You are browsing a read-only backup copy of Wikitech. The live site can be found at wikitech.wikimedia.org

Puppet coding/testing: Difference between revisions

From Wikitech-static
Jump to navigation Jump to search
imported>Tim Landscheidt
m (Fix syntax highlighting errors.)
imported>Dzahn
(Dzahn moved page Puppet coding/testing to Puppet/testing: requested by Brett )
(10 intermediate revisions by 4 users not shown)
Line 1: Line 1:
{{Draft|This is a work in progress. As of January 25th you will need the puppet patch https://gerrit.wikimedia.org/r/#/c/307223/ to experiment running puppet tests.}}
#REDIRECT [[Puppet/testing]]
 
We have a set of helpers to lint, check style and even test the Puppet code we write.  This part cover how to run checks, how to write your own and debug your tests!
 
==== Running tests ====
Some modules have test suites using the ruby test runner rspec and a set of rake tasks to run linting check (to validate manifests, puppet-lint, hiera yaml files, erb templates). The ruby dependencies are listed in <code>Gemfile</code> at the root of the repository and are installed via [https://bundler.io/ bundler]. To install all dependencies and later run a command in that environment:
 
<source lang="bash">
bundle install
bundle exec <some command>
</source>
 
Assuming a module has a Rakefile and tests defined in a <code>spec</code> sub directory, one can run syntax checks, style and tests via the three commands:
 
<source lang="bash">
bundle exec rake syntax
bundle exec rake puppet-lint
bundle exec rake spec
</source>
 
You can list the various rake tasks via <code>bundle exec rake -T</code> or to get a tree of the task dependencies <code>bundle exec rake -P</code>.
 
==== Rake explained ====
 
The <code>/Gemfile</code> asks for the rubygem [https://rubygems.org/gems/puppetlabs_spec_helper puppetlabs_spec_helper] ([www.rubydoc.info/gems/puppetlabs_spec_helper/ doc]) which contains several predefined rake tasks. Hence in a module one just have to create a Rakefile with:
 
<source lang=ruby>
require 'puppetlabs_spec_helper/rake_tasks'
</source>
 
<source lang=console>
$ cd modules/mymodule
$ bundle exec rake -T
rake beaker                # Run beaker acceptance tests
rake beaker_nodes          # List available beaker nodesets
rake build                # Build puppet module package
rake check:dot_underscore  # Fails if any ._ files are present in directory
rake check:git_ignore      # Fails if directories contain the files specified in .gitignore
rake check:symlinks        # Fails if symlinks are present in directory
rake check:test_file      # Fails if .pp files present in tests folder
rake clean                # Clean a built module package
rake compute_dev_version  # Print development version of module
rake coverage              # Generate code coverage information
rake help                  # Display the list of available rake tasks
rake lint                  # Run puppet-lint
rake release_checks        # Runs all nessesary checks on a module in preparation for a release
rake spec                  # Run spec tests in a clean fixtures directory
rake spec_clean            # Clean up the fixtures directory
rake spec_prep            # Create the fixtures directory
rake spec_standalone      # Run spec tests on an existing fixtures directory
rake syntax                # Syntax check Puppet manifests and templates
rake syntax:hiera          # Syntax check Hiera config files
rake syntax:manifests      # Syntax check Puppet manifests
rake syntax:templates      # Syntax check Puppet templates
rake validate              # Check syntax of Ruby files and call :syntax and :metadata_lint
$
</source>
 
The <code>syntax*</code> tasks come from the [https://rubygems.org/gems/puppet-syntax rubygem puppet-syntax].
 
The <code>spec*</code> tasks are helpers to prepare a puppet environment to run rspec into. Notably adding fixtures and module dependencies for the test environment and tearing down the environment on test completions.
 
