You are viewing documentation for Kubernetes version: v1.27
Kubernetes v1.27 documentation is no longer actively maintained. The version you are currently viewing is a static snapshot. For up-to-date information, see the latest version.
Contributing to the Upstream Kubernetes Code
This page shows how to contribute to the upstream kubernetes/kubernetes
project.
You can fix bugs found in the Kubernetes API documentation or the content of
the Kubernetes components such as kubeadm
, kube-apiserver
, and kube-controller-manager
.
If you instead want to regenerate the reference documentation for the Kubernetes
API or the kube-*
components from the upstream code, see the following instructions:
- Generating Reference Documentation for the Kubernetes API
- Generating Reference Documentation for the Kubernetes Components and Tools
Before you begin
-
You need to have these tools installed:
-
Your
GOPATH
environment variable must be set, and the location ofetcd
must be in yourPATH
environment variable. -
You need to know how to create a pull request to a GitHub repository. Typically, this involves creating a fork of the repository. For more information, see Creating a Pull Request and GitHub Standard Fork & Pull Request Workflow.
The big picture
The reference documentation for the Kubernetes API and the kube-*
components
such as kube-apiserver
, kube-controller-manager
are automatically generated
from the source code in the upstream Kubernetes.
When you see bugs in the generated documentation, you may want to consider creating a patch to fix it in the upstream project.
Cloning the Kubernetes repository
If you don't already have the kubernetes/kubernetes repository, get it now:
mkdir $GOPATH/src
cd $GOPATH/src
go get github.com/kubernetes/kubernetes
Determine the base directory of your clone of the
kubernetes/kubernetes repository.
For example, if you followed the preceding step to get the repository, your
base directory is $GOPATH/src/github.com/kubernetes/kubernetes.
The remaining steps refer to your base directory as <k8s-base>
.
Determine the base directory of your clone of the
kubernetes-sigs/reference-docs repository.
For example, if you followed the preceding step to get the repository, your
base directory is $GOPATH/src/github.com/kubernetes-sigs/reference-docs.
The remaining steps refer to your base directory as <rdocs-base>
.
Editing the Kubernetes source code
The Kubernetes API reference documentation is automatically generated from an OpenAPI spec, which is generated from the Kubernetes source code. If you want to change the API reference documentation, the first step is to change one or more comments in the Kubernetes source code.
The documentation for the kube-*
components is also generated from the upstream
source code. You must change the code related to the component
you want to fix in order to fix the generated documentation.
Making changes to the upstream source code
Here's an example of editing a comment in the Kubernetes source code.
In your local kubernetes/kubernetes repository, check out the default branch, and make sure it is up to date:
cd <k8s-base>
git checkout master
git pull https://github.com/kubernetes/kubernetes master
Suppose this source file in that default branch has the typo "atmost":
kubernetes/kubernetes/staging/src/k8s.io/api/apps/v1/types.go
In your local environment, open types.go
, and change "atmost" to "at most".
Verify that you have changed the file:
git status
The output shows that you are on the master branch, and that the types.go
source file has been modified:
On branch master
...
modified: staging/src/k8s.io/api/apps/v1/types.go
Committing your edited file
Run git add
and git commit
to commit the changes you have made so far. In the next step,
you will do a second commit. It is important to keep your changes separated into two commits.
Generating the OpenAPI spec and related files
Go to <k8s-base>
and run these scripts:
./hack/update-codegen.sh
./hack/update-openapi-spec.sh
Run git status
to see what was generated.
On branch master
...
modified: api/openapi-spec/swagger.json
modified: api/openapi-spec/v3/apis__apps__v1_openapi.json
modified: pkg/generated/openapi/zz_generated.openapi.go
modified: staging/src/k8s.io/api/apps/v1/generated.proto
modified: staging/src/k8s.io/api/apps/v1/types_swagger_doc_generated.go
View the contents of api/openapi-spec/swagger.json
to make sure the typo is fixed.
For example, you could run git diff -a api/openapi-spec/swagger.json
.
This is important, because swagger.json
is the input to the second stage of
the doc generation process.
Run git add
and git commit
to commit your changes. Now you have two commits:
one that contains the edited types.go
file, and one that contains the generated OpenAPI spec
and related files. Keep these two commits separate. That is, do not squash your commits.
Submit your changes as a pull request to the master branch of the kubernetes/kubernetes repository. Monitor your pull request, and respond to reviewer comments as needed. Continue to monitor your pull request until it is merged.
PR 57758 is an example of a pull request that fixes a typo in the Kubernetes source code.
staging
directory
in the kubernetes/kubernetes
repository. But in your situation,the staging
directory
might not be the place to find the authoritative source. For guidance, check the
README
files in
kubernetes/kubernetes
repository and in related repositories, such as
kubernetes/apiserver.
Cherry picking your commit into a release branch
In the preceding section, you edited a file in the master branch and then ran scripts to generate an OpenAPI spec and related files. Then you submitted your changes in a pull request to the master branch of the kubernetes/kubernetes repository. Now suppose you want to backport your change into a release branch. For example, suppose the master branch is being used to develop Kubernetes version 1.31, and you want to backport your change into the release-1.30 branch.
Recall that your pull request has two commits: one for editing types.go
and one for the files generated by scripts. The next step is to propose a cherry pick of your first
commit into the release-1.30 branch. The idea is to cherry pick the commit
that edited types.go
, but not the commit that has the results of running the scripts. For instructions, see
Propose a Cherry Pick.
When you have a pull request in place for cherry picking your one commit into the release-1.30 branch, the next step is to run these scripts in the release-1.30 branch of your local environment.
./hack/update-codegen.sh
./hack/update-openapi-spec.sh
Now add a commit to your cherry-pick pull request that has the recently generated OpenAPI spec and related files. Monitor your pull request until it gets merged into the release-1.30 branch.
At this point, both the master branch and the release-1.30 branch have your updated types.go
file and a set of generated files that reflect the change you made to types.go
. Note that the
generated OpenAPI spec and other generated files in the release-1.30 branch are not necessarily
the same as the generated files in the master branch. The generated files in the release-1.30 branch
contain API elements only from Kubernetes 1.30. The generated files in the master branch might contain
API elements that are not in 1.30, but are under development for 1.31.
Generating the published reference docs
The preceding section showed how to edit a source file and then generate
several files, including api/openapi-spec/swagger.json
in the
kubernetes/kubernetes
repository.
The swagger.json
file is the OpenAPI definition file to use for generating
the API reference documentation.
You are now ready to follow the Generating Reference Documentation for the Kubernetes API guide to generate the published Kubernetes API reference documentation.