Skip to content

Latest commit

 

History

History
106 lines (60 loc) · 4.54 KB

README.md

File metadata and controls

106 lines (60 loc) · 4.54 KB

Strimzi Kafka Image with SNAPSHOT Strimzi Kafka OAuth

This is a build of a Docker image based on quay.io/strimzi/kafka:0.44.0-kafka-3.8.0 with added most recently locally built SNAPSHOT version of Strimzi Kafka OAuth libraries.

This image adds a /opt/kafka/libs/oauth directory, and copies the latest jars for OAuth support in it. Then it puts this directory as the first directory on the classpath.

The result is that the specific version of Strimzi Kafka OAuth jars and their dependencies is used, because they appear on the classpath before the ones that are part of quay.io/strimzi/kafka:0.44.0-kafka-3.8.0 which are located in the /opt/kafka/libs directory.

Building

Run mvn install then, use docker build to build the image:

docker build --output type=docker -t strimzi/kafka:latest-kafka-3.8.0-oauth .

You can choose a different tag if you want.

Also, take a look at Dockerfile:

less Dockerfile

Note the FROM directive in the first line. It uses image coordinates to a recent publicly available Strimzi Kafka 3.8.0 image.

You may want to adjust this to a different public image, or to one manually built previously and is only available in your private Docker Registry.

For example, if you want to base your image on Strimzi Kafka 3.7.0 use FROM quay.io/strimzi/kafka:0.43.0-kafka-3.7.0.

Validating

You can start an interactive shell container and confirm that the jars are there.

docker run --rm -ti strimzi/kafka:latest-kafka-3.8.0-oauth /bin/sh
ls -la libs/oauth/
echo "$CLASSPATH"

If you want to play around more within the container you may need to make yourself root.

You achieve that by running the docker session as root user:

docker run --rm -ti --user root strimzi/kafka:latest-kafka-3.8.0-oauth /bin/sh

Pushing the image to a Docker Repository

For Kubernetes to be able to use our image it needs to be pushed to either a public repository or to the private Docker Repository used by your Kubernetes distro.

For example if you are using Kubernetes Kind as described in HACKING.md then your Docker Repository is listening on port 5000 of your local ethernet IP.

# On MacOS
export REGISTRY_IP=$(ifconfig en0 | grep 'inet ' | awk '{print $2}') && echo $REGISTRY_IP 

# On Linux
#export REGISTRY_IP=$(ifconfig docker0 | grep 'inet ' | awk '{print $2}') && echo $REGISTRY_IP 

export DOCKER_REG=$REGISTRY_IP:5000

You need to retag the built image, so you can push it to Docker Registry:

docker tag strimzi/kafka:latest-kafka-3.8.0-oauth $DOCKER_REG/strimzi/kafka:latest-kafka-3.8.0-oauth
docker push $DOCKER_REG/strimzi/kafka:latest-kafka-3.8.0-oauth

Actually, Kubernetes Kind supports an even simpler option how to make an image available to Kubernetes:

kind load docker-image strimzi/kafka:latest-kafka-3.8.0-oauth

If you're using minikube, you'll need to run minikube docker-env before building the image.

Deploying

Via the Strimzi Repository

In order for the operator to use your Kafka image, you have to replace the Kafka image coordinates in packaging/install/cluster-operator/060-Deployment-strimzi-cluster-operator.yaml in your strimzi-kafka-operator project.

This image builds the kafka-3.8.0 replacement image, so we need to replace all occurrences where kafka-3.8.0 is referred to into the proper coordinates to our image:

sed -Ei "s#quay.io/strimzi/kafka:latest-kafka-3.8.0#${DOCKER_REG}/strimzi/kafka:latest-kafka-3.8.0-oauth#" \
    packaging/install/cluster-operator/060-Deployment-strimzi-cluster-operator.yaml

It's best to check the 060-Deployment-strimzi-cluster-operator.yaml file manually to make sure everything is in order:

less packaging/install/cluster-operator/060-Deployment-strimzi-cluster-operator.yaml

You can now deploy Strimzi Kafka Operator following instructions in HACKING.md

Via Helm

You can also run the operator via its Helm chart and set the kafka.image.registry property to your local registry. As an example, if you've built and tagged the image as local.dev/strimzi/kafka:0.44.0-kafka-3.8.0. You can run it using the operator as:

helm repo add strimzi https://strimzi.io/charts/ --force-update
helm upgrade -i -n strimzi strimzi strimzi/strimzi-kafka-operator \
  --version 0.44.0 \
  --set watchNamespaces="{default}" \
  --set generateNetworkPolicy=false \
  --set kafka.image.registry="local.dev" \
  --wait \
  --create-namespace