==== Writing tests ====
 
To tests puppet resources, we rely on [https://github.com/rodjek/rspec-puppet rspec-puppet] an helper on top of the ruby test runner rspec. rspec-puppet provides utilities to setup puppet, compile a catalog and has built-in assert methods to run against the catalog. The recommendation is to point puppet <code>manifest_dir</code> and <code>module_path</code> to an empty fixture directory that is populated via puppetlabs_spec_helper tasks <code>spec_prep</code> (a prerequisites of the task <code>spec</code>).
 
A minimal case requires a Rakefile, an helper file to be read by each test, and a spec. At first the Rakefile reuses the puppetlabs_spec_helper rake tasks described in the previous section:
 
<code>Rakefile</code>:
<source lang=ruby>
require 'puppetlabs_spec_helper/rake_tasks'
</source>
 
The tests are placed in sub directories of <code>spec/</code> based on the type of Puppet resource being tested. That convention lets rspec-puppet properly setup the rspec helpers for the type of puppet resource being tested. rspec finds tests by crawling the hierachy under <code>spec</code> looking for files with the suffix <code>_spec.rb</code>. The hierarchy is:
 
<pre>
spec/
  ├── applications/
  ├── classes/
  ├──── someclass_spec.rb
  ├── defines/
  ├── functions/
  ├──── afunc_spec.rb
  ├── hosts/
  ├── types/
  └── types_aliases/
</pre>
 
We also need common code to initialize Puppet and point it to a fixture directory, that is where puppetlabs_spec_helper will create a dummy <code>manifests/site.pp</code> and eventually inject additional modules required for tests:
 
Create <code>spec/spec_helper.rb</code>:
<source lang=ruby>
require 'rspec-puppet'
 
fixture_path = File.expand_path(File.join(__FILE__, '..', 'fixtures'))
 
RSpec.configure do |c|
  c.module_path = File.join(fixture_path, 'modules')
  c.manifest_dir = File.join(fixture_path, 'manifests')
end
</source>
 
The file will be required by each of the specs using <source  inline lang=ruby>require 'spec_helper'</source>.
 
 
Given a puppet module <code>mymodule</code> consisting of a single class in <code>manifests/init.pp</code>:
<source lang=ruby>
class mymodule {
}
</source>
 
We first have to instruct rspec-puppet to inject our module in the fixture directory. To do so create a <code>.fixtures.yml</code> at the root of the module (eg: <code>/modules/foobar/.fixtures.yml</code>.
 
<source lang=yaml>
fixtures:
    symlinks:
        mymodule: "#{source_dir}"
</source>
 
The puppet labs spec_helper task <code>spec_prep</code> would process that file and symlink our module as <code>spec/fixtures/modules/mymodule</code> as well as create an empty <code>spec/fixtures/manifests/site.pp</code>.
 
Since we will test a class, we create our test file under under <code>spec/classes/</code> as <code>mymodule_spec.rb</code>:
<source lang=ruby>
# Helper from spec/spec_helper.rb that with the puppet configuration for rspec-puppet
require 'spec_helper'
 
# We will act on the resource "my module"
# Defined as a class resource since the file is under spec/classes
describe 'mymodule' do
  # Check whether puppet can compile the catalog for the 'mymodule' class
  it { should.compile }
end
</source>
 
Finally some fancy configuration of rspec via <code>/.rspec</code>:
<pre>
--format doc
--color
</pre>
 
And we can finally run get the test environment prepared and run the spec:
<pre>
$ bundle exec rake spec
 
mymodule
  should compile into a catalogue without dependency cycles
 
Finished in 0.07349 seconds (files took 0.4312 seconds to load)
1 example, 0 failures
$
</pre>
 
Had we had an error in the manifest, for example a missing curly brace:
<pre>
$ bundle exec rake spec
 
mymodule
  should compile into a catalogue without dependency cycles (FAILED - 1)
 
Failures:
 
  1) mymodule should compile into a catalogue without dependency cycles
    Failure/Error: it { should compile }
      error during compilation:
        Syntax error at end of file; expected '}'
        at modules/mymodule/spec/fixtures/modules/mymodule/manifests/init.pp:2 on node johndoe
    # ./spec/classes/mymodule_spec.rb:4:in `block (2 levels) in <top (required)>'
 
Finished in 0.06745 seconds (files took 0.4104 seconds to load)
1 example, 1 failure
 
Failed examples:
 
rspec ./spec/classes/mymodule_spec.rb:4 # mymodule should compile into a catalogue without dependency cycles
</pre>
 
==== Debugging ====
 
A collection of tips to debug spec failures.
 
