Skip to content

Latest commit

 

History

History
161 lines (113 loc) · 4.63 KB

05-kwasm.md

File metadata and controls

161 lines (113 loc) · 4.63 KB

Deploy your Spin application to a WASM enabled Kubernetes cluster using kwasm

kwasm is a Kubernetes operator to install the necessary runwasi shim on existing Kubernetes clusters. At the moment, kwasm supports these Kubernetes distributions:

  • Microsoft AKS
  • Google GKE (only with Ubuntu nodes)
  • Amazon EKS (only with Ubuntu nodes)
  • DigitalOcean DOKS
  • Civo Kubernetes
  • Kind
  • Minikube
  • MicroK8s

Requirements

To install the kwasm-operator into a supported cluster, you will need to have helm installed on your machine, as well as access to a supported cluster. If you do not have one yet, you can create a new cluster with KinD:

kind create cluster

Installation

The following commands will install the kwasm operator and set up the wasmtime-spin runtime class:

# Add HELM repository if not already done
helm repo add kwasm http://kwasm.sh/kwasm-operator/

# Install KWasm operator
helm install -n kwasm --create-namespace kwasm-operator kwasm/kwasm-operator

# Create the runtime class
kubectl apply -f ../apps/05/runtimeclass.yaml

Ensure that the kwasm-operator is running correctly and that the runtimeclass has been created:

kubectl get pods -n kwasm
NAME                             READY   STATUS    RESTARTS   AGE
kwasm-operator-88788fd4c-nqmfm   1/1     Running   0          46s

kubectl get runtimeclass
NAME            HANDLER   AGE
wasmtime-spin   spin      7s

You can now select which nodes you want your Spin workloads to run on by annotating the nodes with kwasm.sh/kwasm-node=true. To annotate all nodes, you can run the following command:

kubectl annotate node --all kwasm.sh/kwasm-node=true

This will trigger a job on each matching node in the cluster to install the containerd shims. Make sure the jobs are finished before proceeding.

kubectl get jobs -n kwasm
NAME                                 COMPLETIONS   DURATION   AGE
kind-control-plane-provision-kwasm   1/1           10s        76s

Deploy a Spin workload to the cluster

You are now ready to run Spin workloads on your cluster. To deploy the demo application, execute this command:

kubectl apply -f ../apps/05/spin.yaml

Let's verify that the service has been created and that the pod is in the Running state.

kubectl get svc
NAME         TYPE           CLUSTER-IP     EXTERNAL-IP   PORT(S)        AGE
kubernetes   ClusterIP      10.96.0.1      <none>        443/TCP        7m33s
wasm-spin    LoadBalancer   10.96.197.81   <pending>     80:31820/TCP   2m23s

kubectl get pods
NAME                         READY   STATUS    RESTARTS   AGE
wasm-spin-74cb5d68c8-rt7h6   1/1     Running   0          5s

Call the Spin application

Notice that the wasm-spin service does not have an external IP. This is due to the KinD cluster running locally without any LoadBalancer provider installed. For the purpose of this demo, we will port-forward this service. If your cluster was able to assign an external IP, feel free to skip the port-forwarding part and use the external IP instead.

kubectl port-forward svc/wasm-spin 8080:80
Forwarding from 127.0.0.1:8080 -> 80
Forwarding from [::1]:8080 -> 80

Now that we are forwarding port 8080 to the Spin service, let's make a request using curl:

curl localhost:8080/hello
Hello world from Spin!

Great! We can see that we got a response from our Spin workload!

Let's check the logs of the pod:

kubectl logs wasm-spin-74cb5d68c8-rt7h6
Hello, world! You should see me in pod logs

Congrats! You now know how to run a WASM workload on a standard Kubernetes cluster using Spin and kwasm-operator!

Deploy the magic 8 ball app to the cluster

We put the final version of the magic 8 ball application into apps/05/magic-8-ball. For now, this is a normal Spin project without any K8s specific configuration.

To build the project, run go mod download, then spin build.

To deploy the application to Kubernetes, we need to build the image and create a deployment manifest. We can do both by running

spin k8s scaffold <image repo>

where image repo is e.g. ghcr.io/<your github username>.

This generates a Dockerfile and a deploy.yaml. Feel free to inspect both files.

To build and push the image, run

spin k8s build

spin k8s push

Make sure your cluster is able to pull the image from the registry (either by creating a pull secret, or by making the image public).

To actually deploy the application to the cluster, run spin k8s deploy.

You can now access the magic 8-ball application as you did with the other application previously.

Cleanup

Feel free to delete the KinD cluster by running

kind delete cluster