Why I can't access my web service from the kubernetes cluster? - docker

Question

I'm trying to execute an application inside a kubernetes cluster.
I used to launch the application with docker-compose without problems, but when I create
my kubernetes deployment files, I am not able to access the service inside the cluster even after exposing them. here is my deployment file:
apiVersion: apps/v1
kind: Deployment
metadata:
name: myapp
spec:
replicas: 3
selector:
matchLabels:
app: myapp
# type: LoadBalancer
template:
metadata:
labels:
app: myapp
spec:
containers:
- name: myapp
image: jksun12/vdsaipro
# command: ["/run.sh"]
ports:
- containerPort: 80
- containerPort: 3306
# volumeMounts:
# - name: myapp-pv-claim
# mountPath: /var/lib/mysql
# volumes:
# - name: myapp-pv-claim
# persistentVolumeClaim:
# claimName: myapp-pv-claim
---
apiVersion: apps/v1
kind: PersistentVolumeClaim
metadata:
name: myapp-pv-claim
labels:
app: myapp
spec:
accesModes:
- ReadWriteMany
resources:
requests:
storage: 4Gi
Here is the result of
kubectl describe service myapp-service
:
Name: myapp-service
Namespace: default
Labels: <none>
Annotations: <none>
Selector: app=myapp
Type: NodePort
IP: 10.109.12.113
Port: port-1 80/TCP
TargetPort: 80/TCP
NodePort: port-1 31892/TCP
Endpoints: 172.18.0.5:80,172.18.0.8:80,172.18.0.9:80
Port: port-2 3306/TCP
TargetPort: 3306/TCP
NodePort: port-2 32393/TCP
Endpoints: 172.18.0.5:3306,172.18.0.8:3306,172.18.0.9:3306
Session Affinity: None
External Traffic Policy: Cluster
Events: <none>
And here are the errors that I get when I try to access them:
curl 172.17.0.2:32393
curl: (1) Received HTTP/0.9 when not allowed
And here is the next result when I try to access the other port
curl 172.17.0.2:31892
curl: (7) Failed to connect to 172.17.0.2 port 31892: Connection refused
curl: (7) Failed to connect to 172.17.0.2 port 31892: Connection refused
I'm running ubuntu server 20.04.1 LTS. The manip is on top of minikube.
Thanks for your help.

Answer 1

If you are accessing the service from inside the cluster use ClusterIP as the IP. So curl should be 10.109.12.113:80 and 10.109.12.113:3306
In case accessing it from outside the cluster then use NODEIP and NODEPORT. So curl should be on <NODEIP>:32393 and <NODEIP>:31892
From inside the cluster I would also use POD IPs directly to understand if the issue is at service level or pod level.
You need to make sure that the application is listening on port 80 and port 3306. Only mentioning containerPort as 80 and 3306 does not make the application listen on those ports.
Also make sure that the application code inside the pod is listening on 0.0.0.0 instead of 127.0.0.1