etc/vbox &&
@ -39,31 +36,29 @@ jobs:
restore-keys: |
vagrant-boxes
- name: Create virtual machines
run: vagrant up
timeout-minutes: 10
- name: Set up Python ${{ env.PYTHON_VERSION }}
uses: actions/setup-python@v2
with:
python-version: ${{ env.PYTHON_VERSION }}
- name: Provision cluster using Ansible
# Since Ansible sets up _all_ machines, it is sufficient to run it only
# once (i.e, for a single node - we are choosing control1 here)
run: vagrant provision control1 --provision-with ansible
timeout-minutes: 25
- name: Install dependencies
run: >-
python3 -m pip install --upgrade pip &&
python3 -m pip install -r requirements.txt
- name: Set up kubectl on the host
run: brew install kubectl &&
mkdir -p ~/.kube &&
vagrant ssh control1 --command "cat ~/.kube/config" > ~/.kube/config
- name: Test with molecule
run: molecule test --scenario-name ${{ matrix.scenario }}
env:
ANSIBLE_K3S_LOG_DIR: ${{ runner.temp }}/logs/k3s-ansible/${{ matrix.scenario }}
- name: Show cluster nodes
run: kubectl describe -A nodes
- name: Show cluster pods
run: kubectl describe -A pods
- name: Test cluster
run: $VAGRANT_CWD/test_cluster.py --verbose --locals
timeout-minutes: 5
- name: Destroy virtual machines
- name: Upload log files
if: always() # do this even if a step before has failed
run: vagrant destroy --force
uses: actions/upload-artifact@3cea5372237819ed00197afe530f5a7ea3e805c8 # 3.1.0
with:
name: logs
path: |
${{ runner.temp }}/logs
- name: Delete old box versions
if: always() # do this even if a step before has failed
run: vagrant box prune --force

2
.gitignore vendored
View File

@ -1 +1 @@
.vagrant
.env/

View File

@ -100,12 +100,12 @@ See the commands [here](https://docs.technotim.live/posts/k3s-etcd-ansible/#test
Be sure to see [this post](https://github.com/techno-tim/k3s-ansible/discussions/20) on how to troubleshoot common problems
### 🔷 Vagrant
### Testing the playbook using molecule
You may want to kickstart your k3s cluster by using Vagrant to quickly build you all needed VMs with one command.
Head to the `vagrant` subfolder and type `vagrant up` to get your environment setup.
After the VMs have got build, deploy k3s using the Ansible playbook `site.yml` by the
`vagrant provision --provision-with ansible` command.
This playbook includes a [molecule](https://molecule.rtfd.io/)-based test setup.
It is run automatically in CI, but you can also run the tests locally.
This might be helpful for quick feedback in a few cases.
You can find more information about it [here](molecule/README.md).
## Thanks 🤝

View File

@ -2,3 +2,4 @@
collections:
- name: community.general
- name: ansible.posix
- name: kubernetes.core

View File

@ -17,9 +17,23 @@ apiserver_endpoint: "192.168.30.222"
# this token should be alpha numeric only
k3s_token: "some-SUPER-DEDEUPER-secret-password"
# change these to your liking, the only required one is--disable servicelb
extra_server_args: "--disable servicelb --disable traefik"
extra_agent_args: ""
# The IP on which the node is reachable in the cluster.
# Here, a sensible default is provided, you can still override
# it for each of your hosts, though.
k3s_node_ip: '{{ ansible_facts[flannel_iface]["ipv4"]["address"] }}'
# these arguments are recommended for servers as well as agents:
extra_args: >-
--flannel-iface={{ flannel_iface }}
--node-ip={{ k3s_node_ip }}
# change these to your liking, the only required one is --disable servicelb
extra_server_args: >-
{{ extra_args }}
--disable servicelb
--disable traefik
extra_agent_args: >-
{{ extra_args }}
# image tag for kube-vip
kube_vip_tag_version: "v0.5.0"

