Tutorial: Knative + Tekton

Linux Foundation Open Source Summit NA 2020

Last Update: 2020/08/27

1. Setup Environment

1.1 Setup Kubernetes Clusters

1.1.1 IBM Free Kubernetes Cluster

Get a free Kubernetes cluster on IBM Cloud, also check out the booth at OSS-NA IBM booth during the conference how to get $200 credit.

Select cluster from IBM Cloud console
Click the drop down Action menu on the top right and select Connect via CLI and follow the commands.

ibmcloud login -a cloud.ibm.com -r <REGION> -g <IAM_RESOURCE_GROUP>

Set the Kubernetes context
```
ibmcloud ks cluster config -c mycluster
```
Verify that you can connect to your cluster.
```
kubectl version --short
```

1.1.2 Kubernetes with Minikube

Install minikube Linux, MacOS, or Windows. This tutorial was tested with version v1.12.0. You print current and latest version number
```
minikube update-check
```
Configure your cluster 2 CPUs, 2 GB Memory, and version of kubernetes v1.18.5. If you already have a minikube with different config, you need to delete it for new configuration to take effect or create a new profile.
```
minikube delete
minikube config set cpus 2
minikube config set memory 2048
minikube config set kubernetes-version v1.18.5
```
Start your minikube cluster
```
minikube start
```
Verify versions if the kubectl, the cluster, and that you can connect to your cluster.
```
kubectl version --short
```

1.1.3 Kubernetes with Kind (Kubernetes In Docker)

Install kind Linux, MacOS, or Windows. This tutorial was tested with version v0.8.1. You can verify version with
```
kind --version
```

A kind cluster manifest file clusterconfig.yaml is already provided, you can customize it. We are exposing port 80 on they host to be later use by the Knative Kourier ingress. To use a different version of kubernetes check the image digest to use from the kind release page

kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
- role: control-plane
  image: kindest/node:v1.18.2@sha256:7b27a6d0f2517ff88ba444025beae41491b016bc6af573ba467b70c5e8e0d85f
  extraPortMappings:
  - containerPort: 31080 # expose port 31380 of the node to port 80 on the host, later to be use by kourier ingress
    hostPort: 80

Create and start your cluster, we specify the config file above

kind create cluster --name knative --config kind/clusterconfig.yaml

Verify the versions of the client kubectl and the cluster api-server, and that you can connect to your cluster.
```
kubectl cluster-info --context kind-knative
```

1.1.4 Kubernetes with Katacoda

For a short version of this tutorial try it out on my Katacoda Scenario

1.2 Setup Command Line Interface (CLI) Tools

Kubernetes CLI kubectl
Knative CLI kn
Tekton CLI tkn

1.3 Setup Container Registry

Get access to a container registry such as quay, dockerhub, or your own private registry instance from a Cloud provider such as IBM Cloud 😉. On this tutorial we are going to use Dockerhub
Set the environment variables REGISTRY_SERVER, REGISTRY_NAMESPACE and REGISTRY_PASSWORD, The REGISTRY_NAMESPACE most likely would be your dockerhub username. For Dockerhub use docker.io as the value for REGISTRY_SERVER
```
REGISTRY_SERVER='docker.io'
REGISTRY_NAMESPACE='REPLACEME_DOCKER_USERNAME_VALUE'
REGISTRY_PASSWORD='REPLACEME_DOCKER_PASSWORD'
```

You can use the file .template.env as template for the variables

cp .template.env .env
# edit the file .env with variables and credentials the source the file
source .env

1.4 Setup Git

Get access to a git server such as gitlab, github, or your own private git instance from a Cloud provider such as IBM Cloud 😉. On this tutorial we are going to use GitHub

Fork this repository https://github.com/csantanapr/knative-tekton
Set the environment variable GIT_REPO_URL to the url of your fork, not mine.
```
GIT_REPO_URL='https://github.com/REPLACEME/knative-tekton'
```

Clone the repository and change directory

git clone $GIT_REPO_URL
cd knative-tekton

You can use the file .template.env as template for the variables

cp .template.env .env
# edit the file .env with variables and credentials the source the file
source .env

2. Install Knative Serving

Install Knative Serving in namespace knative-serving

kubectl apply -f https://github.com/knative/serving/releases/download/v0.17.1/serving-crds.yaml
kubectl apply -f https://github.com/knative/serving/releases/download/v0.17.1/serving-core.yaml
kubectl wait deployment activator autoscaler controller webhook --for=condition=Available -n knative-serving

