The Kubernetes platform.
The Package manager.
The Open Service Broker.
This document is for developers who are interested in working directly on the Drycc codebase. In this guide, we walk you through the process of setting up a development environment that is suitable for hacking on most Drycc components.
We try to make it simple to hack on Drycc components. However, there are necessarily several moving pieces and some setup required. We welcome any suggestions for automating or simplifying this process.
The Drycc team is actively engaged in containerizing Go and Python based development environments tailored specifically for Drycc development in order to minimize the setup required. This work is ongoing. Refer to the drycc/router project for a working example of a fully containerized development environment.
If you're just getting into the Drycc codebase, look for GitHub issues with the label easy-fix. These are more straightforward or low-risk issues and are a great way to become more familiar with Drycc.
In order to successfully compile and test Drycc binaries and build Docker images of Drycc components, the following are required:
For drycc/controller, in particular, you will also need:
sudo pip install virtualenv)
In most cases, you should simply install according to the instructions. There are a few special cases, though. We cover these below.
If your local workstation does not support the
linux/amd64 target environment, you will have to install Go from source with cross-compile support for that environment. This is because some of the components are built on your local machine and then injected into a Docker container.
Homebrew users can just install with cross compiling support:
$ brew install go --with-cc-common
It is also straightforward to build Go from source:
$ sudo su $ curl -sSL https://golang.org/dl/go1.5.src.tar.gz | tar -v -C /usr/local -xz $ cd /usr/local/go/src $ # compile Go for our default platform first, then add cross-compile support $ ./make.bash --no-clean $ GOOS=linux GOARCH=amd64 ./make.bash --no-clean
Once you can compile to
linux/amd64, you should be able to compile Drycc components as normal.
Drycc needs Docker for building images. Docker utilizes a client/server architecture, and while the Docker client is available for Mac OS, the Docker server is dependent upon the Linux kernel. Therefore, in order to use Docker on Mac OS, Docker Machine is used to facilitate running the Docker server within a VirtualBox VM.
Install Docker Machine according to the normal installation instructions, then use it to create a new VM:
$ docker-machine create drycc-docker \ --driver=virtualbox \ --virtualbox-disk-size=100000 \ --engine-insecure-registry 10.0.0.0/8 \ --engine-insecure-registry 172.16.0.0/12 \ --engine-insecure-registry 192.168.0.0/16 \ --engine-insecure-registry 100.64.0.0/10
This will create a new virtual machine named
drycc-docker that will take up as much as 100,000 MB of disk space. The images you build may be large, so allocating a big disk is a good idea.
drycc-docker machine exists, source its values into your environment so your docker client knows how to use the new machine. You may even choose to add this to your bash profile or similar.
$ eval "$(docker-machine env drycc-docker)"
After following these steps, some Docker Machine users report a slight delay (30 - 60 seconds) before the Docker server is ready.
In subsequent steps, you may run a Docker registry within the
drycc-docker VM. Such a registry will not have a valid SSL certificate and will use HTTP instead of HTTPS. Such registries are implicitly untrusted by the Docker server (which is also running on the
drycc-docker VM). In order for the Docker server to trust the insecure registry,
drycc-docker is explicitly created to trust all registries in the IP ranges that that are reserved for use by private networks. The VM (and therefore the registry) will exist within such a range. This will effectively permit Docker pulls and pushes to such a registry.
Once the prerequisites have been met, we can begin to work with Drycc components.
Begin at Github by forking whichever Drycc project you would like to contribute to, then clone that fork locally. Since Drycc is predominantly written in Go, the best place to put it is under
$ mkdir -p $GOPATH/src/github.com/drycc $ cd $GOPATH/src/github.com/drycc $ git clone firstname.lastname@example.org:<username>/<component>.git $ cd <component>
By checking out the forked copy into the namespace
github.com/drycc/<component>, we are tricking the Go toolchain into seeing our fork as the "official" source tree.
