- 1 Introduction
- 2 Puppet Modules
- 3 Getting Help
- 4 Developer documentation
- 4.1 Contributing to the modules
- 4.2 How code gets merged
- 4.3 Continuous Integration
- 4.4 Patches
- 4.5 Rspec puppet tests
- 4.6 Patch abandonment policy
- 4.7 stable branches
- 4.8 Gerrit Dashboard
- 5 Weekly meetings
The Puppet modules for OpenStack are written as a collaborative effort between OpenStack operators using Puppet.
Puppet OpenStack modules are part of OpenStack and most have moved under the OpenStack big tent. Some modules (like ceph) remain in the stackforge namespace.
The following Puppet modules exist:
- stackforge/puppet-openstack (deprecated in Juno release)
Library repository, used for common resources accross all modules:
Optional tools that can be used in composition layer and might be helpful to deploy OpenStack:
The master branch of each modules corresponds to the latest packages available in upstream repositories. So if you submit a patch in master, it should pass beaker jobs that rely on upstream packaging (RDO in CentOS7 and UCA in Ubuntu Trusty).
For example, the current master branch of stackforge/puppet-keystone is targeting OpenStack Kilo, once released, a new stable/kilo branch will be created, from that point onward the master branch should target the OpenStack L release.
Master version of the modules will be released as a new major version to Puppet Forge (forge.puppetlabs.com) when its related version of OpenStack is released.
Each version of OpenStack have a corresponding release on the forge.
|Module Version||OpenStack Version||OpenStack Version Codename||Supported|
|7.z.y||2015.2.x||Liberty||Yes (current master)|
The following OS/version are supported by the Puppet modules:
- Fedora 18 / 19 / 20 / 21 / 22
- RHEL 6.4 / 6.5 / 7
- Ubuntu 12.04 (Precise) / 14.04 (Trusty)
- Debian 7.0 (Wheezy) / Debian 8.0 (Jessie)
The modules have been primarily tested on Puppet 2.7.x and Ruby 1.8.7, although they are also being used with Puppet 3.1.x, 3.0.x, with Ruby 1.9.3.
Puppet 2.6.x is currently NOT supported, it has been EOL'ed by PuppetLabs.
In general, the mailing list is preferred, because it makes the information more readily available so that others who have the same question can search for and find those answers.
- Dev discussions: firstname.lastname@example.org with
- Usage discussions: email@example.com with
You can talk to us directly in IRC on the
#puppet-openstack channel. (freenode.net)
IRC logs can be found here.
Contributing to the modules
We follow the same process as all of the other OpenStack projects.
To contribute, the following docs contain enough information to get started:
- go to : settings > watched projects and add the puppet projects (all of the form stackforge/puppet-*)
- hint: Download all puppet-openstack git repo using mr
Build new modules
If you want to create a new Puppet module for an OpenStack component, please follow this process https://wiki.openstack.org/wiki/Puppet/New_module
How code gets merged
Code is merged based on the voting process of the modules in Gerrit. All submitted patches automatically trigger a job that runs its rspec-puppet tests. This job is considered to be a gate in that no code is allowed to be merged that does not pass these tests. The results of this job are listed for every patch as a +1 Verified vote from Jenkins.
Any users can +/- 1 a commit and add comments on commit, but only members of the puppet-manager-core group have the ability to +2 and approve code to be merged.
Our modules are running CI that test different things against a patch. You can read more about how our CI works here: Puppet/CI.
Patches waiting to be merged can be viewed in Gerrit e.g. for openstack/puppet-keystone
Q. How do I go about submitting a patch for a released branch, what the correct process? Unless is not relevant all patches should be approved for the master branch before you submit them for a stable branch. This ensures we maintain stability in the stable branches and functionality
Downloading a local patch
Clone the relevant module from openstack, ex:
in the patch, find the git checkout or cherry-pick command, and copy it:
git fetch https://review.openstack.org/openstack/puppet-keystone refs/changes/52/29452/9 && git checkout FETCH_HEAD
if you wanted to update an existing patch:
make a topic branch:
git checkout -b my_topic
make your changes:
amend the current commit:
git commit --amend .
Rspec puppet tests
Rspec puppet tests are a requirement for getting code merged into the OpenStack puppet modules.
The best reference for getting started with rspec-puppet can be found here
For a new puppet openstack project, please see this page
Running local tests
The following command can invoked from any if the modules' directories to run their rspec puppet tests.
It assumes that both bundler as well as rubygems (and ruby) are already installed on the system.
mkdir vendor export GEM_HOME=vendor bundle install # bundle exec rake -T bundle exec rake spec
This relies on the file .fixtures.yaml to install all of the external module required for testing. The urls in this file use the git:// protocol, so this may need to be updated if you are behind a proxy.
Patch abandonment policy
If a change is submitted and given a -1, and subsequently the author becomes unresponsive for a few weeks, reviewers should leave reminder comments on the review or attempt to contact the original author via IRC or email. If the change is easy to fix, anyone should feel welcome to check out the change and resubmit it using the same change ID to preserve original authorship. Core reviewers will not abandon such a change.
If a change is submitted and given a -2, or it otherwise becomes clear that the change can not make it in (for example, if an alternate change was chosen to solve the problem), and the author has been unresponsive for at least 3 months, a core reviewer should abandon the change.
This policy is subject to change as we review our bandwidth for taking up forgotten patches and monitor our backlog growth.
Master of the modules should be targeting the trunk of openstack as soon as a release is announced.
When this happens, a stable branch is created for any previous releases.
It is the responsibility for the patch submitted to know if their patch should be backported to previous stable branches.
The general process for this can be found here: https://wiki.openstack.org/wiki/StableBranch#Workflow
create a stable branch
For branching you just go to the project page on Gerrit and make a new branch under the branches option.
When you first create it it doesn't really appear in git until the first commit gets pushed to it or master.