Preparing the Setup
Conceptually, all Gardener components are designated to run inside as a Pod inside a Kubernetes cluster. The API server extends the Kubernetes API via the user-aggregated API server concepts. However, if you want to develop it, you may want to work locally with the Gardener without building a Docker image and deploying it to a cluster each and every time. That means that the Gardener runs outside a Kubernetes cluster which requires providing a Kubeconfig in your local filesystem and point the Gardener to it when starting it (see below).
Further details could be found in
Installing Golang environment
Install latest version of Golang. For MacOS you could use Homebrew:
brew install golang
For other OS, please check Go installation documentation.
Installing kubectl and helm
As already mentioned in the introduction, the communication with the Gardener happens via the Kubernetes (Garden) cluster it is targeting. To interact with that cluster, you need to install
kubectl. Please make sure that the version of
kubectl is at least
On MacOS run
brew install kubernetes-cli
Please check the kubectl installation documentation for other OS.
You may also need to develop Helm charts or interact with Tiller using the Helm CLI:
On MacOS run
brew install kubernetes-helm
On other OS please check the Helm installation documentation.
git as VCS which you need to install.
On MacOS run
brew install git
On other OS, please check the Git installation documentation.
OpenVPN to establish network connectivity from the control plane running in the Seed cluster to the Shoot’s worker nodes running in private networks.
To harden the security we need to generate another secret to encrypt the network traffic (details).
Please install the
openvpn binary. On MacOS run
brew install openvpn export PATH=$(brew --prefix openvpn)/sbin:$PATH
On other OS, please check the OpenVPN downloads page.
You’ll need to have minikube installed and running.
Note: Gardener is working only with self-contained kubeconfig files because of security issue. You can configure your minikube to create self-contained kubeconfig files via:
minikube config set embed-certs true
Alternatively, you can also install Docker for Desktop and kind.
In case you want to use the “Docker for Mac Kubernetes” or if you want to build Docker images for the Gardener you have to install Docker itself. On MacOS, please use Docker for MacOS which can be downloaded here.
On other OS, please check the Docker installation documentation.
iproute2 provides a collection of utilities for network administration and configuration.
On MacOS run
brew install iproute2mac
Installing yaml2json and jq
go get -u github.com/bronze1man/yaml2json brew install jq
[MacOS only] Install GNU core utilities
When running on MacOS you have to install the GNU core utilities:
brew install coreutils gnu-sed
This will create symbolic links for the GNU utilities with
g prefix in
gbase64. To allow using them without the
g prefix please put
/usr/local/opt/coreutils/libexec/gnubin at the beginning of your
PATH environment variable, e.g.,
Apart from Linux distributions and MacOS, the local gardener setup can also run on the Windows Subsystem for Linux 2.
While WSL1, plain docker for windows and various Linux distributions and local Kubernetes environments may be supported, this setup was verified with:
The Gardener repository and all the above-mentioned tools (git, golang, kubectl, …) should be installed in your WSL2 distro, according to the distribution-specific Linux installation instructions.
[Optional] Installing gcloud SDK
In case you have to create a new release or a new hotfix of the Gardener you have to push the resulting Docker image into a Docker registry. Currently, we are using the Google Container Registry (this could change in the future). Please follow the official installation instructions from Google.
Local Gardener setup
This setup is only meant to be used for developing purposes, which means that only the control plane of the Gardener cluster is running on your machine.
Get the sources
Clone the repository from GitHub.
git clone firstname.lastname@example.org:gardener/gardener.git cd gardener
Start the Gardener
:warning: Before you start developing, please ensure to comply with the following requirements:
- You have understood the principles of Kubernetes, and its components, what their purpose is and how they interact with each other.
- You have understood the architecture of Gardener, and what the various clusters are used for.
Start a local kubernetes cluster
For the development of Gardener you need some kind of Kubernetes cluster, which can be used as a “garden” cluster.
I.e. you need a Kubernetes API server on which you can register a
APIService Gardener’s own Extension API Server.
For this you can use a standard tool from the community to setup a local cluster like minikube, kind or the Kubernetes Cluster feature in Docker for Desktop.
However, if you develop and run Gardener’s components locally, you don’t actually a fully fledged Kubernetes Cluster,
i.e. you don’t actually need to run Pods on it. If you want to use a more lightweight approach for development purposes,
you can use the “nodeless Garden cluster setup” residing in
hack/local-garden. This is the easiest way to get your
Gardener development setup up and running.
