To run Kubernetes e2e tests, it is necessary to pull images which stores on Google Container Registry. However, it is impossible for us to pull gcr images in domestic China. This project aims to load all required images to a domestic registry on a foreign server, and load these images on a domestic server in China.

This tool requires a Python environment. Please check that /usr/bin/python does exist on your machine. If the environment variable GOPATH does not exist, please add this variable to your environment with

export GOPATH=YOUR_GOPATH

A Kubernetes git repository is required to get image names. Please place the Kubernetes repository in the file $GOPATH/src/k8s.io/kubernetes. To set up a Kubernetes repository, run

mkdir -p $GOPATH/src/k8s.io
cd $GOPATH/src/k8s.io
git clone https://github.com/kubernetes/kubernetes
cd kubernetes

It is recommended to set the desired Kubernetes version by using

git checkout v1.6.5

To check the current version of the Kubernetes under the git repository, run

git describe

Under this project root, using upload.py to upload images to a registry on a foreign server. To see more information, run

./upload.py -h

To specify the registry, using -r flag. To set username and password, useing -u and -p flags separately. For instance, to login to registry named daocloud.io with the username huangjiuyuan and password password, run

./upload.py -r daocloud.io -u huangjiuyuan -p password

The image will be tagged as daocloud.io/huangjiuyuan/IMAGE_NAME and pushed to registry named daocloud.io.

Under this project root, using load.py to load images to domestic server. To see more information, run

./load.py -h

To specify the registry, using -r flag. To set username and password, useing -u and -p flags separately. For instance, to login to registry named daocloud.io with the username huangjiuyuan and password password, run

./load.py -r daocloud.io -u huangjiuyuan -p password

The image named daocloud.io/huangjiuyuan/IMAGE_NAME willed be pulled and tagged back to its original name.

This project uses kube-conformance as a tool to build the testing image. To build an intermediate testing image, run

docker build -t conformance-builder -f Dockerfile.conformance .

To build the conformance testing image, run

docker run -e KVER=v1.6.5 -v /var/run/docker.sock:/var/run/docker.sock conformance-builder

The default version of Kubernetes testing image is v1.6.5. Set the KVER environment variable by -e flag in the above command to specify the target version. After running the above command, a testing image named gcr.io/heptio-images/kube-conformance will be successfully built.