67
molecule/README.md Normal file
View File

@ -0,0 +1,67 @@
# Test suites for `k3s-ansible`
This folder contains the [molecule](https://molecule.rtfd.io/)-based test setup for this playbook.
## Scenarios
We have these scenarios:
- **default**:
A 3 control + 2 worker node cluster based very closely on the [sample inventory](../inventory/sample/).
## How to execute
To test on your local machine, follow these steps:
### System requirements
Make sure that the following software packages are available on your system:
- [Python 3](https://www.python.org/downloads)
- [Vagrant](https://www.vagrantup.com/downloads)
- [VirtualBox](https://www.virtualbox.org/wiki/Downloads)
### Set up VirtualBox networking on Linux and macOS
_You can safely skip this if you are working on Windows._
Furthermore, the test cluster uses the `192.168.30.0/24` subnet which is [not set up by VirtualBox automatically](https://www.virtualbox.org/manual/ch06.html#network_hostonly).
To set the subnet up for use with VirtualBox, please make sure that `/etc/vbox/networks.conf` exists and that it contains this line:
```
* 192.168.30.0/24`
```
### Install Python dependencies
You will get [Molecule, Ansible and a few extra dependencies](../requirements.txt) via [pip](https://pip.pypa.io/).
Usually, it is advisable to work in a [virtual environment](https://docs.python.org/3/tutorial/venv.html) for this:
```bash
cd /path/to/k3s-ansible
# Create a virtualenv at ".env". You only need to do this once.
python3 -m venv .env
# Activate the virtualenv for your current shell session.
# If you start a new session, you will have to repeat this.
source .env/bin/activate
# Install the required packages into the virtualenv.
# These remain installed across shell sessions.
python3 -m pip install -r requirements.txt
```
### Run molecule
With the virtual environment from the previous step active in your shell session, you can now use molecule to test the playbook.
Interesting commands are:
- `molecule create`: Create virtual machines for the test cluster nodes.
- `molecule destroy`: Delete the virtual machines for the test cluster nodes.
- `molecule converge`: Run the `site` playbook on the nodes of the test cluster.
- `molecule side_effect`: Run the `reset` playbook on the nodes of the test cluster.
- `molecule verify`: Verify that the cluster works correctly.
- `molecule test`: The "all-in-one" sequence of steps that is executed in CI.
This includes the `create`, `converge`, `verify`, `side_effect` and `destroy` steps.
See [`molecule.yml`](default/molecule.yml) for more details.

View File

@ -0,0 +1,75 @@
---
dependency:
name: galaxy
driver:
name: vagrant
platforms:
- &control
name: control1
box: generic/ubuntu2204
memory: 2048
cpus: 2
config_options:
# We currently can not use public-key based authentication on Ubuntu 22.04,
# see: https://github.com/chef/bento/issues/1405
ssh.username: "vagrant"
ssh.password: "vagrant"
groups:
- k3s_cluster
- master
interfaces:
- network_name: private_network
ip: 192.168.30.38
- <<: *control
name: control2
interfaces:
- network_name: private_network
ip: 192.168.30.39
- <<: *control
name: control3
interfaces:
- network_name: private_network
ip: 192.168.30.40
- &node
<<: *control
name: node1
groups:
- k3s_cluster
- node
interfaces:
- network_name: private_network
ip: 192.168.30.41
- <<: *node
name: node2
interfaces:
- network_name: private_network
ip: 192.168.30.42
provisioner:
name: ansible
playbooks:
converge: ../resources/converge.yml
side_effect: ../resources/reset.yml
verify: ../resources/verify.yml
inventory:
links:
group_vars: ../../inventory/sample/group_vars
env:
OVERRIDES_FILE: ../default/overrides.yml
scenario:
test_sequence:
- dependency
- lint
- cleanup
- destroy
- syntax
- create
- prepare
- converge
# idempotence is not possible with the playbook in its current form.
- verify
# We are repurposing side_effect here to test the reset playbook.
# This is why we do not run it before verify (which tests the cluster),
# but after the verify step.
- side_effect
- cleanup
- destroy

