Source code of this series can be found at https://github.com/thangchung/dapr-labs/tree/main/polyglot.
In this final part, we will install k3d, and containerd-wasm-shims which run runwasi inside. It allows us to run WASM/WASI workload with the annotation runtimeClassName: wasmtime-spin, and run a Docker container (containerd format). If you don't declare the previous annotation on the YAML deployment script.
Let's get started with the setup as the image below.
Setup Docker (buildx), k3d
# ref: https://github.com/docker/docker-install
> curl -fsSL https://get.docker.com -o get-docker.sh
> sh get-docker.sh --version 24.0
> docker --version
Docker version 24.0.5, build ced0996
# install k3d
> wget -q -O - https://raw.githubusercontent.com/k3d-io/k3d/main/install.sh | bash
Dockerized apps
# Product API
FROM --platform=${BUILDPLATFORM} rust:1.67 AS build
RUN rustup target add wasm32-wasi
COPY . /product
WORKDIR /product
RUN cargo build --target wasm32-wasi --release
FROM scratch
COPY --from=build /product/target/wasm32-wasi/release/product_api.wasm /target/wasm32-wasi/release/product_api.wasm
COPY ./spin.toml /spin.toml
# Counter API, .NET 8
> dotnet publish ./counter-api/counter-api.csproj --os linux --arch x64 /t:PublishContainer -c Release
docker tag counter-api:latest ghcr.io/thangchung/dapr-labs/counter-api-polyglot:1.0.0
# Barista API
FROM --platform=${BUILDPLATFORM} rust:1.67 AS build
RUN rustup target add wasm32-wasi
COPY . /barista
WORKDIR /barista
RUN cargo build --target wasm32-wasi --release
FROM scratch
COPY --from=build /barista/target/wasm32-wasi/release/barista_api.wasm /target/wasm32-wasi/release/barista_api.wasm
COPY ./spin.toml /spin.toml
# Kitchen API
FROM --platform=${BUILDPLATFORM} rust:1.67 AS build
RUN rustup target add wasm32-wasi
COPY . /kitchen
WORKDIR /kitchen
RUN cargo build --target wasm32-wasi --release
FROM scratch
COPY --from=build /kitchen/target/wasm32-wasi/release/kitchen_api.wasm /target/wasm32-wasi/release/kitchen_api.wasm
COPY ./spin.toml /spin.toml
ENTRYPOINT ["/"]
Build and push it into GitHub artifacts:
> docker login ghcr.io -u <your username>
It asks you to provide the password (PAT), please go to your developer profile to generate it.
> docker buildx build -f Dockerfile --platform wasi/wasm,linux/amd64,linux/arm64 -t ghcr.io/thangchung/dapr-labs/product-api-spin:1.0.0 . --push
> docker push ghcr.io/thangchung/dapr-labs/counter-api-polyglot:1.0.0
> docker buildx build -f Dockerfile --platform wasi/wasm,linux/amd64,linux/arm64 -t ghcr.io/thangchung/dapr-labs/barista-api-spin:1.0.0 . --push
> docker buildx build -f Dockerfile --platform wasi/wasm,linux/amd64,linux/arm64 -t ghcr.io/thangchung/dapr-labs/kitchen-api-spin:1.0.0 . --push
Ship them to Kubernetes
# Product API
apiVersion: apps/v1
kind: Deployment
metadata:
name: product-api
spec:
replicas: 1
selector:
matchLabels:
app: product-api
template:
metadata:
labels:
app: product-api
annotations:
dapr.io/enabled: "true"
dapr.io/app-id: "product-api"
dapr.io/app-port: "80"
dapr.io/enable-api-logging: "true"
spec:
runtimeClassName: wasmtime-spin
containers:
- name: product-api
image: ghcr.io/thangchung/dapr-labs/product-api-spin:1.0.1
command: ["/"]
env:
