pre_post_checks.md

Pre and post flight testing

The module as been updated to be supported by Puppetlabs Litmus.

The module ServerSpec tests for the Bolt Tasks and Plans provided can be used agains a deployment infrastrcuture run as Post flight checks if desired.

Additionally some stand a lone preflight checks will be added presently.

Setup

The module as been converted to Litmus as the directions found at https://puppetlabs.github.io/litmus/Converting-modules-to-use-Litmus.html that means the following files have been added.

\.
├── spec
│   │   
│   ├── spec_helper_acceptance_local.rb
│   ├── spec_helper.rb
│   └── spec_helper_acceptance.rb
└── provision.yaml

An update to .gitignore for the Litmus generated .rerun.json has also been made via .sync.yaml PDK functionality.

Usage

Litmus can provision local testing via provisioning and generation of a Bolt inventory.yaml, see https://puppetlabs.github.io/litmus/Running-acceptance-tests.html for an example.

Inventory

when testing locally with Vagrant or VMpooler you can use the litmus:provision rake task to generate an inventory.yml.

You will normally need to create an inventory.yaml for your target puppet infrastructure, and you may want to group your puppet infrastructure related to the Puppet Console classification node groups. A possible example inventory.yaml is illustrated.

Note that if you are using Litmus against a host once the agent is installed (not for pre deployment checks of peadm) you will want to add features: ['puppet-agent'] to your inventory.yaml this resolves several error messages otherwise encountered.

---
groups:
- name: peserver
  nodes:
  - primary.puppet.example.net
  features: ['puppet-agent']
  config:
    transport: ssh
    ssh:
      host-key-check: false
      user: centos
      run-as: root
      private-key: "~/.ssh/example.pem"
- name: "compilers"
  nodes:
  - compiler00.puppet.example.net
  features: ['puppet-agent']
  config:
    transport: ssh
    ssh:
      host-key-check: false
      user: centos
      run-as: root
      private-key: "~/.ssh/example.pem"
- name: ha
  nodes:
  - ha-primary.puppet.example.net
  features: ['puppet-agent']  
  config:
    transport: ssh
    ssh:
      host-key-check: false
      user: centos
      run-as: root
      private-key: "~/.ssh/example.pem"
- name: windowsagents
  nodes:
  - win0.example.net
  features: ['puppet-agent']
  config:
    transport: winrm
    winrm:
      user: domainadminaccount
      password: "@example"
      ssl: false
- name: linuxagents
  nodes:
  - nix0.example.net
  features: ['puppet-agent']
  config:
    transport: ssh
    ssh:
      host-key-check: false
      user: centos
      run-as: root
      private-key: "~/.ssh/example.pem"

Tests and Checks

One you have the module deployed any Pre or Post deployment checks can be developed as standard ServerSpec tests and orchestrated using Bolt and Litmus.

A default check exists in peadm_spec.rb, note the test is constrained using the os[:family] fact. The example test simply prints it's context and will not fail.

spec
├── acceptance
    └── peadm_spec.rb


require 'spec_helper_acceptance'
# @summary: default test does nothing
def test_peadm()
  
    #return unless os[:family] != 'windows'
    return unless os[:family] != 'Darwin'
end

describe 'default' do
  context 'example acceptance do nothing' do
    it do
        test_peadm()
    end
  end
end

For running the tests review the standard usage of Litmusfor installing and running tests https://puppetlabs.github.io/litmus/Running-acceptance-tests.html#6-run-the-motd-acceptance-tests remember we are doing these actions from within the context of the PDK pdk bundle exec rake has several sub commands.

rake litmus:acceptance:<target || nodes> 
# Run serverspec against targets:name 
rake litmus:acceptance:localhost                                                # Run serverspec against localhost, USE WITH CAUTION, this action can be potentially dangerous
rake litmus:acceptance:parallel                                                 # Run tests in parallel against all machines in the inventory file
rake litmus:acceptance:serial                                                   # Run tests in serial against all machines in the inventory file
rake litmus:install_agent[collection,target_node_name]                          # install puppet agent, [:collection, :target_node_name]
rake litmus:install_module[target_node_name]                                    # install_module - build and install module
rake litmus:install_modules_from_directory[source,target_node_name]             # install_module - build and install module
rake litmus:metadata                                                            # print all supported OSes from metadata
rake litmus:provision[provisioner,platform,inventory_vars]                      # provision container/VM - abs/docker/vagrant/vmpooler eg 'bundle exec rake 'litmus:provision[vmpooler, ubuntu-160...
rake litmus:provision_install[key,collection]                                   # provision_install - provision a list of machines, install an agent, and the module
rake litmus:provision_list[key]                                                 # provision list of machines from provision.yaml file
rake litmus:reinstall_module[target_node_name]                                  # reinstall_module - reinstall module
rake litmus:tear_down[target]                                                   # tear-down - decommission machines
rake litmus:uninstall_module[target_node_name,module_name]                      # uninstall_module - uninstall module