View File

@ -0,0 +1,11 @@
---
- name: Apply overrides
hosts: all
tasks:
- name: Override host variables
ansible.builtin.set_fact:
# See: https://github.com/flannel-io/flannel/blob/67d603aaf45ef80f5dd39f43714fc5e6f8a637eb/Documentation/troubleshooting.md#Vagrant # noqa yaml[line-length]
flannel_iface: eth1
# The test VMs might be a bit slow, so we give them more time to join the cluster:
retry_count: 45

View File

@ -0,0 +1,7 @@
---
- name: Apply overrides
ansible.builtin.import_playbook: >-
{{ lookup("ansible.builtin.env", "OVERRIDES_FILE") }}
- name: Converge
ansible.builtin.import_playbook: ../../site.yml

View File

@ -0,0 +1,7 @@
---
- name: Apply overrides
ansible.builtin.import_playbook: >-
{{ lookup("ansible.builtin.env", "OVERRIDES_FILE") }}
- name: Reset
ansible.builtin.import_playbook: ../../reset.yml

View File

@ -0,0 +1,5 @@
---
- name: Verify
hosts: all
roles:
- verify/from_outside

View File

@ -0,0 +1,9 @@
---
# A host outside of the cluster from which the checks shall be performed
outside_host: localhost
# This kubernetes namespace will be used for testing
testing_namespace: molecule-verify-from-outside
# The directory in which the example manifests reside
example_manifests_path: ../../../../example

View File

@ -0,0 +1,5 @@
---
- name: Clean up kubecfg
ansible.builtin.file:
path: "{{ kubecfg.path }}"
state: absent

View File

@ -0,0 +1,19 @@
---
- name: Create temporary directory for kubecfg
ansible.builtin.tempfile:
state: directory
suffix: kubecfg
register: kubecfg
- name: Gathering facts
delegate_to: "{{ groups['master'][0] }}"
ansible.builtin.gather_facts:
- name: Download kubecfg
ansible.builtin.fetch:
src: "{{ ansible_env.HOME }}/.kube/config"
dest: "{{ kubecfg.path }}/"
flat: true
delegate_to: "{{ groups['master'][0] }}"
delegate_facts: true
- name: Store path to kubecfg
ansible.builtin.set_fact:
kubecfg_path: "{{ kubecfg.path }}/config"

View File

@ -0,0 +1,12 @@
---
- name: Verify
run_once: true
delegate_to: "{{ outside_host }}"
block:
- ansible.builtin.import_tasks: kubecfg-fetch.yml
- name: "TEST CASE: Get nodes"
ansible.builtin.include_tasks: test/get-nodes.yml
- name: "TEST CASE: Deploy example"
ansible.builtin.include_tasks: test/deploy-example.yml
always:
- ansible.builtin.import_tasks: kubecfg-cleanup.yml

View File

@ -0,0 +1,54 @@
---
- name: Deploy example
block:
- name: "Create namespace: {{ testing_namespace }}"
kubernetes.core.k8s:
api_version: v1
kind: Namespace
name: "{{ testing_namespace }}"
state: present
wait: true
kubeconfig: "{{ kubecfg_path }}"
- name: Apply example manifests
kubernetes.core.k8s:
src: "{{ example_manifests_path }}/{{ item }}"
namespace: "{{ testing_namespace }}"
state: present
wait: true
kubeconfig: "{{ kubecfg_path }}"
with_items:
- deployment.yml
- service.yml
- name: Get info about nginx service
kubernetes.core.k8s_info:
kind: service
name: nginx
namespace: "{{ testing_namespace }}"
kubeconfig: "{{ kubecfg_path }}"
vars: &load_balancer_metadata
metallb_ip: status.loadBalancer.ingress[0].ip
metallb_port: spec.ports[0].port
register: nginx_services
- name: Assert that the nginx welcome page is available
ansible.builtin.uri:
url: http://{{ ip }}:{{ port }}/
return_content: yes
register: result
failed_when: "'Welcome to nginx!' not in result.content"
vars:
ip: >-
{{ nginx_services.resources[0].status.loadBalancer.ingress[0].ip }}
port: >-
{{ nginx_services.resources[0].spec.ports[0].port }}
always:
- name: "Remove namespace: {{ testing_namespace }}"
kubernetes.core.k8s:
api_version: v1
kind: Namespace
name: "{{ testing_namespace }}"
state: absent
kubeconfig: "{{ kubecfg_path }}"