Install Knative Layer kourier in namespace kourier-system

kubectl apply -f https://github.com/knative/net-kourier/releases/download/v0.17.0/kourier.yaml
kubectl wait deployment 3scale-kourier-gateway --for=condition=Available -n kourier-system
kubectl wait deployment 3scale-kourier-control --for=condition=Available -n knative-serving

Set the environment variable EXTERNAL_IP to External IP Address of the Worker Node If using minikube:

EXTERNAL_IP=$(minikube ip)
echo EXTERNAL_IP=$EXTERNAL_IP

If using kind:

EXTERNAL_IP="127.0.0.1"

If using IBM Kubernetes:

EXTERNAL_IP=$(kubectl get nodes -o jsonpath='{.items[0].status.addresses[?(@.type=="ExternalIP")].address}')

Verify the value

echo EXTERNAL_IP=$EXTERNAL_IP

Set the environment variable KNATIVE_DOMAIN as the DNS domain using nip.io
```
KNATIVE_DOMAIN="$EXTERNAL_IP.nip.io"
echo KNATIVE_DOMAIN=$KNATIVE_DOMAIN
```
Double check DNS is resolving
```
dig $KNATIVE_DOMAIN
```

Configure DNS for Knative Serving

kubectl patch configmap -n knative-serving config-domain -p "{\"data\": {\"$KNATIVE_DOMAIN\": \"\"}}"

Configure Kourier to listen for http port 80 on the node

If using Kind then use this

cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Service
metadata:
  name: kourier-ingress
  namespace: kourier-system
  labels:
    networking.knative.dev/ingress-provider: kourier
spec:
  type: NodePort
  selector:
    app: 3scale-kourier-gateway
  ports:
    - name: http2
      nodePort: 31080
      port: 80
      targetPort: 8080
EOF

If not using Kind then use this

cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Service
metadata:
  name: kourier-ingress
  namespace: kourier-system
  labels:
    networking.knative.dev/ingress-provider: kourier
spec:
  selector:
    app: 3scale-kourier-gateway
  ports:
    - name: http2
      port: 80
      targetPort: 8080
  externalIPs:
    - $EXTERNAL_IP
EOF

Configure Knative to use Kourier

kubectl patch configmap/config-network \
  --namespace knative-serving \
  --type merge \
  --patch '{"data":{"ingress.class":"kourier.ingress.networking.knative.dev"}}'

Verify that Knative is Installed properly all pods should be in Running state and our kourier-ingress service configured.

kubectl get pods -n knative-serving
kubectl get pods -n kourier-system
kubectl get svc  -n kourier-system kourier-ingress

3. Using Knative to Run Serverless Applications

Set the environment variable SUB_DOMAIN to the kubernetes namespace with Domain name <namespace>.<domainname> this way we can use any kubernetes namespace other than default

CURRENT_CTX=$(kubectl config current-context)
CURRENT_NS=$(kubectl config view -o=jsonpath="{.contexts[?(@.name==\"${CURRENT_CTX}\")].context.namespace}")
if [[ -z "${CURRENT_NS}" ]]; then CURRENT_NS="default" fi
SUB_DOMAIN="$CURRENT_NS.$KNATIVE_DOMAIN"
echo "\n\nSUB_DOMAIN=$SUB_DOMAIN"

3.1 Create Knative Service

Using the Knative CLI kn deploy an application usig a Container Image
```
kn service create hello --image gcr.io/knative-samples/helloworld-go --autoscale-window 15s
```
You can set a lower window. The service is scaled to zero if no request was receivedin during that time.
```
--autoscale-window 10s
```
You can list your service
```
kn service list hello
```
Use curl to invoke the Application
```
curl http://hello.$SUB_DOMAIN
```
It should print
```
Hello World!
```

You can watch the pods and see how they scale down to zero after http traffic stops to the url

kubectl get pod -l serving.knative.dev/service=hello -w

Output should look like this after a few seconds when http traffic stops:

NAME                                     READY   STATUS
hello-r4vz7-deployment-c5d4b88f7-ks95l   2/2     Running
hello-r4vz7-deployment-c5d4b88f7-ks95l   2/2     Terminating
hello-r4vz7-deployment-c5d4b88f7-ks95l   1/2     Terminating
hello-r4vz7-deployment-c5d4b88f7-ks95l   0/2     Terminating

Try to access the url again, and you will see the new pods running again.

