How to Run PostgreSQL on my Local Machine.

I am currently making a project helping to memorize english vocabularies.

At first success, I set up PostgreSQL by using Deployment, Service, Secret. Then I realized this is not a best practice. Because it was stateless, which means all tables disappear if a pod is killed. So this is suitable for stateless service like a API server. As I will plan to deploy the API service, I will make some notes.

Structures: Deployment, Service, Secret

This is my setup details.

Deployment

apiVersion: apps/v1
kind: Deployment
metadata:
  name: postgres-deployment
spec:
  replicas: 1
  selector:
    matchLabels:
      app: postgres
  template:
    metadata:
      labels:
        app: postgres
    spec:
      containers:
      - name: postgres
        image: postgres:13
        ports:
          - containerPort: 5432
        envFrom:
          - secretRef:
              name: postgres-secret
        volumeMounts:
          - name: postgres-storage
            mountPath: /var/lib/postgresql/data
        resources:
          requests:
            memory: "512Mi"
            cpu: "500m"
          limits:
            memory: "1Gi"
            cpu: "1"
      volumes:
        - name: postgres-storage
          emptyDir: {}

Secret

Service

apiVersion: v1
kind: Service
metadata:
  name: postgres-service
spec:
  selector:
    app: postgres
  ports:
    - protocol: TCP
      port: 5432
      targetPort: 5432
  type: NodePort

Run

kubectl apply -f postgre-secret.yaml
kubectl apply -f postgre-deployment.yaml
kubectl apply -f postgre-service.yaml

Access

kubectl get svc 

In the result, you will see the external port like down below. In my case, it was 31477. But if you don't specify NodePort in your service yaml, it will randomly set. Thus yours might be different with mine.

Then, In your dbms program, you can access to the database.

---

However, after some talks with ChatGPT, I learned it should be a stateful service. So, I will try to change this in a Stateful method.

Wow, it was stateless because I didn't set PersistenceVolume. Changing this configuration, it doesn't clear all tables when the pod is killed.

Deployment yaml

...
spec:
  template:
    spec:
      volumes:
        - name: postgres-storage
          persistentVolumeClaim:
            claimName: postgres-pvc
...

PV

apiVersion: v1
kind: PersistentVolume
metadata:
  name: postgres-pv
spec:
  capacity:
    storage: 1Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: /tmp/postgres-data
    type: Directory

PVC

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: postgres-pvc
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi

kubectl apply -f postgre-pv.yaml
kubectl apply -f postgre-pvc.yaml
kubectl apply -f postgre-deployment.yml 

But, this strategy has a disadvantage. I don't know exactly, there might be something related to count of the pod. I think as this deployment runs only one pod, it can be stateful? might be...?

---

Kubenetes kinds usages recommended by ChatGPT

Use Case	Recommended Resource
Stateless apps (e.g., APIs, frontends)	Deployment
Apps needing shared storage (e.g., logs, caches)	Deployment + PVC
Databases, Kafka, Redis, Zookeeper	StatefulSet + PVC
Apps that require stable network identities	StatefulSet