Unable to curl pod IP using containerPort - kubernetes

Question

Sorry if it's a naive question. Please correct me if my understanding is wrong.
Created POD using this command:
kubectl run nginx --image=nginx --port=8888
My understand of this command, nginx (application) container will be exposed/available at port 8888
kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx 1/1 Running 0 10m 10.244.1.2 node01 <none> <none>
curl -v 10.244.1.2:8888 ===> i am wondering why this failed ?
* Trying 10.244.1.2:8888...
* TCP_NODELAY set
* connect to 10.244.1.2 port 8888 failed: Connection refused
* Failed to connect to 10.244.1.2 port 8888: Connection refused
* Closing connection 0
curl: (7) Failed to connect to 10.244.1.2 port 8888: Connection refused
curl -v 10.244.1.2 ===> to my surprise this returned 200 success response
* Trying 10.244.1.2:80...
* TCP_NODELAY set
* Connected to 10.244.1.2 (10.244.1.2) port 80 (#0)
> GET / HTTP/1.1
> Host: 10.244.1.2
> User-Agent: curl/7.68.0
> Accept: */*
>
* Mark bundle as not supporting multiuse
< HTTP/1.1 200 OK
If the application is still referring default 80 port, I am wondering about the significance of container port 8888 ?
OK, so it may be used to expose the POD to the outside world.
Let's see that, I went ahead and created service for the POD:
kubectl expose pod nginx --port=80 --target-port=8888
kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
nginx ClusterIP 10.96.214.161 <none> 80/TCP 13m
$ curl -v 10.96.214.161 ==> here default port (80) didn't work
* Trying 10.96.214.161:80...
* TCP_NODELAY set
* connect to 10.96.214.161 port 80 failed: Connection refused
* Failed to connect to 10.96.214.161 port 80: Connection refused
* Closing connection 0
curl: (7) Failed to connect to 10.96.214.161 port 80: Connection refused
$ curl -v 10.96.214.161:8888 ==> target port didn't work either
* Trying 10.96.214.161:8888...
* TCP_NODELAY set
....waiting forever
Which port do I need to use to make it work? Am I missing anything?

Answer 1

By default, nginx server listen to the port 80. You can see it in their docker image ref.
With kubectl run nginx --image=nginx --port=8888 what you have done here is you have expose another port along with 80. But the server is still listening on the 80 port.
So, try with target port 80. For this reason when you tried with other than port 80 it's not working. Try with set --target-port=8888 to --target-port=80.
Or, If you want to change the server port you need to use configmap along with pod to pass custom config to the server.