NAME                                     READY   STATUS
hello-r4vz7-deployment-c5d4b88f7-rr8cd   0/2     Pending
hello-r4vz7-deployment-c5d4b88f7-rr8cd   0/2     ContainerCreating
hello-r4vz7-deployment-c5d4b88f7-rr8cd   1/2     Running
hello-r4vz7-deployment-c5d4b88f7-rr8cd   2/2     Running

Some people call this Serverless 🎉 🌮 🔥

3.2 Updating the Knative service

Update the service hello with a new environment variable TARGET
```
kn service update hello --env TARGET="World from v1" 
```

Now invoke the service

curl http://hello.$SUB_DOMAIN

It should print

Hello World from v1!

3.3 Knative Service Traffic Splitting

Update the service hello by updating the environment variable TARGET, tag the previous version v1, send 25% traffic to this new version and leaving 75% of the traffic to v1

kn service update hello \
 --env TARGET="Knative from v2" \
 --tag $(kubectl get ksvc hello --template='{{.status.latestReadyRevisionName}}')=v1 \
 --traffic v1=75,@latest=25

Describe the service to see the traffic split details

kn service describe  hello

Should print this

Name:       hello
Namespace:  debug
Age:        6m
URL:        http://hello.$SUB_DOMAIN

Revisions:  
  25%  @latest (hello-mshgs-3) [3] (26s)
        Image:  gcr.io/knative-samples/helloworld-go (pinned to 5ea96b)
  75%  hello-tgzmt-2 #v1 [2] (6m)
        Image:  gcr.io/knative-samples/helloworld-go (pinned to 5ea96b)

Conditions:  
  OK TYPE                   AGE REASON
  ++ Ready                  21s 
  ++ ConfigurationsReady    24s 
  ++ RoutesReady            21s

Invoke the service usign a while loop you will see the message Hello Knative from v2 25% of the time

while true; do
curl http://hello.$SUB_DOMAIN
sleep 0.5
done

Should print this

Hello World from v1!
Hello Knative from v2!
Hello World from v1!
Hello World from v1!

Update the service this time dark launch new version v3 on a specific url, zero traffic will go to this version from the main url of the service

kn service update hello \
    --env TARGET="OSS NA 2020 from v3" \
    --tag $(kubectl get ksvc hello --template='{{.status.latestReadyRevisionName}}')=v2 \
    --tag @latest=v3 \
    --traffic v1=75,v2=25,@latest=0

Describe the service to see the traffic split details, v3 doesn't get any traffic

kn service describe  hello

Should print this

Revisions:  
    +  @latest (hello-wkyty-4) #v3 [4] (1m)
        Image:  gcr.io/knative-samples/helloworld-go (pinned to 5ea96b)
25%  hello-fbzqf-3 #v2 [3] (6m)
        Image:  gcr.io/knative-samples/helloworld-go (pinned to 5ea96b)
75%  hello-kcspq-2 #v1 [2] (7m)
        Image:  gcr.io/knative-samples/helloworld-go (pinned to 5ea96b)

The latest version of the service is only available with an url prefix v3-, go ahead and invoke the latest directly.
```
curl http://v3-hello.$SUB_DOMAIN
```
It shoud print this
```
Hello OSS NA from v3!
```
We are happy with our darked launch version of the application, lets turn it live to 100% of the users on the default url
```
kn service update hello --traffic @latest=100
```

Describe the service to see the traffic split details, @latest now gets 100% of the traffic

kn service describe  hello

Should print this

Revisions:  
100%  @latest (hello-wkyty-4) #v3 [4] (4m)
        Image:  gcr.io/knative-samples/helloworld-go (pinned to 5ea96b)
    +  hello-fbzqf-3 #v2 [3] (8m)
        Image:  gcr.io/knative-samples/helloworld-go (pinned to 5ea96b)
    +  hello-kcspq-2 #v1 [2] (9m)
        Image:  gcr.io/knative-samples/helloworld-go (pinned to 5ea96b)

If we invoke the service in a loop you will see that 100% of the traffic is directed to version v3 of our application

while true; do
curl http://hello.$SUB_DOMAIN
sleep 0.5
done

Should print this

Hello OSS NA 2020 from v3!
Hello OSS NA 2020 from v3!
Hello OSS NA 2020 from v3!
Hello OSS NA 2020 from v3!

