Nginx routing for kubernetes services

Question

I am having same requirements like metioned in this post - https://tech.holidayextras.com/routing-to-internal-kubernetes-services-using-proxies-and-ingress-controllers-e7eb44954d53

But I am not sure of what is that ipaddresses of backends mentioned there. I assumed it is the ips of my master and worker nodes in my cluster

My kubernetes has master node ip as 10.118.6.35 and worker node ip as 10.118.2.215 which are AWS ec2 instances.

So when I configured like below in my nginx.conf(please refer below), I am getting index.html rendered when I do
curl https://10.118.6.35
As well from my browser it is working fine.

But when I do https://10.118.6.35/hello-kenzan
I am getting nginx error page default 404 page
But I expect to route it to my kubernetes services running in NodePort 80:30854

I have followed steps as mentioned in that post.
FYI, I am showing the echo-ing.yaml below -

    apiVersion: extensions/v1beta1
    kind: Ingress
    metadata:
      name: echomap
      annotations: {
        'kubernetes.io/ingress.class': nginx
       }
    spec:
      rules:
      - host: ip-10-118-6-35.ec2.internal
        http:
          paths:
          - path: /hello-kenzan
            backend:
               service and: hello-kenzan
               service port: 80

And below is nginx.conf -

    http {
           ssl_certificate ...
            ....
            ....
      server {
       listen     443 SSL;
       server_name www.ip-10.118-6-35.ec2.internal.com;
       root /usr/share/nginx/html;
       ssl_certificate ...
       ...
       ...
       include /etc/nginx/default.d/*.conf;
    
       location / {
       }
       error_page 404 /404.html;
          location = /40x.html {
        }
        error_page 500 502 503 504 /50x.html;
          location = /50x.html {
        }
    }
    upstream backend_nodes {
        server 10.118.6.35:31001;
        server 10.118.2.215:31001;
      }
    upstream backend_nodes_ssl {
        server 10.118.6.35:32001;
        server 10.118.2.215:32001;
      }
    server {
        listen 80;
        server_name backends.nodes;
         location / {
           proxy_pass https://backend_nodes;
          }
       }
     server {
        listen 443;
        server_name backends.nodes.ssl;
         location / {
           proxy_pass https://backend_nodes_ssl;
          }
       }
    }

Please suggest what must go in the backend nodes ip.

Answer 1

Hey, backend is a service running on your cluster and when asked about the IP address of the backend, it refers to the IP address of the service. To get this IP addres execute the following command:

kubectl get svc

It'll list down all the services with their clusterIP and external IP's. If used the service NodePort, the service will not have any external IP. In such a case use the clusterIP to access the service. If you've used LoadBalancer as your service type then you can access the service using the ExternalIP. 

Comments

Thanks for the reply..

**Update**

Hey finally made it to work when I try with IP address from browser instead of DNS name -

https://10.118.6.35/hello-kenzan  <- this working now
https://myservice.myorg.com/hello-kenzan <- this not working by giving nginx error page temporarily unavailable.

But https://myservice.myorg.com <- works by taking to welcome to nginx index page.

This is my modified nginx.conf -

````
 
 http {
           ssl_certificate ...
            ....
            ....

server {
       listen     80;
       server_name ip-10.118-6-35.ec2.internal;
       root /usr/share/nginx/html;
       
       include /etc/nginx/default.d/*.conf;
    
       location / {
       }
       location /hello-kenzan {
           proxy_pass https://backend_nodes;
          }
       error_page 404 /404.html;
          location = /40x.html {
        }
        error_page 500 502 503 504 /50x.html;
          location = /50x.html {
        }
    }
      server {
       listen     443 SSL;
       server_name ip-10.118-6-35.ec2.internal;
       root /usr/share/nginx/html;
       ssl_certificate ...
       ...
       ...
       include /etc/nginx/default.d/*.conf;
    
       location / {
       }
       location /hello-kenzan {
           proxy_pass https://backend_nodes_ssl;
       }
       error_page 404 /404.html;
          location = /40x.html {
        }
        error_page 500 502 503 504 /50x.html;
          location = /50x.html {
        }
    }
    upstream backend_nodes {
        server 10.96.88.237:80
      }
    upstream backend_nodes_ssl {
        server 10.96.88.237:443
      }
   
  }
  
 ````

Any thoughts??

---

Ahh, that looks good. Exposing the service on a particular port made it work.

---

Yeah thanks again now DNS also is working!