cookbook 'pipeline', '= 2.1.1'
pipeline (4) Versions 2.1.1 Follow11
Installs/Configures a Jenkins based chef delivery pipeline
cookbook 'pipeline', '= 2.1.1', :supermarket
knife supermarket install pipeline
knife supermarket download pipeline
Continuous delivery pipeline for Chef.
As you commit Chef artifacts to git (i.e. cookbooks, data bags, environments and roles) the pipeline will test and load them into Chef automatically.
The pipeline is Jenkins based and relies on a git repository of your Chef items of the following structure:
- Contains all the cookbooks in your project.
- Used to generate cookbook jobs for cookbooks that need testing
- Used to upload community cookbooks that don't need testing
environments/: Directory of environment files in a format accepted by
roles/: Directory of role files in a format accepted by
data_bags_/: Directory of data bag files in a format accepted by
Both of the following job types listen on their respective repositories and upload to Chef when changes are committed to git.
Once the pipeline is up and fully flushed out you will have N+1 cookbooks in your Jenkins dashboard, where N is the number of cookbooks you have that require testing. And the +1 is the chef_repo job which is responsible for uploading all the roles, environments and data bags as well as cookbooks that bypass testing.
Chef Repo Job
- Uploads roles, environments, data bags and community cookbooks from the git repo to the Chef server
- Lint and test the respective cookbook
- Upload respective cookbook from git to the Chef server
Note that there is a dependency between the chef-client and a Jenkins job. The cookbook jobs are generated by the chef-client from the chef_repo that is kept up to date by Jenkins. Because of this the chef-client will need to run twice on initial set up of the pipeline server to flush out all of the cookbook jobs.
From the command line, run the following twice:
kitchen converge centos-7
Now check out your local CD pipeline demo on http://localhost:9090
Update the following to a chef_repo of your own:
default['pipeline']['chef-repo']['url'] = "https://github.com/stephenlauck/pipeline_chef.git"
- There is a .chef folder in the jenkins user's home. You will need to put an appropriate .pem file in there for a pipeline user with admin rights to Chef.
- This uses the Chef embedded version of Ruby for knife and Berkshelf etc.
- Since this uses Berkshelf for cookbook uploads, all cookbook versions are frozen by default. Failure to update the metadata version number should result in a failed job.
- Since this uses Berkshelf to upload your cookbooks to Chef, there is an expectation that every cookbook to be tested contain the usual Berkshelf workflow tools such as kitchen, vagrant and berks.
- Ubuntu 12.04
- CentOS 6.5
|apt >= 0.0.0
|yum >= 0.0.0
|git >= 0.0.0
|java >= 0.0.0
|jenkins >= 0.0.0
|chef-zero >= 0.0.0
|emacs >= 0.0.0
|sudo >= 0.0.0
|chef-dk >= 0.0.0
|resolver >= 0.0.0
There are no cookbooks that are contingent upon this one.
pipeline Cookbook CHANGELOG
This file is used to list changes made in each version of the pipeline cookbook.
- Add lint recipe to setup lint testing on jobs
- pin dependent cookbook versions
- job commands as partials
broke out recipes from default into berkshelf, jenkins, jobs
added build command attr in order to step thru cookbook job -- may belong in wrapper
added LWRP for creating chef_server configuration ala berks config
added chef-zero support
2.1.1 failed this metric
FC034: Unused template variables: /tmp/cook/1c639b31f156d4d44598eced/pipeline/templates/default/job-config.xml.erb:1
2.1.1 failed this metric