Filter-Event Service

This project contains a Quarkus REST service for the filter-events of my osmosis water-filter. The micro-controller of the water-filter will send all flushing and filtering events as HTTP REST calls to the service. The events are stored in a PostgreSQL database.

Both the REST service and the PostgreSQL database are hosted on a k3d Kubernetes cluster which runs on my Raspberry Pi4. The k3s cluster comes with its own container registry.

Used Quarkus Extensions

• RestEasy Reactive
• Hibernate Reactive with Panache
• Reactive Postgres Client
• Docker Container Image

Build & Deploy to raspberry-pi

# Java Build, Docker Build & Push to k3d registry
mvn package

# Deploy PostgreSQL database on Kubernetes
kubectl apply -f k8s-database.yaml

# Deploy Quarkus service on Kubernetes
kubectl apply -f k8s-service.yaml

Test

Send the HTTP requests located in the file src/test/resources/requests.http using VSCode extension REST Client.

k3d setup

1. Software Prerequisites

   • Ubuntu 20.04 LTS
   • Docker https://omar2cloud.github.io/rasp/rpidock/
   • Patch the /boot/firmware/cmdline.txt and add cgroup_enable=cpuset cgroup_enable=memory cgroup_memory=1 to the end of the file
   • k3d https://k3d.io/
   • kubectl - https://kubernetes.io/docs/tasks/tools/install-kubectl-linux/

2. Setup a new k3d cluster on Raspberry Pi 4

   # set some variables
   raspberry_ip=192.168.178.41
   cluster_name=pi4-cluster
   registry_name=pi4-registry
   
   # create the directory on the host where we will persist the data
   mkdir -p /tmp/k3dvol
   
   # create a new k3d cluster
   k3d cluster create $cluster_name --volume /tmp/k3dvol:/tmp/k3dvol --api-port $raspberry_ip:6550 -p "8081:80@loadbalancer" --agents 2 --registry-create $registry_name:0.0.0.0:5000
   
   # get the kubeconfig - copy to the local machine
   k3d kubeconfig get -a

3. Create DNS entries for the local machine

   Domain	IP address
   pi4-registry	192.168.178.41
   ubuntu-pi4	192.168.178.41

4. Set docker insecure registry on the local machine

   "insecure-registries" : ["pi4-registry:5000"]

Further Links

• Great k3d tutorial: https://ian-says.com/articles/k3d-k8s-kubernetes/

TODOs

• Mount Raspberry Pi volume into k3d to keep DB persistence on cluster restart