Yugastore in Kubernetes

This is an implementation of a sample ecommerce app. This microservices-based retail marketplace or eCommerce app is composed of microservices written in Java (Spring), a GUI based on React and YugabyteDB as the distributed SQL database.

If you're using this demo app, please ⭐ this repository! We appreciate your support.

Trying it out

This repo contains all the instructions you need to run the app on your laptop.

You can also try the app out online, it is hosted on Pivotal Web Services.

Versions

• Java 17
• Spring Boot 2.6.3
• Spring Cloud 2021.0.0
• Yugabyte Java Driver 4.6.0-yb-10
• Python 3 (for data Loading)

Features

• Written fully in Spring Framework
• Desgined for multi-region and Kubernetes-native deployments
• Features Spring Boot microservices
• Uses a discovery service for the microservices
• Sample data has over 6K products in the store

Architecture

The architecture diagram of Yugastore is shown below.

Minikube Setup

This mode puts Docker images directly into minikube's container repository using skaffold for automation.

1. Start minikube.

   minikube start --cpus 4 --memory 5120 --vm-driver virtualbox
   

2. Tell skaffold to use minikube's docker container registry. It's magic.

   eval $(minikube -p minikube docker-env)
   

   If you previously used minikube and now want to switch to GKE, unset any minikube Docker environment variables, if present:

   eval $(minikube -p minikube docker-env -u)
   

GKE and Minikube

Start here for GKE. Continue here if you are using minikube.

1. Ensure your helm charts are up to date so you can run YugabyteDB in k8s.

   helm repo update
   

2. Create OSS YugabyteDB, and give it a minute or two to start up. Note: if you use Platform instead, you will need to edit env.yaml in the k8s-manifests directory.

   for namespace in yb-demo
   do
   helm install $namespace yugabytedb/yugabyte \
   --set resource.master.requests.cpu=0.5,resource.master.requests.memory=0.5Gi,\
   resource.tserver.requests.cpu=0.5,resource.tserver.requests.memory=0.5Gi,\
   replicas.master=1,replicas.tserver=1,enableLoadBalancer=False,\
   istioCompatibility.enabled=true \
   --create-namespace --namespace $namespace
   done
   

3. Enable port forwarding to the database so you can populate data.

   kubectl port-forward -n yb-demo svc/yb-tservers 5444:5433 &
   

4. Create the schemas and load product catalog data into YugabyteDB.

   cd resources
   ./dataload.sh 5444
   

5. Build the Docker containers and deploy them.

   Minikube:

   skaffold run --skip-tests=true
   

   GKE:

   skaffold run --skip-tests=true --default-repo gcr.io/dataengineeringdemos/yugabyte
   

   • You will need to tell skaffold the name of your GCP project's image registry in the step above. The registry name can be determined by navigating to your project's Container Registry and clicking on the copy icon next to the repository to get the full name.

6. Populate the Redis cache.

   kubectl port-forward redis 6380:6379 &
   ./json2redis.py products.json
   

7. Enable port forwarding for your browser, and browse to http://localhost:8080/

   kubectl port-forward svc/yugastore-ui 8080:8080