View File

@ -0,0 +1,17 @@
---
- name: Get all nodes in cluster
kubernetes.core.k8s_info:
kind: node
kubeconfig: "{{ kubecfg_path }}"
register: cluster_nodes
- name: Assert that the cluster contains exactly the expected nodes
ansible.builtin.assert:
that: found_nodes == expected_nodes
success_msg: "Found nodes as expected: {{ found_nodes }}"
fail_msg: "Expected nodes {{ expected_nodes }}, but found nodes {{ found_nodes }}"
vars:
found_nodes: >-
{{ cluster_nodes | json_query('resources[*].metadata.name') | unique }}
expected_nodes: >-
{{ (groups['master'] + groups['node']) | unique }}

7
requirements.txt Normal file
View File

@ -0,0 +1,7 @@
ansible-core>=2.13.2
jmespath
jsonpatch
kubernetes>=12.0.0
molecule-vagrant>=1.0.0
molecule>=4.0.1
pyyaml>=3.11

View File

@ -2,10 +2,10 @@
ansible_user: root
server_init_args: >-
{% if groups['master'] | length > 1 %}
{% if ansible_host == hostvars[groups['master'][0]]['ansible_host'] | default(groups['master'][0]) %}
{% if ansible_hostname == hostvars[groups['master'][0]]['ansible_hostname'] %}
--cluster-init
{% else %}
--server https://{{ hostvars[groups['master'][0]]['ansible_host'] | default(groups['master'][0]) }}:6443
--server https://{{ hostvars[groups['master'][0]].k3s_node_ip }}:6443
{% endif %}
--token {{ k3s_token }}
{% endif %}

View File

@ -0,0 +1,28 @@
---
# Download logs of k3s-init.service from the nodes to localhost.
# Note that log_destination must be set.
- name: Fetch k3s-init.service logs
ansible.builtin.command:
cmd: journalctl --all --unit=k3s-init.service
changed_when: false
register: k3s_init_log
- name: Create {{ log_destination }}
delegate_to: localhost
run_once: true
become: false
ansible.builtin.file:
path: "{{ log_destination }}"
state: directory
mode: "0755"
- name: Store logs to {{ log_destination }}
delegate_to: localhost
become: false
ansible.builtin.template:
src: content.j2
dest: "{{ log_destination }}/k3s-init@{{ ansible_hostname }}.log"
mode: 0644
vars:
content: "{{ k3s_init_log.stdout }}"

