Ingress controller to route TCP traffic
I answered a similar question on another thread. How to use nginx ingress TCP service on different namespace
Basically, you can specify the port and backend for your service in configmap.
The following is the link to the document. https://github.com/kubernetes/ingress-nginx/blob/master/docs/user-guide/exposing-tcp-udp-services.md
rasthiya
Updated on June 04, 2020Comments
-
rasthiya almost 4 years
I'm trying to setup an ingress controller(nginx) to forward some TCP traffic to a kubernetes service(GCP). There's this tutorial that shows how to route HTTP traffic to services (based on the paths) using nginx. I want to have a similar setup to forward TCP traffic.
In my cluster, I have a pod running a TCP echo server written in python using sockets. There is a service attached to the pod. If I set the service type of this service to LoadBalancer, I can run my client as follows and get the echo from the cluster.
python client.py --host <EXTERNAL-IP-OF-LOAD-BALANCER> --port <PORT>
Similar to the echo server, I have other TCP services in my cluster that serves other pods. Currently I have set all of them to LoadBalancers. So, they have external IPs and listen for traffic on different ports. However, I do not want to create LoadBalancers to all of these services. How would I use the nginx to route the TCP traffic to different services based on the port numbers. If nginx cannot do this, are there other options that I can use to achieve this?
UPDATE: Following the HangDu's answer I created the following files.
apiVersion: v1 kind: ConfigMap metadata: name: tcp-services namespace: default data: 9000: "default/echo-service:50000"
and
apiVersion: v1 kind: Service metadata: name: ingress-nginx namespace: default labels: app.kubernetes.io/name: ingress-nginx app.kubernetes.io/part-of: ingress-nginx spec: type: LoadBalancer ports: - name: http port: 80 targetPort: 80 protocol: TCP - name: https port: 443 targetPort: 443 protocol: TCP - name: proxied-tcp-9000 port: 9000 targetPort: 9000 protocol: TCP selector: app.kubernetes.io/name: ingress-nginx app.kubernetes.io/part-of: ingress-nginx
Then I used
kubectl create -f <FILE_NAME>
to create the config map and the service. So I was hoping I could use the external IP of the newly created service and the port 9000 and runpython client.py --host <EXTERNAL-IP-OF-LOAD-BALANCER> --port 9000
to run my echo client. However, I get a connection refused error when I do that. Am I doing something wrong? -
AndreyS about 3 yearscan you please assist me on my issue? I desrbibed it completely... Please if you can assist me on similar issue. Although I installed "ingress" as Service in completely different way using the Helm that is why I am confused... Thank you!!! stackoverflow.com/questions/66190275/…