Using the nodeless cluster setup
Setting up a local nodeless Garden cluster is quite simple. The only prerequisite is a running docker daemon. Just use the provided Makefile rules to start your local Garden:
make local-garden-up [...] Starting gardener-dev kube-etcd cluster..! Starting gardener-dev kube-apiserver..! Starting gardener-dev kube-controller-manager..! Starting gardener-dev gardener-etcd cluster..! namespace/garden created clusterrole.rbac.authorization.k8s.io/gardener.cloud:admin created clusterrolebinding.rbac.authorization.k8s.io/front-proxy-client created [...]
This will start all minimally required components of a Kubernetes cluster (
etcd Instance for the
gardener-apiserver as Docker containers.
To tear down the local Garden cluster and remove the Docker containers, simply run:
Alternatively, spin up a cluster with minikube with this command:
minikube start --embed-certs # `--embed-certs` can be omitted if minikube has already been set to create self-contained kubeconfig files. 😄 minikube v1.8.2 on Darwin 10.15.3 🔥 Creating virtualbox VM (CPUs=2, Memory=2048MB, Disk=20000MB) ... [...] 🏄 Done! Thank you for using minikube!
Prepare the Gardener
Now, that you have started your local cluster, we can go ahead and register the Gardener API Server.
Just point your
KUBECONFIG environment variable to the local cluster you created in the previous step and run:
make dev-setup Found Minikube ... namespace/garden created namespace/garden-dev created deployment.apps/etcd created service/etcd created service/gardener-apiserver created service/gardener-controller-manager created endpoints/gardener-apiserver created endpoints/gardener-controller-manager created apiservice.apiregistration.k8s.io/v1alpha1.core.gardener.cloud created apiservice.apiregistration.k8s.io/v1beta1.core.gardener.cloud created validatingwebhookconfiguration.admissionregistration.k8s.io/gardener-controller-manager created
Optionally, you can switch off the
Logging feature gate of Gardenlet to save resources:
sed -i -e 's/Logging: true/Logging: false/g' dev/20-componentconfig-gardenlet.yaml
The Gardener exposes the API servers of Shoot clusters via Kubernetes services of type
In order to establish stable endpoints (robust against changes of the load balancer address), it creates DNS records pointing to these load balancer addresses. They are used internally and by all cluster components to communicate.
You need to have control over a domain (or subdomain) for which these records will be created.
Please provide an internal domain secret (see this for an example) which contains credentials with the proper privileges. Further information can be found here.
kubectl apply -f example/10-secret-internal-domain-unmanaged.yaml secret/internal-domain-unmanaged created
Run the Gardener
Next, run the Gardener API Server, the Gardener Controller Manager (optionally), the Gardener Scheduler (optionally), and the Gardenlet in different terminal windows/panes using rules in the
make start-apiserver Found Minikube ... I0306 15:23:51.044421 74536 plugins.go:84] Registered admission plugin "ResourceReferenceManager" I0306 15:23:51.044523 74536 plugins.go:84] Registered admission plugin "DeletionConfirmation" [...] I0306 15:23:51.626836 74536 secure_serving.go:116] Serving securely on [::]:8443 [...]
(Optional) Now you are ready to launch the Gardener Controller Manager.
make start-controller-manager time="2019-03-06T15:24:17+02:00" level=info msg="Starting Gardener controller manager..." time="2019-03-06T15:24:17+02:00" level=info msg="Feature Gates: " time="2019-03-06T15:24:17+02:00" level=info msg="Starting HTTP server on 0.0.0.0:2718" time="2019-03-06T15:24:17+02:00" level=info msg="Acquired leadership, starting controllers." time="2019-03-06T15:24:18+02:00" level=info msg="Starting HTTPS server on 0.0.0.0:2719" time="2019-03-06T15:24:18+02:00" level=info msg="Found internal domain secret internal-domain-unmanaged for domain nip.io." time="2019-03-06T15:24:18+02:00" level=info msg="Successfully bootstrapped the Garden cluster." time="2019-03-06T15:24:18+02:00" level=info msg="Gardener controller manager (version 1.0.0-dev) initialized." time="2019-03-06T15:24:18+02:00" level=info msg="ControllerRegistration controller initialized." time="2019-03-06T15:24:18+02:00" level=info msg="SecretBinding controller initialized." time="2019-03-06T15:24:18+02:00" level=info msg="Project controller initialized." time="2019-03-06T15:24:18+02:00" level=info msg="Quota controller initialized." time="2019-03-06T15:24:18+02:00" level=info msg="CloudProfile controller initialized." [...]
(Optional) Now you are ready to launch the Gardener Scheduler.
make start-scheduler time="2019-05-02T16:31:50+02:00" level=info msg="Starting Gardener scheduler ..." time="2019-05-02T16:31:50+02:00" level=info msg="Starting HTTP server on 0.0.0.0:10251" time="2019-05-02T16:31:50+02:00" level=info msg="Acquired leadership, starting scheduler." time="2019-05-02T16:31:50+02:00" level=info msg="Gardener scheduler initialized (with Strategy: SameRegion)" time="2019-05-02T16:31:50+02:00" level=info msg="Scheduler controller initialized." [...]
