In this article we will show you how to set up the deployment of an example application (a cool voting app in our case) using werf. It is better to start with a short introduction first if you haven’t read it yet.

Prepare your host

  1. Install dependencies (Docker and Git).
  2. Install multiwerf and werf.

Make sure you have werf command available in your shell before proceeding to the next step:

werf version

Prepare your Kubernetes and Docker Registry

You should have access to the Kubernetes cluster and be able to push images to your Docker Registry. Docker Registry should also be accessible from the Kubernetes cluster to pull images.

If your Kubernetes and Docker Registry are running already:

  1. Perform the standard docker login procedure into your Docker Registry from your host.
  2. Make sure your Kubernetes cluster is accessible from your host (if the kubectl tool is already set up and running, then werf will also work just fine).


Or use one of the following instructions to set up the local Kubernetes cluster and Docker Registry in your OS:

Windows
  1. Install minikube.

  2. Start minikube:

    minikube start --driver=docker

    IMPORTANT: If you have already started minikube make sure it uses docker driver, restart minikube otherwise (by using minikube delete and start command shown above).

  3. Enable the minikube registry addon:

    minikube addons enable registry

    Output should contain a line like:

    ...
* Registry addon on with docker uses 32769 please use that instead of default 5000
...

    Remember the port 32769.

  4. Run the following port forwarder in the separate terminal replacing 32769 port with the port from the previous step:

    docker run -ti --rm --network=host alpine ash -c "apk add socat && socat TCP-LISTEN:5000,reuseaddr,fork TCP:host.docker.internal:32769"

  5. Run a service with a binding to 5000 port:

    kubectl -n kube-system expose rc/registry --type=ClusterIP --port=5000 --target-port=5000 --name=werf-registry --selector=actual-registry=true

  6. Run the following port forwarder in the separate terminal:

    kubectl port-forward --namespace kube-system service/werf-registry 5000
MacOS
  1. Install minikube.

  2. Start minikube:

    minikube start --driver=docker

    IMPORTANT: If you have already started minikube make sure it uses docker driver, restart minikube otherwise (by using minikube delete and then start command shown above).

  3. Enable the minikube registry addon:

    minikube addons enable registry

    Output should contain a line like:

    ...
* Registry addon on with docker uses 32769 please use that instead of default 5000
...

    Remember the port 32769.

  4. Run the following port forwarder in the separate terminal replacing 32769 port with the port from the step 3:

    docker run -ti --rm --network=host alpine ash -c "apk add socat && socat TCP-LISTEN:5000,reuseaddr,fork TCP:host.docker.internal:32769"

  5. Run the following port forwarder in the separate terminal replacing 32769 port with the port from the step 3:

    brew install socat
socat TCP-LISTEN:5000,reuseaddr,fork TCP:host.docker.internal:32769
Linux
  1. Install minikube.

  2. Start minikube:

    minikube start --driver=docker

    IMPORTANT: If you have already started minikube make sure it uses docker driver, restart minikube otherwise (by using minikube delete and then start command shown above).

  3. Enable the minikube registry addon:

    minikube addons enable registry

  4. Run the following port forwarder in the separate terminal:

    sudo apt-get install -y socat
socat -d -d TCP-LISTEN:5000,reuseaddr,fork TCP:$(minikube ip):5000

Deploy an example application

  1. Clone the example application’s repository:

    git clone https://github.com/werf/quickstart-application
cd quickstart-application

  2. Run the converge command using your Docker Registry for storing images (localhost:5000/quickstart-application repository in the case of a local Docker Registry).

    werf converge --repo localhost:5000/quickstart-application

NOTE: werf uses the same settings to connect to the Kubernetes cluster as the kubectl tool does: the ~/.kube/config file and the KUBECONFIG environment variable. werf also supports --kube-config and --kube-config-base64 parameters for specifying custom kubeconfig files.

Check the result

When the converge command is successfully completed, it is safe to assume that our application is up and running. Let’s check it!

Our application is a basic voting system. Go to the following URL to vote:

minikube service --namespace quickstart-application --url vote

Go to the following URL to check the result of voting:

minikube service --namespace quickstart-application --url result

How it works

To deploy an application using werf, we should define the desired state in the Git (as set out in the introduction).

  1. We have the following Dockerfiles in our repository:

    vote/Dockerfile
result/Dockerfile
worker/Dockerfile

  2. The werf.yaml file references those Dockerfiles:

    configVersion: 1
project: quickstart-application
---
image: vote
dockerfile: vote/Dockerfile
context: vote
---
image: result
dockerfile: result/Dockerfile
context: result
---
image: worker
dockerfile: worker/Dockerfile
context: worker

  3. Kubernetes templates for vote, db, redis, result and worker components of the application are described in the files of a .helm/templates/ directory. Components interact with each other as shown in the diagram:

architecture

  • A front-end web app in Python or ASP.NET Core lets you vote for one of the two options;
  • A Redis or NATS queue collects new votes;
  • A .NET Core, Java or .NET Core 2.1 worker consumes votes and stores them in…
  • A Postgres or TiDB database backed by a Docker volume;
  • A Node.js or ASP.NET Core SignalR web-app shows the results of the voting in real-time.

What’s next?

Check the “Using werf with CI/CD systems” article or refer to the guides.