cct - Cucumber Cloud Testsuite

This is the testsuite for SUSE Openstack Cloud.

The master branch contains the testsuite for yet to be released version 6 of SOC. The testsuite for other versions of cloud are available in respective branches.

Topics

Installation
Dependencies
Usage
How-To's
Configuration
Features

Quick start

1. Check system dependencies
2. git clone [email protected]:suse-cloud/cct
3. cd cct && bundle install
4. Add missing data into config/development.yml file (create one by copying config/development.yml.example file)
5. rake feature:admin

Dependencies

zypper in rubygem-bundler ruby-devel
zypper in gcc make

List of required rubygems can be found in file cct.gemspec and in Gemfile.

Installation

Make sure you have installed all dependencies

After you have cloned the repository, install the required rubygems:

 bundle install

All rubygems will be installed into the directory vendor/bundle within the repo.

This is the default setup, if you use one of Ruby version managers like RVM, you might want to create a new gemset and override the bundle config in .bundle/config with:

 bundle install --system

Usage

rake is one way how to run the tests and everything else. To get a list of commands you can use, type:

 bundle exec rake help

To get rid of the annoying rake command prefixing with bundle exec, having done

 alias rake="bundle exec rake"

might be useful. The command bundle exec takes care about locating and loading the correct rubygems from the pre-configured path as set by .bundle/config.

If you get an error like

rake aborted!
LoadError: cannot load such file -- cucumber
/home/path/to/code/cct/Rakefile:6:in `<top (required)>'

the rubygems installed in path vendor/bundle are not visible to rake.

All files with tasks are located in the directory tasks/ and have suffix .rake All are loaded automatically.

A single file usually contains several tasks within some well chosen namespace.

If you don't like rake you can still use the plain cucumber command. Typing bundle exec cucumber will run all features. If you want to limit the testsuite to some specific feature, put the path to the feature file after the command, like cucumber features/admin_node.feature. In case you want to limit the test run to a one or more scenarios, you must specify tags for those scenarios: cucumber features/admin_node.feature -t @services,@packages.

Useful commands

Get some help:

 rake help
 rake h
 rake -T
 rake -T keyword

Run unit tests for code inside lib directory:

 rake spec

How-To's

Make use of rake console

Once you have added your configuration for the admin node, it's handy to get some live data from the deployed cloud in an ruby's irb session.

rake console will give you access to admin node, control node, crowbar API and some other useful stuff used in the cucumber step definitions.

To make it more verbose, call it with rake console --verbose

Write a new feature

There is a dedicated rake task available for creating all the necessaries:

rake add:feature name='Awesome Stuff' task='awesome_stuff'  

Try to keep the name of the rake task short but still expressive. Multiple words must be combined with underscore.

Two new files will be created for a new feature:

* new cucumber feature file in path `features/` that must have `.feature` extension  
* new rake task file in path `tasks/` with `.rake` extension  

Look into the directory templates/ where the templates for both files are stored.

Write a new scenario

Every cucumber feature is composed of one or more scenarios. A scenaro consists of its name and steps that describe what is being tested.

A scenario is a concept based on Givens, Whens and Thens.

The purpose of Given is to put the system in a known state or to find out whether the system is in state ready for testing.

The purpose of When is to describe the key action the actor (system or user) performs.

The purpose of Then step is to observe outcomes. The observations should be related to the feature description.

Beside these you can use And and But keywords to make the scenario more descriptive or specific.

Don't forget to add a tag above the scenario name in the form of a single word that expresses the scenario name, like @packages . This tag will be used inside of a feature rake task to call that single scenario only.

If you are still not sure how to write a scenario, better look at the existing features and make your own judgment.

Use scenario configuration

If you find yourself writing similar scenarios with different configurations, you might want to make use of scenario specific config values. You can put these into config/features/FEATURE_NAME.yml

This file must have structure like this:

features:
  admin:
    ntp:
      key_one: abcd
      key_two: efcg

You can access these config values either manually by config["features"]["admin"]["ntp"]["key_one"] or by using a method with block:

# Put this into some step definition block
with_scenario_config do |config|
  puts config["key_one"]
  puts config["key_two"]
end

If the configuration for this scenario is empty, the code within the block will be not executed. (Maybe raising an exception like ConfigurationNotFound would be better?).

This configuration is loaded in a Before hook in features/support/env.rb and is based on the tags used for the feature and scenario.

It works only when a feature has a single tag. At the scenario level the first tag is used, the rest is ignored.

Write a step definition

Once the scenario is written, you should adapt the respective feature rake task to let it run via rake (more on this here).

When the steps does not match any already existing step definition the scenario will be marked as undefined with this notice:

You can implement step definitions for undefined steps with these snippets:

Given(/^I have my system in the okay state$/) do
  pending # Write code here that turns the phrase above into concrete actions
end

Now you copy the text above starting with Given, go to the directory features/step_definitions/YOUR_FEATURE, create a new file with name matching the scenario name and with suffix _steps.rb. Insert the copied text from above.

The last step is to replace the pending method with your test implementation. More about that you can find in the section about commands in step definitions .

Run a single scenario with rake

To be able to run a single scenario, you need to mark it with a unique @tag. Then in rake you must implement a specific task that will call the feature with that @tag. A rake task for a feature might look like this:

namespace :feature do
  feature_name "Admin node"

  namespace :admin do
    desc "Test NTP Server availability"
    feature_task :ntp, tags: :@ntp
  end
end

It looks up the feature with name Admin node, creates a rake task with name :ntp within the namespace :admin. To call this task you need to type:

rake feature:admin:ntp

Run a complete feature with rake

There needs to be a standalone rake task implemented to make a feature run complete. It's done by a separate task without description like this:

namespace :feature do
  feature_name "Controller node"

  namespace :controller do
    desc "Essential system requirements"
      feature_task :system, tags: :@system
      feature_task :all # Define it here like this
    end

    # One task to rule them all
    desc "Complete verification of 'Controller node' feature"
    task :controller => "controller:all"
  end

Now call the whole feature with:

  rake feature:controller

We created a new task without any specific @tags, that means everything what is found within the feature is run.