View File

@ -20,7 +20,7 @@
owner: root
group: root
mode: 0644
when: ansible_host == hostvars[groups['master'][0]]['ansible_host'] | default(groups['master'][0])
when: ansible_hostname == hostvars[groups['master'][0]]['ansible_hostname']
- name: Copy vip rbac manifest to first master
template:
@ -29,7 +29,7 @@
owner: root
group: root
mode: 0644
when: ansible_host == hostvars[groups['master'][0]]['ansible_host'] | default(groups['master'][0])
when: ansible_hostname == hostvars[groups['master'][0]]['ansible_hostname']
- name: Copy vip manifest to first master
template:
@ -38,7 +38,7 @@
owner: root
group: root
mode: 0644
when: ansible_host == hostvars[groups['master'][0]]['ansible_host'] | default(groups['master'][0])
when: ansible_hostname == hostvars[groups['master'][0]]['ansible_hostname']
# these will be copied and installed now, then tested later and apply config
- name: Copy metallb namespace to first master
@ -48,7 +48,7 @@
owner: root
group: root
mode: 0644
when: ansible_host == hostvars[groups['master'][0]]['ansible_host'] | default(groups['master'][0])
when: ansible_hostname == hostvars[groups['master'][0]]['ansible_hostname']
- name: Copy metallb namespace to first master
template:
@ -57,7 +57,7 @@
owner: root
group: root
mode: 0644
when: ansible_host == hostvars[groups['master'][0]]['ansible_host'] | default(groups['master'][0])
when: ansible_hostname == hostvars[groups['master'][0]]['ansible_hostname']
- name: Init cluster inside the transient k3s-init service
command:
@ -80,6 +80,12 @@
delay: 10
changed_when: false
always:
- name: Save logs of k3s-init.service
include_tasks: fetch_k3s_init_logs.yml
when: log_destination
vars:
log_destination: >-
{{ lookup('ansible.builtin.env', 'ANSIBLE_K3S_LOG_DIR', default=False) }}
- name: Kill the temporary service used for initialization
systemd:
name: k3s-init

View File

@ -0,0 +1,5 @@
{#
This is a really simple template that just outputs the
value of the "content" variable.
#}
{{ content }}

79
vagrant/Vagrantfile vendored
View File

@ -1,79 +0,0 @@
# -*- mode: ruby -*-
# vi: set ft=ruby :
Vagrant.configure("2") do |config|
# General configuration
config.vm.box = "generic/ubuntu2204"
config.vm.synced_folder ".", "/vagrant", disabled: true
config.ssh.insert_key = false
config.vm.provider :virtualbox do |v|
v.memory = 2048
v.cpus = 2
v.linked_clone = true
end
# Control Node 1
config.vm.define "control1" do |control1|
control1.vm.hostname = "control1"
control1.vm.network "private_network", ip: "192.168.30.38"
end
# Control Node 2
config.vm.define "control2" do |control2|
control2.vm.hostname = "control2"
control2.vm.network "private_network", ip: "192.168.30.39"
end
# Control Node 3
config.vm.define "control3" do |control3|
control3.vm.hostname = "control3"
control3.vm.network "private_network", ip: "192.168.30.40"
end
# Worker Node 1
config.vm.define "node1" do |node1|
node1.vm.hostname = "node1"
node1.vm.network "private_network", ip: "192.168.30.41"
end
# Worker Node 2
config.vm.define "node2" do |node2|
node2.vm.hostname = "node2"
node2.vm.network "private_network", ip: "192.168.30.42"
end
config.vm.provision "ansible",type: "ansible", run: "never" do |ansible|
ansible.playbook = "../site.yml"
ansible.limit = "all"
ansible.groups = {
"master" => ["control1", "control2", "control3"],
"node" => ["node1", "node2"],
"k3s_cluster:children" => ["master", "node"],
"k3s_cluster:vars" => {"k3s_version" => "v1.24.4+k3s1",
"ansible_user" => "vagrant",
"systemd_dir" => "/etc/systemd/system",
"flannel_iface" => "eth1",
"apiserver_endpoint" => "192.168.30.222",
"k3s_token" => "supersecret",
"extra_server_args" => "--node-ip={{ ansible_eth1.ipv4.address }} --flannel-iface={{ flannel_iface }} --no-deploy servicelb --no-deploy traefik",
"extra_agent_args" => "--flannel-iface={{ flannel_iface }}",
"kube_vip_tag_version" => "v0.5.0",
"metal_lb_speaker_tag_version" => "v0.13.4",
"metal_lb_controller_tag_version" => "v0.13.4",
"metal_lb_ip_range" => "192.168.30.80-192.168.30.90",
"retry_count" => "60"}
}
ansible.host_vars = {
"control1" => {
"server_init_args" => "--cluster-init --token {{ k3s_token }} {{ extra_server_args | default('') }}"
},
"control2" => {
"server_init_args" => "--server https://192.168.30.38:6443 --token {{ k3s_token }} {{ extra_server_args | default('') }}"
},
"control3" => {
"server_init_args" => "--server https://192.168.30.38:6443 --token {{ k3s_token }} {{ extra_server_args | default('') }}"
}
}
end
end