(Optional) Now you are ready to launch the Gardenlet.
make start-gardenlet time="2019-11-06T15:24:17+02:00" level=info msg="Starting Gardenlet..." time="2019-11-06T15:24:17+02:00" level=info msg="Feature Gates: HVPA=true, Logging=true" time="2019-11-06T15:24:17+02:00" level=info msg="Acquired leadership, starting controllers." time="2019-11-06T15:24:18+02:00" level=info msg="Found internal domain secret internal-domain-unmanaged for domain nip.io." time="2019-11-06T15:24:18+02:00" level=info msg="Gardenlet (version 1.0.0-dev) initialized." time="2019-11-06T15:24:18+02:00" level=info msg="ControllerInstallation controller initialized." time="2019-11-06T15:24:18+02:00" level=info msg="Shoot controller initialized." time="2019-11-06T15:24:18+02:00" level=info msg="Seed controller initialized." [...]
:warning: The Gardenlet will handle all your seeds for this development scenario, although, for productive usage it is recommended to run it once per seed, see this document for more information. See the Appendix on how to configure the Seed clusters for the local development scenario.
Please checkout the Gardener Extensions Manager to install extension controllers - make sure that you install all of them required for your local development. Also, please refer to this document for further information about how extensions are registered in case you want to use other versions than the latest releases.
The Gardener should now be ready to operate on Shoot resources. You can use
kubectl get shoots No resources found.
to operate against your local running Gardener API Server.
Note: It may take several seconds until the
minikubecluster recognizes that the Gardener API server has been started and is available.
No resources foundis the expected result of our initial development setup.
Limitations of local development setup
You can run Gardener (API server, controller manager, scheduler, gardenlet) against any local Kubernetes cluster, however, your seed and shoot clusters must be deployed to a “real” provider. Currently, it is not possible to run Gardener entirely isolated from any cloud provider. We are planning to support a setup that can run completely locally (see this for details), however, it does not yet exist. This means that - after you have setup Gardener - you need to register an external seed cluster (e.g., one created in AWS). Only after that step you can start creating shoot clusters with your locally running Gardener.
Some time ago, we had a local setup based on VirtualBox/Vagrant. However, as we have progressed with the Extensibility epic we noticed that this implementation/setup does no longer fit into how we envision external providers to be. Moreover, it hid too many things and came with a bunch of limitations, making the development scenario too “artificial”:
- No integration with machine-controller-manager.
- The Shoot API Server is exposed via a NodePort. In a cloud setup a LoadBalancer would be used.
- It was not possible to create Shoot clusters consisting of more than one worker node. Cluster auto-scaling therefore is not supported.
- It was not possible to create two or more Shoot clusters in parallel.
- The communication between the Seed and the Shoot Clusters uses VPN tunnel. In this setup tunnels are not needed since all components run on localhost.
To make sure that a specific Seed cluster will be chosen, specify the
.spec.seedName field (see here for an example Shoot manifest).
Please take a look at the example manifests folder to see which resource objects you need to install into your Garden cluster.
Configure Seed clusters for local development
When using the Gardenlet in a local development scenario with
make start-gardenlet then the Gardenlet component configuration is setup with a seed selector that targets all available Seed clusters.
Seed resource needs to be configured to allow being reconciled by a Gardenlet which such a configuration.
When deploying the Gardenlet to reconcile only one Seed cluster (using component configuration
the Gardenlet either needs to be supplied with a kubeconfig for the particular Seed cluster, or acquires one via bootstrapping.
Having said that, if the Gardenlet is configured to manage multiple Seed clusters based on a label selector, it needs to fetch the kubeconfig of each Seed cluster at runtime from somewhere.
That is why the
Seed resource needs to be configured with an additional secret reference that contains the kubeconfig of the Seed cluster.
Create a secret containing the base64 encoded kubeconfig of the Seed cluster (the scope of the permissions should be identical to the kubeconfig that the Gardenlet creates during bootstrapping - for now,
cluster-admin privileges are recommended).
Create the secret in the Garden cluster:
apiVersion: v1 kind: Secret metadata: name: sweet-seed namespace: garden type: Opaque data: kubeconfig: <base64-seed-kubeconfig>.
Seed resource to reference the secret in
spec.secretRef like so:
apiVersion: core.gardener.cloud/v1beta1 kind: Seed metadata: name: my-sweet-seed spec: ... secretRef: name: sweet-seed namespace: garden