By using tags the custom urls with tag prefix are still available, in case you want to access an old revision of the application

curl http://v1-hello.$SUB_DOMAIN 
curl http://v2-hello.$SUB_DOMAIN 
curl http://v3-hello.$SUB_DOMAIN

It should print

Hello World from v1!
Hello Knative from v2!
Hello OSS NA 2020 from v3!

Now that you have your service configure and deploy, you want to reproduce this using a kubernetes manifest using YAML in a different namespace or cluster. You can define your Knative service using the following YAML you can use the command kn service export

Show me the Knative YAML

---
apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: hello
spec:
  template:
    metadata:
      name: hello-v1
    spec:
      containers:
        - image: gcr.io/knative-samples/helloworld-go
          env:
            - name: TARGET
              value: World from v1
---
apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: hello
spec:
  template:
    metadata:
      name: hello-v2
    spec:
      containers:
        - image: gcr.io/knative-samples/helloworld-go
          env:
            - name: TARGET
              value: Knative from v2
---
apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: hello
spec:
  template:
    metadata:
      name: hello-v3
    spec:
      containers:
        - image: gcr.io/knative-samples/helloworld-go
          env:
            - name: TARGET
              value: OSS NA 2020 from v3
  traffic:
    - latestRevision: false
      percent: 0
      revisionName: hello-v1
      tag: v1
    - latestRevision: false
      percent: 0
      revisionName: hello-v2
      tag: v2
    - latestRevision: true
      percent: 100
      tag: v3

If you want to deploy usign YAML, delete the Application with kn and redeploy with kubectl

kn service delete hello
kubectl apply -f knative/v1.yaml
kubectl apply -f knative/v2.yaml
kubectl apply -f knative/v3.yaml

Try the service again

while true; do
curl http://hello.$SUB_DOMAIN 
done

Delete the Application and all it's revisions
```
kn service delete hello
```

4. Install Tekton Pipelines

4. Install Tekton

4.1 Install Tekton Pipelines

Install Tekton Pipelines in namespace tekton-pipelines

kubectl apply -f https://github.com/tektoncd/pipeline/releases/download/v0.14.1/release.yaml
kubectl wait deployment tekton-pipelines-controller tekton-pipelines-webhook --for=condition=Available -n tekton-pipelines

4.2 Install Tekton Dashboard (Optional)

Install Tekton Dashboard in namespace tekton-pipelines

kubectl apply -f https://github.com/tektoncd/dashboard/releases/download/v0.7.1/tekton-dashboard-release.yaml
kubectl wait deployment tekton-dashboard --for=condition=Available -n tekton-pipelines

We can access the Tekton Dashboard serving using the Kourier Ingress using the KNATIVE_DOMAIN

cat <<EOF | kubectl apply -f -
apiVersion: networking.internal.knative.dev/v1alpha1
kind: Ingress
metadata:
  name: tekton-dashboard
  namespace: tekton-pipelines
  annotations:
    networking.knative.dev/ingress.class: kourier.ingress.networking.knative.dev
spec:
  rules:
  - hosts:
    - dashboard.tekton-pipelines.$KNATIVE_DOMAIN
    http:
      paths:
      - splits:
        - appendHeaders: {}
          serviceName: tekton-dashboard
          serviceNamespace: tekton-pipelines
          servicePort: 9097
    visibility: ExternalIP
  visibility: ExternalIP
EOF

Set an environment variable TEKTON_DASHBOARD_URL with the url to access the Dashboard

TEKTON_DASHBOARD_URL=http://dashboard.tekton-pipelines.$KNATIVE_DOMAIN
echo TEKTON_DASHBOARD_URL=$TEKTON_DASHBOARD_URL

5. Using Tekton to Build and Deploy Applications

5. Using Tekton to Build Applications

Tekton helps create composable DevOps Automation by putting together Tasks, and Pipelines

5.1 Configure Credentials and ServiceAccounts for Tekton

5.1 Configure Access for Tekton

We need to package our application in a Container Image and store this Image in a Container Registry. Since we are going to need to create secrets with the registry credentials we are going to create a ServiceAccount pipelines with the associated secret regcred. Make sure you setup your container credentials as environment variables. Checkout the Setup Container Registry in the Setup Environment section on this tutorial. This commands will print your credentials make sure no one is looking over, the printed command is what you need to run.
```
echo ""
echo kubectl create secret docker-registry regcred \
  --docker-server=\'${REGISTRY_SERVER}\' \
  --docker-username=\'${REGISTRY_NAMESPACE}\' \
  --docker-password=\'${REGISTRY_PASSWORD}\'
echo "\nRun the above command manually ^^ this avoids problems with certain charaters in your password on  the shell"
```
NOTE: If you password have some characters that are interpreted by the shell, then do NOT use environment variables, explicit enter your values in the command wrapped by single quotes '
Verify the secret regcred was created
```
kubectl describe secret regcred
```