What commands to use in step definitions

The ruby commands available in the step definitions are defined by code in the file features/support/env.rb :

World do
  Cct::Cloud::World.new
end

The actual ruby code is available in lib/cct/cloud/world.rb

Helper commands are also loaded by this:

World(
  StepHelpers
  FeatureHelpers
  ...
)

The code is available in features/support/*_helpers.rb. There should be only step or scenario specific code placed here, the right place for new commands is in lib/cct/commands/.

The main command available for executing local or remote bash is exec!. You need to call it on the respective node object to let the code run remotely:

admin_node.exec! "crowbar ntp list"
control_node.exec! "cat", ".openrc"

The exec! command is available even for local code execution, it's what it does when you call it without any receiver.

There are several useful commands implemented for common use, you will find them at lib/cct/commands/remote.rb :

There are available, for example, these (and more to come):

read_file expects a path on the remote machine as parameter
rpm_q expects a package name as parameter

exec! commands returns a struct object with three attributes: output, success? and exit_code.

If the execution of the exec! command has failed, an exception is thrown. You should not rescue these exceptions as it's important to see and log the source of the primary problem.

Additionally, you can use built-in matchers of rspec unit test framework. There is a big variety of the matchers for various situations, you rather look at them at rspec matchers docs . Here is a bit more detailed doc about how to customize the matchers.

Openstack client commands

For testing the cloud functionality we use the python-openstackclient interface, more about that please look at OpenStackClient docs.

These commands is available at the control_node object:

control_node.openstack.user.list          # returns list of users
control_node.openstack.image.list         # returns list of images
control_node.openstack.image.show "jeos1" # returns object with properties about the image

The recommended way of making yourself familiar with those is rake console.

Please, don't use the specialized openstack components' commands (like keystone role-list or glance image-create) as these are going to be deprecated in some of the coming releases of Openstack.

Add new command for the step definitions

The rule of thumb is when you use a remote or local command on more than 3 places, let's it implement as a predefined command within the lib/cct/commands. Don't forget to add unit test to the new command.

Add new code to lib/

If you find a bug or you want to propose an improvement, please create a pull request. Every new code for the lib/ directory must have rspec unit tests.

Configuration

Show the current configuration:

 rake config

Default config files

By default the testsuite contains two configuration files:

 config/default.yml
 config/development.yml.example

The first one, config/default.yml contains general information about the product, used by the features and test cases in the testsuite. You don't alter this data for your custom configuration. If you think they need to be changed, create a pull request.

The second one, config/development.yml.example is a template configuration file you might use for testing your cloud instance. You may uncomment and use the template data, however they might not match your expactations. Feel free to change the data as needed to make the testsuite detect your cloud:

* ip address for the admin node
* password for the admin node HTTP API
* SSH password for the admin node unless you use key based authentication

All configuration files but the default.yml file are ignored by git.

Proxy configuration

If your cloud instance is located on some remote machine behind firewall, you can run tests there by updating the proxy section of the configuration file. The development.yml.example file contains an example of how to do that. However, it works only if you have set up SSH keys with that remote domain.

Nodes' configuration

The development.yml config file contains this section:

nodes:
  - name:
    ssh:
      user: root
      password:

Here you can configure the ssh credentials for the respective nodes (beside admin node which is configured in its own section).

If you want to add a configuration for some specific node, please copy the whole section beggining with name and keep the default one untouched.

At least one section should keep the name attribute empty, this setup is considered as default and will be used for the rest of nodes who share the same configuration for ssh part.

Usually the only thing you might want to change is the password attribute. For the name attribute you can use either a hostname or a FQDN.

Custom config files

If you think you might need to add more configuration files, add them into the config/ directory, they will be ignored by git. To make the testsuite load them for you, add this line to the config/development.yml file:

autoload_config_files:
 - your_file.yml

The config files are loaded in the order as specified in the list. Don't forget that if you have the same sections in your config files, the last loaded file wins.

If the file you added does not exist, the testsuite will fail.

Custom config data as json or yaml

Additionally, you can provide configuration data in json or yaml format to every rake command like this:

rake feature:admin cct_config='{"admin_node":{"remote":{"ip":"192.168.199:10"}}}'

These are merged with the data after all configuration files have been loaded.

Turn off the fancy colors!

If you don't like the colored cucumber output or for some other reason you want to have it off, provide a bash variable nocolor to the rake command:

rake feature:admin nocolor=true

Features

The testsuite is powered by cucumber which is a tool for running automated tests written in plain language. More information is available at the Cucumber Github Wiki.

Feature files

All features are stored in files with .feature suffix in the directory features/. Every file describes a single feature with one or more scenarios using Gherkin. It is a Domain Specific Language that lets you describe behaviour without detailing how that behaviour is implemented.

Run a feature!

The command line interface is powered by rake which is a task management tool, more information is available at the rake Github repo.

Feature tasks begin with feature keyword, to test only a particular scenario you need to run some specific rake task:

rake feature:admin       # Run all scenarios for admin node
rake feature:admin:ntp   # Test NTP Server availability on admin node
rake feature:admin:os    # Check operating system support for admin node

Disclaimer

This file contains information and formulations as found on cucumber github wiki.

Contributing

1. Fork it!
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Add some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create a new Pull Request