[Kubernetes]Accessing cluster nginx, dashboard and other services on the external public IP

DevOps Kubernetes
#21

Whether you need that or not depends on your platform. In a cloud use the cloud facilities. On-premise you will need something like MetalLB to cater for this.

#22

(Generally speaking it may not be worth implementing for this sole purpose.)

#23

I think that you need to learn more about service. https://kubernetes.io/es/docs/concepts/services-networking/service/

#24

I don’t want the service to run constantly at external IP but would like to know if it works at external IP for which port-forward is the solution I understood. but if I do the port-forward it does not show at the public IP.

#25

Cloud does not have LoadBalancer service so I have to have my service.

#26

Which cloud service are you referring to?

#27

Tried something like this?

#28

You won’t really learn unless you realize you need to learn.

#29

You have NOT understood, no.

#30

“Port-forward” is something completely different and no one has referred you to it in this thread.

#31

But due diligence says it’s your turn now, not ours.

#32

Yes, I did tried this hetzner solution but did not get it working, as you said you are right some understanding is required for me.

#33

Is it possible to have Ingress-controller without having floating IPs, Loadbalancer ?

#34

Please, read the text thatI post before.

#35

Floating IP is a terminology belonging elsewhere. Yes, you can expose an ingress controller using for example hostPort. And indeed, please read what you respond to.

#36

(I also had an IC running as a DaemonSet fronted by an external load balancer.)

#37

I suppose I have left one ingress-controller service without properly deleting and it is obstructing for other things. How to get rid of this svc

error: failed to create ingress: Internal error occurred: failed calling webhook "[validate.nginx.ingress.kubernetes.io](http://validate.nginx.ingress.kubernetes.io)": failed to call webhook: Post "[https://m-ingress-ingress-nginx-controller-admission.mondas-ns-102.svc:443/networking/v1/ingresses?timeout=10s](https://m-ingress-ingress-nginx-controller-admission.mondas-ns-102.svc:443/networking/v1/ingresses?timeout=10s)": service "m-ingress-ingress-nginx-controller-admission" not found```
#38

I have setup a local vagrant setup there I have setup working with metallb and expected that hetzner is also VM environment should work with metlalb but it is failing with the same procedure.

NAME                          READY   STATUS    RESTARTS       AGE
pod/nginx-6799fc88d8-nmz75    1/1     Running   0              4d4h
pod/nginx2-5fc4444698-xm4kc   1/1     Running   2 (4d4h ago)   5d3h

NAME                 TYPE           CLUSTER-IP       EXTERNAL-IP     PORT(S)        AGE
service/kubernetes   ClusterIP      10.96.0.1        <none>          443/TCP        13d
service/nginx        LoadBalancer   10.100.157.191   172.16.16.241   80:30975/TCP   5d4h
service/nginx2       LoadBalancer   10.100.26.178    172.16.16.242   80:31563/TCP   5d3h

NAME                     READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/nginx    1/1     1            1           5d4h
deployment.apps/nginx2   1/1     1            1           5d3h

NAME                                DESIRED   CURRENT   READY   AGE
replicaset.apps/nginx-6799fc88d8    1         1         1       5d4h
replicaset.apps/nginx2-5fc4444698   1         1         1       5d3h```
#39

What setup is it you think you have working? because there’s no ingress here, you have exposed your nginx instances via LB-type services that worked owing to MetalLB acting as CCM

#40

Plus the above is not an “ingress-controller service”, it’s an Ingress