Create a ServiceAccount pipeline that contains the secret regsecret that we just created

apiVersion: v1
kind: ServiceAccount
metadata:
  name: pipeline
secrets:
  - name: regcred

Run the following command with the provided YAML

kubectl apply -f tekton/sa.yaml

We are going to be using Tekton to deploy the Knative Service, we need to configure RBAC to provide edit access to the current namespace default to the ServiceAccount pipeline if you are using a different namespace than default edit the file tekton/rbac.yaml and provide the namespace where to create the Role and the RoleBinding fo more info check out the RBAC docs. Run the following command to grant access to sa pipelines
```
cat tekton/rbac.yaml | sed "s/namespace: default/namespace: $CURRENT_NS/g" | kubectl apply -f -
```

5.2 The Build Tekton Task

In this repository we have a sample application, you can see the source code in ./nodejs/app.js This application is using JavaScript to implement a web server, but you can use any language you want.

const app = require("express")()
const server = require("http").createServer(app)
const port = process.env.PORT || "8080"
const message = process.env.TARGET || 'Hello World'

app.get('/', (req, res) => res.send(message))
server.listen(port, function () {
    console.log(`App listening on ${port}`)
});

I provided a Tekton Task that can download source code from git, build and push the Image to a registry.

Show me the Build Task YAML

apiVersion: tekton.dev/v1beta1
kind: Task
metadata:
  name: build
spec:
  params:
    - name: repo-url
      description: The git repository url
    - name: revision
      description: The branch, tag, or git reference from the git repo-url location
      default: master
    - name: image
      description: "The location where to push the image in the form of <server>/<namespace>/<repository>:<tag>"
    - name: CONTEXT
      description: Path to the directory to use as context.
      default: .
    - name: BUILDER_IMAGE
      description: The location of the buildah builder image.
      default: quay.io/buildah/stable:v1.14.8
    - name: STORAGE_DRIVER
      description: Set buildah storage driver
      default: overlay
    - name: DOCKERFILE
      description: Path to the Dockerfile to build.
      default: ./Dockerfile
    - name: TLSVERIFY
      description: Verify the TLS on the registry endpoint (for push/pull to a non-TLS registry)
      default: "false"
    - name: FORMAT
      description: The format of the built container, oci or docker
      default: "oci"
  steps:
    - name: git-clone
      image: alpine/git
      script: |
        git clone $(params.repo-url) /source
        cd /source
        git checkout $(params.revision)
      volumeMounts:
        - name: source
          mountPath: /source
    - name: build-image
      image: $(params.BUILDER_IMAGE)
      workingdir: /source
      script: |
        echo "Building Image $(params.image)"
        buildah --storage-driver=$(params.STORAGE_DRIVER) bud --format=$(params.FORMAT) --tls-verify=$(params.TLSVERIFY) -f $(params.DOCKERFILE) -t $(params.image) $(params.CONTEXT)
        
        echo "Pushing Image $(params.image)"
        buildah  --storage-driver=$(params.STORAGE_DRIVER) push --tls-verify=$(params.TLSVERIFY) --digestfile ./image-digest $(params.image) docker://$(params.image)
      securityContext:
        privileged: true
      volumeMounts:
        - name: varlibcontainers
          mountPath: /var/lib/containers
        - name: source
          mountPath: /source
  volumes:
    - name: varlibcontainers
      emptyDir: {}
    - name: source
      emptyDir: {}

Install the provided task build like this.
```
kubectl apply -f tekton/task-build.yaml
```
You can list the task that we just created using the tkn CLI
```
tkn task ls
```
We can also get more details about the build Task using tkn task describe
```
tkn task describe build
```
Let's use the Tekton CLI to test our build Task you need to pass the ServiceAccount pipeline to be use to run the Task. You will need to pass the GitHub URL to your fork or use this repository. You will need to pass the directory within the repository where the application in our case is nodejs. The repository image name is knative-tekton
```
tkn task start build --showlog \
  -p repo-url=${GIT_REPO_URL} \
  -p image=${REGISTRY_SERVER}/${REGISTRY_NAMESPACE}/knative-tekton \
  -p CONTEXT=nodejs \
  -s pipeline 
```

You can check out the container registry and see that the image was pushed to repository a minute ago, it should return status Code 200

curl -s -o /dev/null -w "%{http_code}\n" https://index.$REGISTRY_SERVER/v1/repositories/$REGISTRY_NAMESPACE/knative-tekton/tags/latest

5.3 The Deploy Tekton Task

I provided a Deploy Tekton Task that can run kubectl to deploy the Knative Application using a YAML manifest.

