Unable to access Nginx Ingress Load Balancer IP from my local

Question

I have created an on-premise kubernetes cluster using kubespray. The cluster consists of 3 master and 3 worker nodes. I have followed the documentation from this website : https://www.linuxtechi.com/install-kubernetes-using-kubespray/

Then I have configured MetalLB and nginx-ingress controller using this github repo : https://github.com/morrismusumi/kubernetes/blob/main/clusters/homelab-k8s/apps/metallb-plus-nginx-ingress/README.md

Right now all my nodes are available and they are working fine but when i deploy the example webapp in the link https://github.com/morrismusumi/kubernetes/blob/main/clusters/homelab-k8s/apps/metallb-plus-nginx-ingress/web-app-ingress.yml i cannot reach it from my local machine unless i make the "type: LoadBalancer" in my web-app-deployment.yaml

My cluster ips are as follows

• M01 - 11.11.120.50
• M02 - 11.11.120.51
• M03 - 11.11.120.52
• W01 - 11.11.120.53
• W02 - 11.11.120.54
• W03 - 11.11.120.55

and for my metal-lb address pool i used 11.11.120.150-11.11.120.170

When i deploy an example nginx web app with "LoadBalancer" type it can assign an External IP and when i curl to this external IP i can successfully see the output "welcome to my first website" from my local. when i change the service type from LoadBalancer to Cluster IP and create ingress yaml, i can see the External IP address in "kubectl get ingress" output. But in this scenario, when i try to curl from my local machine i can not reach to this address. if i try from cluster nodes, i can access.

Manifests files for example web-app are as follows:

web-deployment.yaml:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: web-app
  labels:
    app.kubernetes.io/name: web-app
  name: web-app
spec:
  replicas: 1
  selector:
    matchLabels:
      app.kubernetes.io/name: web-app
  template:
    metadata:
      labels:
        app.kubernetes.io/name: web-app
    spec:
      containers:
      - image: nginx
        name: web-app
        command: 
          - /bin/sh
          - -c
          - "echo 'welcome to my web app!' > /usr/share/nginx/html/index.html && nginx -g 'daemon off;'"
      dnsConfig:
            options:
              - name: ndots
                value: "2"

---

apiVersion: v1
kind: Service
metadata:
  name: web-app
  labels:
    app.kubernetes.io/name: web-app
spec:
  selector:
    app.kubernetes.io/name: web-app
  ports:
  - name: http
    port: 80
    protocol: TCP
    targetPort: 80
  type: ClusterIP

web-app-ingress.yaml:

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: webapp
  annotations:
    kubernetes.io/ingress.class: "nginx"
spec:
  rules:
  - host: example-webapp.com
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: web-app
            port:
              number: 80

and i added hostname and IP to my /etc/hosts file like "11.11.120.151 example-webapp.com"

my ingress output:

kubectl get ingress
NAME      CLASS    HOSTS                ADDRESS         PORTS   AGE
web-app   <none>   example-webapp.com   11.11.120.151   80      63m

my service output:

kubectl get services
NAME         TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)   AGE
kubernetes   ClusterIP   10.233.x.x      <none>        443/TCP   42h
web-app      ClusterIP   10.233.x.x      <none>        80/TCP    70m

By the way, i changed my web-app service externalTrafficPolicy from Cluster to Local. But nothing has changed.

The only way to reach it from my local machine is making the "type: LoadBalancer" as i started before. While i can access via LoadBalancer, i cannot access with Ingress. Can you help me to solve this problem?

Thanks in advance.

Answer 1

I solved the issue;

Instead of the image in the official nginx installation document, I used the image on Kubernetes' own site: https://kubernetes.github.io/ingress-nginx

(changed the nginx helm repo from oci://ghcr.io/nginxinc/charts/nginx-ingress to https://kubernetes.github.io/ingress-nginx)

When I followed this doc my problem was fixed. In fact, the difference was that it created only one controller and a default-backend deployment. When I applied it as in the sample ingress yaml (writing ingressClassName under spec instead of annotation), my problem was solved and I was able to access ingress resources successfully.