- name: RUST_BACKTRACE
value: "1"
resources: # limit the resources to 128Mi of memory and 100m of CPU
limits:
cpu: 100m
memory: 128Mi
requests:
cpu: 100m
memory: 128Mi
---
apiVersion: v1
kind: Service
metadata:
name: product-api
spec:
type: LoadBalancer
ports:
- protocol: TCP
port: 5001
targetPort: 80
selector:
app: product-api
# Counter API
apiVersion: apps/v1
kind: Deployment
metadata:
name: counter-api
spec:
replicas: 1
selector:
matchLabels:
app: counter-api
template:
metadata:
labels:
app: counter-api
annotations:
dapr.io/enabled: "true"
dapr.io/app-id: "counter-api"
dapr.io/app-port: "8080"
dapr.io/enable-api-logging: "true"
spec:
containers:
- name: counter-api
image: ghcr.io/thangchung/dapr-labs/counter-api-polyglot:1.0.0
env:
- name: ProductCatalogAppDaprName
value: "product-api"
resources: # limit the resources to 128Mi of memory and 100m of CPU
limits:
cpu: 100m
memory: 128Mi
requests:
cpu: 100m
memory: 128Mi
---
apiVersion: v1
kind: Service
metadata:
name: counter-api
spec:
type: LoadBalancer
ports:
- protocol: TCP
port: 5002
targetPort: 8080
selector:
app: counter-api
# Barista API
apiVersion: apps/v1
kind: Deployment
metadata:
name: barista-api
spec:
replicas: 1
selector:
matchLabels:
app: barista-api
template:
metadata:
labels:
app: barista-api
annotations:
dapr.io/enabled: "true"
dapr.io/app-id: "barista-api"
dapr.io/app-port: "80"
dapr.io/enable-api-logging: "true"
spec:
runtimeClassName: wasmtime-spin
containers:
- name: barista-api
image: ghcr.io/thangchung/dapr-labs/barista-api-spin:1.0.0
# imagePullPolicy: Always
command: ["/"]
env:
- name: RUST_BACKTRACE
value: "1"
- name: DAPR_URL
value: "http://localhost:3500"
resources: # limit the resources to 128Mi of memory and 100m of CPU
limits:
cpu: 100m
memory: 128Mi
requests:
cpu: 100m
memory: 128Mi
---
apiVersion: v1
kind: Service
metadata:
name: barista-api
spec:
type: LoadBalancer
ports:
- protocol: TCP
port: 5003
targetPort: 80
selector:
app: barista-api
# Kitchen API
apiVersion: apps/v1
kind: Deployment
metadata:
name: kitchen-api
spec:
replicas: 1
selector:
matchLabels:
app: kitchen-api
template:
metadata:
labels:
app: kitchen-api
annotations:
dapr.io/enabled: "true"
dapr.io/app-id: "kitchen-api"
dapr.io/app-port: "80"
dapr.io/enable-api-logging: "true"
spec:
runtimeClassName: wasmtime-spin
containers:
- name: kitchen-api
image: ghcr.io/thangchung/dapr-labs/kitchen-api-spin:1.0.0
# imagePullPolicy: Always
command: ["/"]
env:
- name: RUST_BACKTRACE
value: "1"
- name: DAPR_URL
value: "http://localhost:3500"
resources: # limit the resources to 128Mi of memory and 100m of CPU
limits:
cpu: 100m
memory: 128Mi
requests:
cpu: 100m
memory: 128Mi
---
apiVersion: v1
kind: Service
metadata:
name: kitchen-api
spec:
type: LoadBalancer
ports:
- protocol: TCP
port: 5004
targetPort: 80
selector:
app: kitchen-api
# ingress.yaml
# Middleware
# Strip prefix /spin
apiVersion: traefik.containo.us/v1alpha1
kind: Middleware
metadata:
name: strip-prefix
spec:
stripPrefix:
forceSlash: false
prefixes:
- /p
- /c
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: polyglot-wasm-ingress
annotations:
ingress.kubernetes.io/ssl-redirect: "false"
kubernetes.io/ingress.class: traefik
