With the revamping of Jenkins 2.0 and Blue Ocean, having a "drop in" cookbook development Jenkins job seemed like a no brainer. This is a generic enough drop in example with only a couple tweaks to make it work in your environment.
This file was verified against the sudo cookbook from the chef-cookbooks GitHub repository. It's supported by the Awesome Chef Community Engineering team, and normally is verified via Travis. The Travis workflow works awesome in a public setting, but this Jenkinsfile is a way to bring that same workflow in house. I chose the Docker workflow to help create a quick disposable environment, so you have a "clean" test bed every time you run your Pipeline.
This is only a framework, you will need to configure a few of the settings for your specific use case, but this should be the lions share of the work you need. This Jenkinsfile is intended to be used as a "multi branch pipeline" job type in Jenkins, and live alongside your code.
Here's an example of the Blue Ocean Pipeline:
And the "default" Pipeline plugin for Jenkins:
I'll walk through each section explain what each does, to make sure we are on the same
page. I consider a section {..}
, and I'll name it via the line number too. Lets go!
pipeline {
- This is the main declaration of the specific pipeline, everything goes in here.agent {
- This sets up the agent that does the work in the workspace. As you can see this command uses the standard chef/chefdk
container with the chefdk already in it. The -u root
allows us to install some dependencies into the container, and because we use kitchen-dokken
later on, we have "Docker in Docker" which is the bind mount of /var/run/docker.sock
.triggers {
- This line is how often the pipeline should run. I put an H
in for the min, which is a "random number". It also uses pollSCM
to see if the git sha has changed, and only pulls down if it has. This helps with network overhead with larger repositories.stages {
- The location for all the "build stages" to be declared.stage('\u27A1 Dependencies for Docker and ChefDK') {
- The first place where real work is done. This stage updates the apt
repositories, installs things required for Docker to be able to be built.stage('\u27A1 Install Docker-CE') {
- This installs standard Docker-CE, with it you can run kitchen-dokken
.stage('\u27A1 Start Docker') {
- In my testing while installing Docker in Docker, the service sometimes didn't start. This stage sends a start command to the service to make sure it is up and running.stage('\u27A1 Verify Docker') {
- Verifying Docker is a safe sanity check in this pipeline. Running the docker run hello-world
shows you can call out to the Public Repos, run the docker command, and have an output.stage('\u27A1 Verify ChefDK') {
- This verifies that the ChefDK commands run as expected. Also it helps with debugging version numbers if something goes off the rails.stage('\u27A1 Verify Kitchen') {
- Because we are focusing on using kitchen-dokken
as our driver, this outputs all the versions of the suites and platforms. It also verifies that test-kitchen
is properly set up and can take commands.stage('\u27A1 Run test-kitchen') {
- The true meat of this file. We run kitchen-dokken
for each of the five platforms in parallel (Line 55). This allows for concurrency and faster verification. Also using test
as the command, it cleans up the previous containers, and run the Inspec validation after each of the converges allowing for better visibility.stage('\u27A1 Upload to Chef Server') {
- And as a final stage, we can upload the cookbook to our Chef server. Now, if you've read this far, you'll probably be wondering how to do this, we haven't done any configuration to this. That's our next section.Now that you've configured your Jenkinsfile with the generic chef/chefdk
you'll
probably want to publish to your Chef server if everything has passed. This next
section is a tad bit more complex, but pretty straight forward.
Go ahead and spin up the chef/chefdk
Docker container:
$ docker run -it chef/chefdk /bin/bash
root@7425261cb693:/#
In another window/terminal run the following command to figure out the docker container id:
$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
7425261cb693 chef/chefdk "/bin/bash" 54 seconds ago Up 53 seconds laughing_lovelace
$
Back on the chef/chefdk
docker container, create a .chef
directory at $HOME
which is /
:
root@7425261cb693:/# mkdir .chef
Then copy in your knife.rb
, key.pem
, validator.pem
from your workstation to that directory:
$ docker cp ~/.chef/knife.rb 7425261cb693:/.chef/
$ docker cp ~/.chef/my-key.pem 7425261cb693:/.chef/
$ docker cp ~/.chef/my-validator.pem 7425261cb693:/.chef/
You should be able to run knife status
from your docker container now, it should be able to talk to your
chef server.
root@7425261cb693:/# chef exec knife status
20 minutes ago, web01
Awesome! OK, so now we need to commit and save this container. Occasionally, you'll need to go through these
steps when a new chef/chefdk
base container is released, back to the versioning
step.
Back on the workstation, we are going to set up a local registry, inject the container into it.
$ docker run -d -p 5000:5000 --restart=always --name registry registry:2 # start up registry container if not already started
$ docker commit -m "keys included for Chef Server" 7425261cb693 # commit the changes to the container
$ docker images -a # list images to verify that it has been committed
$ docker tag 7425261cb693 localhost:5000/chefdkkeysv1 # tag the image you committed
$ docker push localhost:5000/chefdkkeysv1 # push to local registry
You should be able to take the container and ship it to where you run Docker from your Jenkins box. There are a ton of variants here so I'll go ahead an let you figure that one out. NOTE: Being your keys are in this container now, realize that it should NOT GO TO THE PUBLIC REGISTRY.
Now that you have it in a location that your Jenkins instance can reach it, all you need to do is change the following setting and you should be able to pull the container down now:
docker {
reuseNode false
args '-u root -v /var/run/docker.sock:/var/run/docker.sock'
image 'chefdkkeysv1'
}
With this, your final stage of stage('\u27A1 Upload to Chef Server') {
should start publishing to your
chef server if everything passes. Note this only happens on the production branch (see the when
statement).
If you would like to add a gate to Approval
or Disapproval
it's a simple line to add!
stage('\u27A1 Upload to Chef Server') {
steps {
input 'Do you want to proceed to the Deployment?'
sh 'chef exec knife cookbook upload sudo -o ../'
}
}
Here's an image of what it looks like on Blue Ocean!
On line 51:
when {
anyOf { branch 'master'; branch 'staging' }
}
specifies that the 'Run test-kitchen' stage should run only on the master, staging or production branches. No need to do this work on other experimental branches. Likewise, on the upload stage, it will be skipped unless the production branch is updated. You can specify a single branch or multiple branches to run on.
Hopefully this will give you the framework you need to get your own Jenkinsfile configured for your own environment.
There are a few configurations you'll need to make in the process, namely the keys
and knife.rb
but assuming you
just want to verify you can drop the final example stage
. If you have questions never hesitate to reach out via
the Chef-Community Slack, @jj
.
Author:: JJ Asghar (jj@chef.io)
Copyright:: Copyright (c) 2017 JJ Asghar
License:: Apache License, Version 2.0
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.