Show me the Deploy Task YAML

apiVersion: tekton.dev/v1beta1
kind: Task
metadata:
  name: deploy
spec:
  params:
    - name: repo-url
      description: The git repository url
    - name: revision
      description: The branch, tag, or git reference from the git repo-url location
      default: master
    - name: dir
      description: Path to the directory to use as context.
      default: .
    - name: yaml
      description: Path to the directory to use as context.
      default: ""
    - name: image
      description: Path to the container image
      default: ""
    - name: KUBECTL_IMAGE
      description: The location of the kubectl image.
      default: docker.io/csantanapr/kubectl
  steps:
    - name: git-clone
      image: alpine/git
      script: |
        git clone $(params.repo-url) /source
        cd /source
        git checkout $(params.revision)
      volumeMounts:
        - name: source
          mountPath: /source
    - name: kubectl-apply
      image: $(params.KUBECTL_IMAGE)
      workingdir: /source
      script: |

        if [ "$(params.image)" != "" ] && [ "$(params.yaml)" != "" ]; then
          yq w -i $(params.dir)/$(params.yaml) "spec.template.spec.containers[0].image" "$(params.image)"
          cat $(params.dir)/$(params.yaml)
        fi

        kubectl apply -f $(params.dir)/$(params.yaml)

      volumeMounts:
        - name: source
          mountPath: /source
  volumes:
    - name: source
      emptyDir: {}

Install the provided task deploy like this.
```
kubectl apply -f tekton/task-deploy.yaml
```
You can list the task that we just created using the tkn CLI
```
tkn task ls
```
We can also get more details about the deploy Task using tkn task describe
```
tkn task describe deploy
```

I provided a Task YAML that defines our Knative Application in knative/service.yaml

apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: demo
spec:
  template:
    spec:
      containers:
        - image: docker.io/csantanapr/knative-tekton
          imagePullPolicy: Always
          env:
            - name: TARGET
              value: Welcome to the Knative Meetup

Let's use the Tekton CLI to test our deploy Task you need to pass the ServiceAccount pipeline to be use to run the Task. You will need to pass the GitHub URL to your fork or use this repository. You will need to pass the directory within the repository where the application yaml manifest is located and the file name in our case is knative and service.yaml .
```
tkn task start deploy --showlog \
  -p image=${REGISTRY_SERVER}/${REGISTRY_NAMESPACE}/knative-tekton \
  -p repo-url=${GIT_REPO_URL} \
  -p dir=knative \
  -p yaml=service.yaml \
  -s pipeline 
```
You can check out that the Knative Application was deploy
```
kn service list demo
```

5.4 The Build and Deploy Pipeline

If we want to build the application image and then deploy the application, we can run the Tasks build and deploy by defining a Pipeline that contains the two Tasks, deploy the Pipeline build-deploy.

Show me the Pipeline YAML

apiVersion: tekton.dev/v1beta1
kind: Pipeline
metadata:
  name: build-deploy
spec:
  params:
    - name: repo-url
      default: https://github.com/csantanapr/knative-tekton
    - name: revision
      default: master
    - name: image
    - name: image-tag
      default: latest
    - name: CONTEXT
      default: nodejs
  tasks:
    - name: build
      taskRef:
        name: build
      params:
        - name: image
          value: $(params.image):$(params.image-tag)
        - name: repo-url
          value: $(params.repo-url)
        - name: revision
          value: $(params.revision)
        - name: CONTEXT
          value: $(params.CONTEXT)
    - name: deploy
      runAfter: [build]
      taskRef:
        name: deploy
      params:
        - name: image
          value: $(params.image):$(params.image-tag)
        - name: repo-url
          value: $(params.repo-url)
        - name: revision
          value: $(params.revision)
        - name: dir
          value: knative
        - name: yaml
          value: service.yaml

