Deploy to hybrid Linux/Windows Kubernetes clusters

How to run Dapr apps on Kubernetes clusters with windows nodes

Dapr supports running on kubernetes clusters with windows nodes. You can run your Dapr microservices exclusively on Windows, exclusively on Linux, or a combination of both. This is helpful to users who may be doing a piecemeal migration of a legacy application into a Dapr Kubernetes cluster.

Kubernetes uses a concept called node affinity so that you can denote whether you want your application to be launched on a Linux node or a Windows node. When deploying to a cluster which has both Windows and Linux nodes, you must provide affinity rules for your applications, otherwise the Kubernetes scheduler might launch your application on the wrong type of node.

Pre-requisites

You will need a Kubernetes cluster with Windows nodes. Many Kubernetes providers support the automatic provisioning of Windows enabled Kubernetes clusters.

  1. Follow your preferred provider’s instructions for setting up a cluster with Windows enabled

  1. Once you have set up the cluster, you should see that it has both Windows and Linux nodes available

    1. kubectl get nodes -o wide
    2. NAME                                STATUS   ROLES   AGE     VERSION   INTERNAL-IP    EXTERNAL-IP      OS-IMAGE                         KERNEL-VERSION      CONTAINER-RUNTIME
    3. aks-nodepool1-11819434-vmss000000   Ready    agent   6d      v1.17.9   10.240.0.4     <none>        Ubuntu 16.04.6    LTS               4.15.0-1092-azure   docker://3.0.10+azure
    4. aks-nodepool1-11819434-vmss000001   Ready    agent   6d      v1.17.9   10.240.0.35    <none>        Ubuntu 16.04.6    LTS               4.15.0-1092-azure   docker://3.0.10+azure
    5. aks-nodepool1-11819434-vmss000002   Ready    agent   5d10h   v1.17.9   10.240.0.129   <none>        Ubuntu 16.04.6    LTS               4.15.0-1092-azure   docker://3.0.10+azure
    6. akswin000000                        Ready    agent   6d      v1.17.9   10.240.0.66    <none>        Windows Server 2019    Datacenter   10.0.17763.1339     docker://19.3.5
    7. akswin000001                        Ready    agent   6d      v1.17.9   10.240.0.97    <none>        Windows Server 2019    Datacenter   10.0.17763.1339     docker://19.3.5

Installing the Dapr control plane

If you are installing using the Dapr CLI or via a helm chart, simply follow the normal deployment procedures: Installing Dapr on a Kubernetes cluster

Affinity will be automatically set for kubernetes.io/os=linux. This will be sufficient for most users, as Kubernetes requires at least one Linux node pool.

Note: Dapr control plane containers are built and tested for both windows and linux, however, we generally recommend using the linux control plane containers. They tend to be smaller and have a much larger user base.

If you understand the above, but want to deploy the Dapr control plane to Windows, you can do so by setting:

  1. helm install dapr dapr/dapr --set global.daprControlPlaneOs=windows

Installing Dapr applications

Windows applications

In order to launch a Dapr application on Windows, you’ll first need to create a Docker container with your application installed. For a step by step guide see Get started: Prep Windows for containers. Once you have a docker container with your application, create a deployment YAML file with node affinity set to kubernetes.io/os: windows.

  1. Create a deployment YAML

    Here is a sample deployment with nodeAffinity set to “windows”. Modify as needed for your application.

    1. apiVersion: apps/v1
    2. kind: Deployment
    3. metadata:
    4.   name: yourwinapp
    5.   labels:
    6.     app: applabel
    7. spec:
    8.   replicas: 1
    9.   selector:
    10.     matchLabels:
    11.       app: applablel
    12.   template:
    13.     metadata:
    14.       labels:
    15.         app: applabel
    16.       annotations:
    17.         dapr.io/enabled: "true"
    18.         dapr.io/id: "addapp"
    19.         dapr.io/port: "6000"
    20.         dapr.io/config: "appconfig"
    21.     spec:
    22.       containers:
    23.       - name: add
    24.         image: yourreponsitory/your-windows-dapr-container:your-tag
    25.         ports:
    26.         - containerPort: 6000
    27.         imagePullPolicy: Always
    28.       affinity:
    29.         nodeAffinity:
    30.           requiredDuringSchedulingIgnoredDuringExecution:
    31.             nodeSelectorTerms:
    32.               - matchExpressions:
    33.                 - key: kubernetes.io/os
    34.                   operator: In
    35.                   values:
    36.                   - windows

    This deployment yaml will be the same as any other dapr application, with an additional spec.template.spec.affinity section as shown above.

  2. Deploy to your Kubernetes cluster

    1. kubectl apply -f deploy_windows.yaml

Linux applications

If you have already got a dapr application with runs on Linux, you’ll still need to add affinity rules as above, but choose linux affinity instead.

  1. Create a deployment YAML

    Here is a sample deployment with nodeAffinity set to “linux”. Modify as needed for your application.

    1. apiVersion: apps/v1
    2. kind: Deployment
    3. metadata:
    4.   name: yourlinuxapp
    5.   labels:
    6.     app: yourlabel
    7. spec:
    8.   replicas: 1
    9.   selector:
    10.     matchLabels:
    11.       app: yourlabel
    12.   template:
    13.     metadata:
    14.       labels:
    15.         app: yourlabel
    16.       annotations:
    17.         dapr.io/enabled: "true"
    18.         dapr.io/id: "addapp"
    19.         dapr.io/port: "6000"
    20.         dapr.io/config: "appconfig"
    21.     spec:
    22.       containers:
    23.       - name: add
    24.         image: yourreponsitory/your-application:your-tag
    25.         ports:
    26.         - containerPort: 6000
    27.         imagePullPolicy: Always
    28.       affinity:
    29.         nodeAffinity:
    30.           requiredDuringSchedulingIgnoredDuringExecution:
    31.             nodeSelectorTerms:
    32.               - matchExpressions:
    33.                 - key: kubernetes.io/os
    34.                   operator: In
    35.                   values:
    36.                   - linux

  2. Deploy to your Kubernetes cluster

    1. kubectl apply -f deploy_linux.yaml

Cleanup

  1. kubectl delete -f deploy_linux.yaml
  2. kubectl delete -f deploy_windows.yaml
  3. helm uninstall dapr

Last modified September 20, 2021 : Merge pull request #1800 from greenie-msft/gRPC_proxying_video (36dff3c)