View File

@ -1,114 +0,0 @@
#!/usr/bin/env python3
# Perform a few tests on a cluster created with this playbook.
# To simplify test execution, the scripts does not depend on any third-party
# packages, only the Python standard library.
import json
import subprocess
import unittest
from pathlib import Path
from time import sleep
from warnings import warn
VAGRANT_DIR = Path(__file__).parent.absolute()
PLAYBOOK_DIR = VAGRANT_DIR.parent.absolute()
class TestK3sCluster(unittest.TestCase):
def _kubectl(self, args: str, json_out: bool = True) -> dict | None:
cmd = "kubectl"
if json_out:
cmd += " -o json"
cmd += f" {args}"
result = subprocess.run(cmd, capture_output=True, shell=True, check=True)
if json_out:
return json.loads(result.stdout)
else:
return None
def _curl(self, url: str) -> str:
options = [
"--silent", # no progress info
"--show-error", # ... but errors should still be shown
"--fail", # set exit code on error
"--location", # follow redirects
]
cmd = f'curl {" ".join(options)} "{url}"'
result = subprocess.run(cmd, capture_output=True, shell=True, check=True)
output = result.stdout.decode("utf-8")
return output
def _apply_manifest(self, manifest_file: Path) -> dict:
apply_result = self._kubectl(
f'apply --filename="{manifest_file}" --cascade="background"'
)
self.addCleanup(
lambda: self._kubectl(
f'delete --filename="{manifest_file}"',
json_out=False,
)
)
return apply_result
@staticmethod
def _retry(function, retries: int = 5, seconds_between_retries=1):
for retry in range(1, retries + 1):
try:
return function()
except Exception as exc:
if retry < retries:
sleep(seconds_between_retries)
continue
else:
raise exc
def _get_load_balancer_ip(
self,
service: str,
namespace: str = "default",
) -> str | None:
svc_description = self._kubectl(
f'get --namespace="{namespace}" service "{service}"'
)
ip = svc_description["status"]["loadBalancer"]["ingress"][0]["ip"]
return ip
def test_nodes_exist(self):
out = self._kubectl("get nodes")
node_names = {item["metadata"]["name"] for item in out["items"]}
self.assertEqual(
node_names,
{"control1", "control2", "control3", "node1", "node2"},
)
def test_ip_address_pool_exists(self):
out = self._kubectl("get --all-namespaces IpAddressPool")
pools = out["items"]
self.assertGreater(len(pools), 0)
def test_nginx_example_page(self):
# Deploy the manifests to the cluster
deployment = self._apply_manifest(PLAYBOOK_DIR / "example" / "deployment.yml")
service = self._apply_manifest(PLAYBOOK_DIR / "example" / "service.yml")
# Assert that the dummy page is available
metallb_ip = self._retry(
lambda: self._get_load_balancer_ip(service["metadata"]["name"])
)
# Now that an IP address was assigned, let's reload the service description:
service = self._kubectl(f'get service "{service["metadata"]["name"]}"')
metallb_port = service["spec"]["ports"][0]["port"]
response_body = self._retry(
lambda: self._curl(f"http://{metallb_ip}:{metallb_port}/")
)
self.assertIn("Welcome to nginx!", response_body)
if __name__ == "__main__":
unittest.main()