Know your actual purchasers: Use PROXY protocol on Pink Hat OpenShift on IBM Cloud

As of 14 June 2023, PROXY protocol is supported for Ingress Controllers in Pink Hat OpenShift on IBM Cloud clusters hosted on VPC infrastructure.

Introduction

Fashionable software program architectures usually embody a number of layers of proxies and cargo balancers. Preserving the IP tackle of the unique consumer via these layers is difficult, however is likely to be required on your use instances. A possible answer for the issue is to make use of PROXY Protocol.

Beginning with Pink Hat OpenShift on IBM Cloud model 4.13, PROXY protocol is now supported for Ingress Controllers in clusters hosted on VPC infrastructure.

In case you are keen on utilizing PROXY protocol for Ingress Controllers on IBM Cloud Kubernetes Service clusters, yow will discover extra data in our earlier weblog publish.

Establishing PROXY protocol for OpenShift Ingress Controllers

When utilizing PROXY protocol for supply tackle preservation, all proxies that terminate TCP connections within the chain have to be configured to ship and obtain PROXY protocol headers after initiating L4 connections. Within the case of Pink Hat OpenShift on IBM Cloud clusters working on VPC infrastructure, we’ve got two proxies: the VPC Software Load Balancer (ALB) and the Ingress Controller.

On OpenShift clusters, the Ingress Operator is chargeable for managing the Ingress Controller cases and the load balancers used to reveal the Ingress Controllers. The operator watches IngressController assets on the cluster and makes changes to match the specified state.

Due to the Ingress Operator, we are able to allow PROXY protocol for each of our proxies directly. All we have to do is to alter the endpointPublishingStrategy configuration on our IngressController useful resource:

endpointPublishingStrategy:
  sort: LoadBalancerService
  loadBalancer:
    scope: Exterior
    providerParameters:
      sort: IBM
      ibm:
        protocol: PROXY

Once you apply the earlier configuration, the operat,or switches the Ingress Controller into PROXY protocol mode and provides the service.kubernetes.io/ibm-load-balancer-cloud-provider-enable-features: "proxy-protocol" annotation to the corresponding LoadBalancer typed Service useful resource, enabling PROXY protocol for the VPC ALB.

Instance

On this instance, we deployed a check utility in a single-zone Pink Hat OpenShift on IBM Cloud 4.13 cluster that makes use of VPC technology 2 compute. The appliance accepts HTTP connections and returns details about the obtained requests, such because the consumer tackle. The appliance is uncovered by the default-router created by the OpenShift Ingress Operator on the echo.instance.com area.

Shopper data with out utilizing PROXY protocol

By default, the PROXY protocol isn’t enabled. Let’s check accessing the applying:

$ curl https://echo.instance.com

Hostname: test-application-cd7cd98f7-9xbvm

Pod Info:
    -no pod data available-

Server values:
    server_version=nginx: 1.13.3 - lua: 10008

Request Info:
    client_address=172.24.84.165
    methodology=GET
    actual path=/
    question=
    request_version=1.1
    request_scheme=http
    request_uri=http://echo.instance.com:8080/

Request Headers:
    settle for=*/*
    forwarded=for=10.240.128.45;host=echo.instance.com;proto=https
    host=echo.instance.com
    user-agent=curl/7.87.0
    x-forwarded-for=10.240.128.45
    x-forwarded-host=echo.instance.com
    x-forwarded-port=443
    x-forwarded-proto=https

Request Physique:
    -no physique in request-

As you may see, the tackle within the x-forwarded-for header 10.240.128.45 doesn’t match your tackle. That’s the employee node’s tackle that obtained the request from the VPC load balancer. Which means we can’t get well the unique tackle of the consumer:

$ kubectl get nodes
NAME            STATUS   ROLES           AGE     VERSION
10.240.128.45   Prepared    grasp,employee   5h33m   v1.26.3+b404935
10.240.128.46   Prepared    grasp,employee   5h32m   v1.26.3+b404935

Enabling PROXY protocol on the default ingress controller

First, edit the Ingress Controller useful resource:

oc -n openshift-ingress-operator edit ingresscontroller/default

Within the Ingress controller useful resource, discover the spec.endpointPublishingStrategy.loadBalancer part and outline the next providerParameters values:

endpointPublishingStrategy:
  loadBalancer:
    providerParameters:
      sort: IBM
      ibm:
        protocol: PROXY
    scope: Exterior
  sort: LoadBalancerService

Then, save and apply the useful resource.

Shopper data utilizing PROXY protocol

Wait till the default-router pods are recycled and check entry to the applying once more:

$ curl https://echo.instance.com


Hostname: test-application-cd7cd98f7-9xbvm

Pod Info:
    -no pod data available-

Server values:
    server_version=nginx: 1.13.3 - lua: 10008

Request Info:
    client_address=172.24.84.184
    methodology=GET
    actual path=/
    question=
    request_version=1.1
    request_scheme=http
    request_uri=http://echo.instance.com:8080/

Request Headers:
    settle for=*/*
    forwarded=for=192.0.2.42;host=echo.instance.com;proto=https
    host=echo.instance.com
    user-agent=curl/7.87.0
    x-forwarded-for=192.0.2.42
    x-forwarded-host=echo.instance.com
    x-forwarded-port=443
    x-forwarded-proto=https

Request Physique:
    -no physique in request-

This time, yow will discover the precise consumer tackle 192.0.2.42 within the request headers, which is the precise public IP tackle of the unique consumer.

Limitations

The PROXY protocol characteristic on Pink Hat OpenShift on IBM Cloud is supported for under VPC technology 2 clusters that run 4.13 OpenShift model or later.

Extra data

For extra data, take a look at our official documentation about exposing apps with load balancers, enabling PROXY protocol for Ingress Controllers or the Pink Hat OpenShift documentation.