Install the Pipeline with this command

kubectl apply -f tekton/pipeline-build-deploy.yaml

You can list the pipeline that we just created using the tkn CLI
```
tkn pipeline ls
```
We can also get more details about the build-deploy Pipeline using tkn pipeline describe
```
tkn pipeline describe build-deploy
```
Let's use the Tekton CLI to test our build-deploy Pipeline you need to pass the ServiceAccount pipeline to be use to run the Tasks. You will need to pass the GitHub URL to your fork or use this repository. You will also pass the Image location where to push in the the registry and where Kubernetes should pull the image for the Knative Application. The directory and filename for the Kantive yaml are already specified in the Pipeline definition.
```
tkn pipeline start build-deploy --showlog \
  -p image=${REGISTRY_SERVER}/${REGISTRY_NAMESPACE}/knative-tekton \
  -p repo-url=${GIT_REPO_URL} \
  -s pipeline 
```
You can inpect the results and duration by describing the last PipelineRun
```
tkn pipelinerun describe --last
```
Check that the latest Knative Application revision is ready
```
kn service list demo
```

Run the Application using the url

curl http://demo.$SUB_DOMAIN

It shoudl print

Welcome to OSS NA 2020

6. Automate the Tekton Pipeline using Git Web Hooks Triggers

6. Automate the Tekton Pipeline using Git Web Hooks

6.1 Install Tekton Triggers

Install Tekton Triggers in namespace tekton-pipelines

kubectl apply -f https://github.com/tektoncd/triggers/releases/download/v0.6.1/release.yaml
kubectl wait deployment tekton-triggers-controller tekton-triggers-webhook --for=condition=Available -n tekton-pipelines

6.2 Create TriggerTemplate, TriggerBinding

When the Webhook invokes we want to start a Pipeline, we will a TriggerTemplate to use a specification on which Tekton resources should be created, in our case will be creating a new PipelineRun this will start a new Pipeline install.

Show me the TriggerTemplate YAML

apiVersion: triggers.tekton.dev/v1alpha1
kind: TriggerTemplate
metadata:
  name: build-deploy
spec:
  params:
    - name: gitrevision
      description: The git revision
      default: master
    - name: gitrepositoryurl
      description: The git repository url
    - name: gittruncatedsha
    - name: image
      default: REPLACE_IMAGE
  resourcetemplates:
    - apiVersion: tekton.dev/v1beta1
      kind: PipelineRun
      metadata:
        generateName: build-deploy-run-
      spec:
        serviceAccountName: pipeline
        pipelineRef:
          name: build-deploy
        params:
          - name: revision
            value: $(params.gitrevision)
          - name: repo-url
            value: $(params.gitrepositoryurl)
          - name: image-tag
            value: $(params.gittruncatedsha)
          - name: image
            value: $(params.image)

Install the TriggerTemplate

cat tekton/trigger-template.yaml | sed "s/REPLACE_IMAGE/$REGISTRY_SERVER\/$REGISTRY_NAMESPACE\/knative-tekton/g" | kubectl apply -f -

When the Webhook invokes we want to extract information from the Web Hook http request sent by the Git Server, we will use a TriggerBinding this information is what gets passed to the TriggerTemplate.

Show me the TriggerBinding YAML

apiVersion: triggers.tekton.dev/v1alpha1
kind: TriggerBinding
metadata:
  name: build-deploy
spec:
  params:
    - name: gitrevision
      value: $(body.head_commit.id)
    - name: gitrepositoryurl
      value: $(body.repository.url)
    - name: gittruncatedsha
      value: $(body.extensions.truncated_sha)

Install the TriggerBinding

kubectl apply -f tekton/trigger-binding.yaml

6.3 Create Trigger EventListener

To be able to handle the http request sent by the GitHub Webhook, we need a webserver. Tekton provides a way to define this listeners that takes the TriggerBinding and the TriggerTemplate as specification. We can specify Interceptors to handle any customization for example I only want to start a new Pipeline only when push happens on the main branch.