If you are going to be issuing pull requests to the upstream repository from which you forked, we suggest configuring Git such that you can easily rebase your code to the upstream repository's main branch. There are various strategies for doing this, but the most common is to add an
$ git remote add upstream https://github.com/drycc/<component>.git
For the sake of simplicity, you may want to point an environment variable to your Drycc code - the directory containing one or more Drycc components:
$ export DRYCC=$GOPATH/src/github.com/drycc
Throughout the rest of this document,
$DRYCC refers to that location.
A number of Drycc contributors prefer to pull directly from
drycc/<component>, but push to
<username>/<component>. If that workflow suits you better, you can set it up this way:
$ git clone email@example.com:drycc/<component>.git $ cd drycc $ git config remote.origin.pushurl firstname.lastname@example.org:<username>/<component>.git
In this setup, fetching and pulling code will work directly with the upstream repository, while pushing code will send changes to your fork. This makes it easy to stay up to date, but also make changes and then issue pull requests.
With your development environment set up and the code you wish to work on forked and cloned, you can begin making your changes.
Drycc components each include a comprehensive suite of automated tests, mostly written in Go. See testing for instructions on running the tests.
Although writing and executing tests are critical to ensuring code quality, most contributors will also want to deploy their changes to a live environment, whether to make use of those changes or to test them further. The remainder of this section documents the procedure for running officially released Drycc components in a development cluster and replacing any one of those with your customizations.
To run a Kubernetes cluster locally or elsewhere to support your development activities, refer to Drycc installation instructions here.
To facilitate deploying Docker images containing your changes to your Kubernetes cluster, you will need to make use of a Docker registry. This is a location to where you can push your custom-built images and from where your Kubernetes cluster can retrieve those same images.
If your development cluster runs locally (in Minikube, for instance), the most efficient and economical means of achieving this is to run a Docker registry locally as a Docker container.
To facilitate this, most Drycc components provide a make target to create such a registry:
$ make dev-registry
In a Linux environment, to begin using the registry:
export DRYCC_REGISTRY=<IP of the host machine>:5000
In non-Linux environments:
export DRYCC_REGISTRY=<IP of the drycc-docker Docker Machine VM>:5000
If your development cluster runs on a cloud provider such as Google Container Engine, a local registry such as the one above will not be accessible to your Kubernetes nodes. In such cases, a public registry such as DockerHub or quay.io will suffice.
To use DockerHub for this purpose, for instance:
$ export DRYCC_REGISTRY="docker.io" $ export IMAGE_PREFIX=<your DockerHub username>
To use quay.io:
$ export DRYCC_REGISTRY=quay.io $ export IMAGE_PREFIX=<your quay.io username>
Note the importance of the trailing slash.
With a functioning Kubernetes cluster and the officially released Drycc components installed onto it, deployment and further testing of any Drycc component you have made changes to is facilitated by replacing the officially released component with a custom built image that contains your changes. Most Drycc components include Makefiles with targets specifically intended to facilitate this workflow with minimal friction.
In the general case, this workflow looks like this:
make buildto build a new Docker image
make dev-releaseto generate Kubernetes manifest(s)
make deployto restart the component using the updated manifest
This can be shortened to a one-liner using just the
$ make deploy
Once your customized Drycc component has been deployed, here are some helpful commands that will allow you to inspect your cluster and to troubleshoot, if necessary:
$ kubectl --namespace=drycc get pods
This is often useful for troubleshooting pods that are in pending or crashed states:
$ kubectl --namespace=drycc describe -f <pod name>
$ kubectl --namespace=drycc logs -f <pod name>
Specific to drycc/controller
$ kubectl --namespace=drycc exec -it <pod name> -- python manage.py shell
Have commands other Drycc contributors might find useful? Send us a PR!
Satisfied with your changes? Share them!
Please read Submitting a Pull Request. It contains a checklist of things you should do when proposing a change to any Drycc component.