traefik.ingress.kubernetes.io/router.middlewares: default-strip-prefix@kubernetescrd
spec:
rules:
- http:
paths:
- path: /p
pathType: Prefix
backend:
service:
name: product-api
port:
number: 5001
- path: /c
pathType: Prefix
backend:
service:
name: counter-api
port:
number: 5002
# create Kubernetes cluster for WASM/WASI
> k3d cluster create wasm-cluster --image ghcr.io/deislabs/containerd-wasm-shims/examples/k3d:v0.9.0 -p "8081:80@loadbalancer" --agents 2
# add redis
> helm install my-redis oci://registry-1.docker.io/bitnamicharts/redis --set architecture=standalone --set global.redis.password=P@ssw0rd
# for demo only, otherwise need use Dapr with Helm chart
> dapr init -k --runtime-version 1.11.2
# remember to edit dapr component YAML, and put password for redis
> kubectl apply -f components-k8s
> kubectl apply -f iac/kind-spin
Make sure you can query:
> kubectl get component
NAME AGE
baristapubsub 2d1h
kitchenpubsub 2d1h
statestore 2d1h
> kubectl get subscription
NAME AGE
barista-ordered-subscription 2d1h
barista-updated-subscription 2d1h
kitchen-ordered-subscription 2d1h
kitchen-updated-subscription 2d1h
> kubectl get po
NAME READY STATUS RESTARTS AGE
my-redis-master-0 1/1 Running 24 (55m ago) 10d
counter-api-8bdc488b7-h96pw 2/2 Running 16 (55m ago) 2d1h
product-api-8ccbc56b-ql5bt 2/2 Running 12 (54m ago) 2d1h
kitchen-api-54f7c588cb-ffnzl 2/2 Running 12 (55m ago) 2d1h
barista-api-6f5b4d6fb8-pxbqc 2/2 Running 12 (55m ago) 2d1h
> kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S)
AGE
kubernetes ClusterIP 10.43.0.1 <none> 443/TCP
10d
my-redis-headless ClusterIP None <none> 6379/TCP
10d
my-redis-master ClusterIP 10.43.109.123 <none> 6379/TCP
10d
barista-api-dapr ClusterIP None <none> 80/TCP,50001/TCP,50002/TCP,9090/TCP 2d1h
counter-api-dapr ClusterIP None <none> 80/TCP,50001/TCP,50002/TCP,9090/TCP 2d1h
kitchen-api-dapr ClusterIP None <none> 80/TCP,50001/TCP,50002/TCP,9090/TCP 2d1h
product-api-dapr ClusterIP None <none> 80/TCP,50001/TCP,50002/TCP,9090/TCP 2d1h
product-api LoadBalancer 10.43.58.9 172.19.0.2,172.19.0.4,172.19.0.5 5001:30896/TCP
2d1h
kitchen-api LoadBalancer 10.43.105.225 172.19.0.2,172.19.0.4,172.19.0.5 5004:32611/TCP
2d1h
counter-api LoadBalancer 10.43.19.86 172.19.0.2,172.19.0.4,172.19.0.5 5002:30630/TCP
2d1h
barista-api LoadBalancer 10.43.105.93 172.19.0.2,172.19.0.4,172.19.0.5 5003:31258/TCP
2d1h
> kubectl get ing
NAME CLASS HOSTS ADDRESS PORTS AGE
polyglot-wasm-ingress <none> * 172.19.0.2,172.19.0.4,172.19.0.5 80 2d1h
If everything is okay, then we can play around with Rest Client at client.k3d.http.
Summary
So we have already walked through 4 parts of how can we build the polyglot apps with Docker, WebAssembly (Spin), Dapr, and Kubernetes (k3d). There are still obstacles, but the future is bright for WASM apps on Kubernetes due to the very active process from the community.
What's next? The WebAssembly/WASI has a clear roadmap. In WasmCon 2023 recently, they mentioned Component Model and I think that is a final abstraction in the computing unit. It helps a lot in building what Luke Wagner calls Modularity without microservices, and we will invest time on it as well. See you in the next posts to dive more into this kind of component model.
Stay stun!
Top comments (0)