Muokkaa

Jaa

Migrate a simple app from Service Fabric to AKS

  • Artikkeli

This article provides an example workload migration to help you implement some of the conceptual information provided in Migrate your workload from Service Fabric to AKS. That article provides information about Azure Kubernetes Service (AKS) and a comparison of AKS with Azure Service Fabric. It also describes considerations to take into account when you migrate your workloads.

This example focuses on Windows-based Service Fabric applications that have already been containerized. Azure Service Fabric and Azure Kubernetes Service both support Windows and Linux containers. If your application isn't containerized, consider investigating whether you can containerize it. Building a container image for your application is a prerequisite for deploying the application to Azure Kubernetes Service. If the application depends on Service Fabric programming models (Reliable Services, Reliable Actors, ASP.NET Core, and guest executables), you'll probably need to do some refactoring.

For information about containerizing your application, see Prepare an application for AKS. For information about containerizing an ASP.NET application, see ASP.NET app containerization and migration to AKS.

Prerequisites

Before you start the migration, you need:

  • An application container image that's stored in Azure Container Registry.

  • A Bash environment that you can use to configure your Azure resources.

  • The kubectl Kubernetes command-line tool. If it's not already available in your environment, you can install it by running this command:

    az aks install-cli

Migration steps

The first step is to set up the resources that you need to build a Windows node pool in Kubernetes. To do that, follow the guidance in Create a Windows Server container on an AKS cluster, but be sure to stop when you reach the "Deploy the application" section. At that point, follow the instructions in this article.

The translation of the Service Fabric configuration manifest to an AKS manifest is an important step. The following sections show:

  • Service manifest XML that you might use for a basic Service Fabric deployment.
  • A functionally equivalent AKS manifest that creates Kubernetes Deployment and Service objects.

The two manifests don't map one-to-one because they're based on the functional paradigms that are specific to each service, but their intents are the same. (In these samples, variables use the format <VARIABLE DESCRIPTION>.)

In the AKS manifest, a Deployment object provides declarative updates for Pods and ReplicaSets. A Service object exposes an application that's running on a set of pods as a network service.

Sample Service Fabric service manifest

<?xml version="1.0" encoding="utf-8"?>

<ServiceManifest Name="<APP NAME>"
                 Version="1.0.0"
                 xmlns="http://schemas.microsoft.com/2011/01/fabric"
                 xmlns:xsd="https://www.w3.org/2001/XMLSchema"
                 xmlns:xsi="https://www.w3.org/2001/XMLSchema-instance">
  <ServiceTypes>
    <StatelessServiceType ServiceTypeName="<SERVICE NAME>" UseImplicitHost="true" />
  </ServiceTypes>
 
  <!-- Code package is your service executable file. -->
  <CodePackage Name="code" Version="1.0.0">
    <EntryPoint>
      <ContainerHost>
        <ImageName><YOUR IMAGE></ImageName>
        <Commands></Commands>
      </ContainerHost>
    </EntryPoint>
    <!-- Pass environment variables to your container. -->
    <EnvironmentVariables>
      <EnvironmentVariable Name="HttpGatewayPort" Value=""/>
      <EnvironmentVariable Name="BackendServiceName" Value=""/>
    </EnvironmentVariables>
 
  </CodePackage>
 
  <ConfigPackage Name="Config" Version="1.0.0" />
 
  <Resources>
    <Endpoints>
      <Endpoint Name="<HTTP ENDPOINT NAME>" UriScheme="http" Port="80" Protocol="http"/>
    </Endpoints>
  </Resources>
</ServiceManifest>

Sample AKS manifest

apiVersion: apps/v1
kind: Deployment
metadata:
  name: <APP NAME>
  labels:
    app: <APP NAME>
spec:
  replicas: 1
  template:
    metadata:
      name: <APP NAME>
      labels:
        app: <APP NAME>
    spec:
      nodeSelector:
        "kubernetes.io/os": windows
      containers:
      - name: <SERVICE NAME>
        image: <YOUR IMAGE>
        resources:
          limits:
            cpu: 1
            memory: 800M
        ports:
          - containerPort: 80
    - env:
        - name: HttpGatewayPort
          value: ""
        - name: BackendServiceName
          value: ""
  selector:
    matchLabels:
      app: <APP NAME>
---
apiVersion: v1
kind: Service
metadata:
  name: <SERVICE NAME>
spec:
  type: LoadBalancer
  ports:
  - protocol: TCP
    port: 80
  selector:
    app: <SERVICE NAME>

Kubernetes provides a large set of configuration options, which is useful for experienced developers. But manifests can become large and complex when you use too many of them. To learn about implementing a simple migration, we recommend that you review Deployments and YAML manifests.

After you have your manifest, you just need to apply it, and you can watch your app:

kubectl apply -f <YOUR MANIFEST>.yaml
kubectl get deploy <APP NAME>
kubectl get service <SERVICE NAME> --watch

Note

This example uses the default Kubernetes namespace, which is generally used only for basic scenarios. In Kubernetes, namespaces provide a mechanism for isolating groups of resources within a single cluster. Namespaces are important for enforcing security, networking, and resource boundaries. To determine a configuration that works best for your application, see the Kuberetes namespaces documentation.

Contributors

This article is maintained by Microsoft. It was originally written by the following contributors.

Principal authors:

Other contributors:

To see non-public LinkedIn profiles, sign in to LinkedIn.

Next steps