Deployment Manager lets you define and deploy simple declarative configuration for your Kubernetes resources (e.g., pods, replication controllers, services, etc.).
You can also use Python or Jinja to create powerful parameterizable abstract types called Templates. You can create general abstract building blocks to reuse, like a Replicated Service, or create more concrete types like a Redis cluster.
You can find more examples of Templates and configurations in our examples.
Deployment Manager uses the same concepts and languages as Google Cloud Deployment Manager, but creates resources within your Kubernetes cluster, not on the Google Cloud Platform.
Please join us on the Google Group and/or in the Slack chat room for the Kubernetes configuration SIG.
There are two ways to get started...
-
The quick way simply installs Deployment Manager in your cluster using kubectl. This is the fastest way to get started and takes only a few seconds.
-
The interesting way bootstraps Deployment Manager, by building and running a local instance on your machine, and then using it to install another instance in your cluster. You might want to go this way if you're interested in contributing to Deployment Manager.
Both assume that you have a Kubernetes cluster up and running, and that you can
run kubectl commands against it. They both also assume that that you're working
with a clone of the repository installed in the src folder of your GOPATH, per
convention.
Instructions for the quick install follow here. Instructions for bootstrapping Deployment Manager can be found in examples/bootstrap/README.md.
For the quick install, you're going to use kubectl to create the replication
controllers and services that comprise a Deployment Manager instance from a predefined
configuration file, as follows:
kubectl create -f install.yaml
That's it. You should now be able to see Deployment Manager running in your cluster using:
kubectl get pod,rc,service
If you see replication controllers named expandybird-rc, manager-rc and resourcifier-rc with pods that are READY, and services with corresponding names, then Deployment Manager is up and running.
Note that you can also tear down Deployment Manager using the same file, with:
kubectl delete -f install.yaml
The easiest way to interact with Deployment Manager, now that it's up and running,
is to use a kubectl proxy:
kubectl proxy --port=8001 &
This command will start a proxy that lets you interact with the Kubernetes api server through port 8001 on you local host. However, there are other ways to access Deployment Manager. We won't go into them here, but if you know how to access services running on Kubernetes, you should be able to use any of the supported methods to access Deployment Manager.
Next, you're going to deploy the canonical guestbook example to your Kubernetes cluster.
client --name guestbook --service=http://localhost:8001/api/v1/proxy/namespaces/default/services/manager-service:manager examples/guestbook/guestbook.yaml
You should now have guestbook up and running. To verify, get the list of services running on the cluster:
kubectl get service
You should see frontend-service running. If your cluster supports external load balancing, it will have an external IP assigned to it, and you should be able to navigate to it in your browser to see the guestbook in action.
This project runs Deployment Manager on Kubernetes as three replicated services. By default, prebuilt images stored in Google Container Registry are used to create them. However, you can build your own container images and push them to your own project in the registry.
To build and push your own images to Google Container Registry, first set the environment variable PROJECT to the name of a project known to gcloud. Then, run the following command:
make push
There is a more detailed design document available.
The project is still under active development, so you might run into issues. If you do, please don't be shy about letting us know, or better yet, contributing a fix or feature. We use the same contribution conventions as the main Kubernetes repository.