README.adoc

ImageBuilder - Knative build

This example demonstrates how to build images with Knative build, starting from source stored in a GitHub repository.

Two variations are offered, one with a plain Knative Build, the other using a BuildTemplate. First, you need to have Knative installed. Please refer to the installation instructions for the options you have for installing Knative build

Both examples require a registry to push the created image into. For the simplicities sake, we install an insecure and unprotected registry directly into the current namespace.

kubectl create -f registry.yml

This registry is only available within from the current namespace as registry. Of course, feel free to adapt the examples below to use another registry which is more accessible than that.

Simple Build

To start a simple build with the Jib Maven Plugin you create the build with

kubectl create -f build-with-jib.yml

As always, you might want to have a look into the descriptor, it contains some useful comments.

To monitor the build check the pods:

kubectl get pods

NAME                                    READY   STATUS     RESTARTS   AGE
random-generator-with-jib-pod-b68481   0/1     Init:1/3   0          7s
registry-59df4ddcdc-jhkdg               1/1     Running    0          42s

and the logs

kubectl logs random-generator-with-jib-pod-b68481 -c build-step-build-and-push

[INFO] Scanning for projects...
Downloading: https://repo.maven.apache.org/maven2/....
....
[INFO]
[INFO] Containerizing application to registry:80/k8spatterns/random-generator...
[WARNING] Base image 'gcr.io/distroless/java' does not use a specific image digest - build may not be reproducible
[INFO] Retrieving registry credentials for registry:80...
[INFO] Getting base image gcr.io/distroless/java...
[INFO] Building dependencies layer...
[INFO] Building resources layer...
[INFO] Building classes layer...
[INFO]
[INFO] Container entrypoint set to [java, -cp, /app/resources:/app/classes:/app/libs/*, io.k8spatterns.examples.RandomGeneratorApplication]
[INFO]
[INFO] Built and pushed image as registry:80/k8spatterns/random-generator
[INFO]
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 24.203 s
[INFO] Finished at: 2019-01-22T11:24:16Z
[INFO] Final Memory: 39M/403M
[INFO] ------------------------------------------------------------------------

In order to deploy see this application, see below.

Build Template

As alternative you can build the application with a BuildTemplate.

First, install the template:

kubectl create -f maven-kaniko-buildtemplate.yml

You can verify the template with

kubectl get buildtemplate
NAME           AGE
maven-kaniko   7s

Now start the build which uses this template:

kubectl create -f build-with-template.yml

You can watch the progress of the build with

kubectl get pods
NAME                                    READY   STATUS      RESTARTS   AGE
random-generator-with-template-pod-0ae221    0/1     Completed   0          1m

kubectl logs random-generator-with-template-pod-0ae221 -c build-step-maven-build
.....
kubectl logs random-generator-with-template-pod-0ae221 -c build-step-image-build-and-push
.....

Deploy application

When you have build the random-generator with one of the methods above, you can deploy it easily. Unfortunately, for picking up images, Kubernetes (or at least Minikube) doesn’t support service DNS lookups in the image: field of a container specification.

So we need to insert the IP adress directly which you can do while deploying, like in:

cat random-generator-deploy.yml | \
  sed -e "s/registry:80/$(kubectl get svc registry -o jsonpath={.spec.clusterIP}):80/" | \
  kubectl create -f -

This will also install an Ingress object which allows you access the service easily from outside the cluster. If you are testing with Minikube, you should be sure that you have the Ingress addon enabled with minikube addons enable ingress before deploying the application.

On Minikube you then access the service with

curl -sL http://random-service.$(minikube ip).nip.io | jq .