Show me the Trigger Eventlistener YAML

apiVersion: triggers.tekton.dev/v1alpha1
kind: EventListener
metadata:
  name: cicd
spec:
  serviceAccountName: pipeline
  triggers:
    - name: cicd-trig
      bindings:
        - ref: build-deploy
      template:
        name: build-deploy
      interceptors:
        - cel:
            filter: "header.match('X-GitHub-Event', 'push') && body.ref == 'refs/heads/master'"
            overlays:
              - key: extensions.truncated_sha
                expression: "body.head_commit.id.truncate(7)"

Install the Trigger EventListener

kubectl apply -f tekton/trigger-listener.yaml

The Eventlister creates a deployment and a service you can list both using this command
```
kubectl get deployments,eventlistener,svc -l eventlistener=cicd
```

6.4 Get URL for Git WebHook

If you are using the IBM Free Kubernetes cluster a public IP Address is alocated to your worker node and we will use this one for this part of the tutorial. It will depend on your cluster and how traffic is configured into your Kubernetes Cluster, you would need to configure an Application Load Balancer (ALB), Ingress, or in case of OpenShift a Route. If you are running the Kubernetes cluster on your local workstation using something like minikube, kind, docker-desktop, or k3s then I recommend a Cloud Native Tunnel solution like inlets by the open source contributor Alex Ellis.

Expose the EventListener with Kourier

cat <<EOF | kubectl apply -f -
apiVersion: networking.internal.knative.dev/v1alpha1
kind: Ingress
metadata:
  name: el-cicd
  namespace: $CURRENT_NS
  annotations:
    networking.knative.dev/ingress.class: kourier.ingress.networking.knative.dev
spec:
  rules:
  - hosts:
    -  el-cicd.$CURRENT_NS.$KNATIVE_DOMAIN
    http:
      paths:
      - splits:
        - appendHeaders: {}
          serviceName: el-cicd
          serviceNamespace: $CURRENT_NS
          servicePort: 8080
    visibility: ExternalIP
  visibility: ExternalIP
EOF

Get the url using using CURRENT_NS and KNATIVE_DOMAIN
```
GIT_WEBHOOK_URL=http://el-cicd.$CURRENT_NS.$KNATIVE_DOMAIN
echo GIT_WEBHOOK_URL=$GIT_WEBHOOK_URL
```
WARNING: Take into account that this URL is insecure is using http and not https, this means you should not use this type of URL for real work environments, In that case you would need to expose the service for the eventlistener using a secure connection using https
Add the Git Web Hook url to your Git repository
1. Open Settings in your Github repository
2. Click on the side menu Webhooks
3. Click on the top right Add webhook
4. Copy and paste the $GIT_WEBHOOK_URL value into the Payload URL
5. Select from the drop down Content type application/json
6. Select Send me everything to handle all types of git events.
7. Click Add webhook
Now make a change to the application manifest such like changing the message in knative/service.yaml to something like My First Serveless App @ OSS NA 2020 🎉 🌮 🔥 🤗! and push the change to the master branch
(Optional) If you can't receive the git webhook, for example if using minikube you can emulate the git web hook using by sending a http request directly with git payload. You can edit the file tekton/hook.json to use a different git commit value.
```
curl -H "X-GitHub-Event:push" -d @tekton/hook.json $GIT_WEBHOOK_URL
```
A new Tekton PipelineRun gets created starting a new Pipeline Instance. You can check in the Tekton Dashboard for progress of use the tkn CLI
```
tkn pipeline logs -f --last
```
To see the details of the execution of the PipelineRun use the tkn CLI
```
tkn pipelinerun describe --last
```
The Knative Application Application is updated with the new Image built using the tag value of the 7 first characters of the git commit sha, describe the service using the kn CLI
```
kn service describe demo
```

Invoke your new built revision for the Knative Application

curl http://demo.$SUB_DOMAIN

It should print

My First Serveless App @ OSS NA 2020  🎉 🌮 🔥 🤗!

rugwirobaker / knative-tekton

Tutorial: Knative + Tekton

Linux Foundation Open Source Summit NA 2020

1. Setup Environment

2. Install Knative Serving

3. Using Knative to Run Serverless Applications

3.1 Create Knative Service

3.2 Updating the Knative service

3.3 Knative Service Traffic Splitting

4. Install Tekton

4.2 Install Tekton Dashboard (Optional)

5. Using Tekton to Build Applications

5.1 Configure Access for Tekton

5.2 The Build Tekton Task

5.3 The Deploy Tekton Task

5.4 The Build and Deploy Pipeline

6. Automate the Tekton Pipeline using Git Web Hooks

6.1 Install Tekton Triggers

6.2 Create TriggerTemplate, TriggerBinding

6.3 Create Trigger EventListener

6.4 Get URL for Git WebHook

About

Languages