===== Puppet debug log =====
Enable puppet debug log to the console. In the <code>spec_helper.rb</code> add:
<source lang=ruby>
if ENV['PUPPET_DEBUG']
  Puppet::Util::Log.level = :debug
  Puppet::Util::Log.newdestination(:console)
</source>
 
Then run tests with <code>PUPPET_DEBUG=1 bundle exec rake spec</code>
 
Credits: [https://gist.github.com/maxlinc/6382696 maxlinc@github gist]).
 
===== Run a single spec / example =====
 
At first prepare the fixture environment:
<source lang="bash">
bundle exec rake spec_prep
</source>
 
Then run in the bundle environment, run rspec on a specific spec:
<source lang="bash">
bundle exec rspec spec/classes/someclass_spec.rb
</source>
 
Or you can filter based on the spec name:
<source lang="bash">
bundle exec rspec --example mymodule::someclass
</source>
 
See rspec help for more details.
 
===== ruby debugger =====
 
 
You can use the gem <code>pry</code> to break on error and get shown a console in the context of the failure.  To your Gemfile add <code>gem 'pry'</code> and install it with <code>bundle install</code> then to break inside a spec:
<source lang=ruby>
require 'spec_helper'
require 'pry'
 
describe 'mymodule::someclass' do
 
  it {
    # enable debugger
    binding.pry
    # compilation that fails:
    should compile
  }
end
</source>
 
You will then be in a console before the breakage that let you inspect the environment (<code>ls</code>). See https://github.com/pry/pry for details.
 
[https://www.mediawiki.org/wiki/Continuous_integration/Entry_points#ruby_debug_tip reference]
 
==== Integration with Jenkins ====
 
''Note: as of January 25th patches have not been merged. See [https://gerrit.wikimedia.org/r/#/q/topic:rspec-puppet+project:operations/puppet+NOT+is:abandoned Gerrit topic:rspec-puppet].''
 
Jenkins job simply runs <code>rake test</code> ([https://www.mediawiki.org/wiki/Continuous_integration/Entry_points#Ruby CI entry point]) from the root of the operations/puppet.git. The checks we want to run automatically are marked as prerequisites of the test task, for example:
<source lang=ruby>
task test: [:rubocop, :puppetlint_head, :syntax_head, :spec]
</source>
 
The tasks suffixed with <code>_head</code> are optimized to have the utility to only run on files changed in the proposed patch. Typically puppet-lint takes minutes to run against all the puppet manifests, when for CI we only are interested in the manifests that are actually being changed.
 
The rakefile add a task for each module having a <code>spec</code> directory. The task is named after the module and put under the namespace <code>spec</code>. Hence as soon as you create a basic structure for a module mymodule, you can run it from the root of the repository with:
<source lang="bash">
bundle exec rake spec:mymodule
</source>
And it is dynamically made a prerequisite of the <code>spec</code> task which is run by CI.  To say it otherwise, once a spec directory is created, Jenkins will try to run the spec.
 
As of January 25, patches are pending consensus.  For the CI integration, at first the job will probably only be run on demand (by commenting <code>check experimental</code> in Gerrit).  Notably the puppet catalog compilation is a bit slow (1.2sec per class) and we should run them in parallel.
 
==== Resources ====
 
* http://rspec-puppet.com/ (might not be up-to-date)
* https://github.com/rodjek/rspec-puppet
* https://github.com/puppetlabs/puppetlabs_spec_helper#puppet-labs-spec-helper (among others: [https://github.com/puppetlabs/puppetlabs_spec_helper#using-fixtures doc about fixtures]).

Revision as of 18:02, 29 June 2022

Redirect to: