This project provides a docker image with the google SDK to ease making management operations on Google's kubernetes implementation.

The image contains both gcloud (for authentication and infra needs) as well as kubectl of a version that matches the version of kubernetes that google has deployed at the time the image is built.

Left up to the user is mounting credential information, such as a service account key, into the container, to provide the container the authority to manage the kubernetes cluster(s) and other cloud resources in Google Cloud.

Does not require any initialization and is suitable for when one wants ephemeral use, such as in a CI product.

There is just one requirement, that you have populated a gcloud credentials file with the service account information you want to use, and mounted it into the container at: /root/.config/gcloud/credentials

The benefit of stateful configuration is that you can separate configuration from use, into distinct lifecycle steps, suitable for running on one's private laptop to save on needless typing. The downside is that your risk clobbering the configuration if you frequently need to swap between configurations (such as in a CI system).

Stateful use of the container requires an initial configuration of kubectl and gcloud, so that the SDK knows where your cluster is and what credential to use.

The following assumptions have been made:

• A service account for use with Google Container Engine already exists.
• A private key for the service account has been saved in the project folder as secrets/key.json.

To initialize the use of kubectl for a kubernetes cluster, run the following:

docker run --rm -ti -v $PWD/secrets:/root/secrets tomologic/kubeadmin \
  initialize <project-id> <region> <cluster> secrets/key.json

• project-id is the ID of the project that contains the kubernetes cluster
• availability-zone is the zone where the cluster resides
• cluster is the name of the container cluster
• secrets/key.json is the path to the file in the container containing the service account private key.

Example use of SDK with stateful context, after initialization:

alias c="docker run --rm -i -v $PWD/secrets/.config:/root/.config -v $PWD/secrets/.kube:/root/.kube tomologic/kubeadmin"
c kubectl get pods
c gcloud container clusters list

Example of stateless use:

docker run --rm -i \
-v $HOME/.config/gcloud/credentials:/root/.config/gcloud/credentials:ro \
-e CLOUDSDK_COMPUTE_ZONE=europe-west1-a \
-e CLOUDSDK_COMPUTE_REGION=europe-west1 \
tomologic/kubeadmin gcloud \
--project my-project \
--account [email protected] \
compute instances list

Note that the account used needs to have been imported into the host's credentials file first. Refer to Google's docs on the sibject for how to use their tools.

Example of stateless, using a service account key file to list available buckets

docker run --rm -v ${SERVICE_ACCOUNT_KEY}:/key.json:ro tomologic/kubeadmin sh -c \
"gcloud auth activate-service-account --key-file=/key.json && gcloud config set project ${PROJECT} && gsutil ls"