Kubernetes internal network and addresses not working for all internal names for same pod/service

Question

I am setting up a k8s cluster using k3s. Currently I have a Jenkins pod running, set up according to this guide.

I have run into an issue setting up the kubernetes cluster for Jenkins workers I can't seem to get my head wrapped around:

If I go to my Jenkins pod and start the console to test that the JNLP is working, not all the different names I can use work:

curl localhost:50000 ✔ Works
curl `cat /etc/hostname`:50000 ✔ Works
curl jenkins-service:50000 ❌ Times out

BUT !!!

curl localhost:8080 ✔ Works
curl `cat /etc/hostname`:8080 ✔ Works
curl jenkins-service:8080 ✔ Works

I then went back to the YAML and took a look at it. Appart from there being more references to 8080 than 50000 (8080 is a part of several urls and liveliness checks and so on), I do not see anything amiss

      containers:
        ...
        ports:
        - containerPort: 8080
          name: httpport
          protocol: TCP
        - containerPort: 50000
          name: jnlpport
          protocol: TCP

I have looked at these defects and the issues seem to not be the same, causing me to think that it must be Kubernetes related.

Answer 1

I found a detailed guide which lead me to the answer...

In the service section only port 8080 was forwarded to 32000.

So I changed the service.yaml to be

apiVersion: v1
kind: Service
metadata:
  name: jenkins-service
  namespace: devops-tools
  annotations:
      prometheus.io/scrape: 'true'
      prometheus.io/path:   /
      prometheus.io/port:   '8080'
spec:
  selector:
    app: jenkins-server
  type: NodePort
  ports:
    - name: httpport
      port: 8080
      targetPort: 8080
      nodePort: 32000
    - name: jnlpport
      port: 50000
      targetPort: 50000

and executed kubectl apply -f service.yaml which seems to have had an immediate effect.