Privately exposing apps with Ingress

Step 1: Deploy apps and create app services

Start by deploying your apps and creating Kubernetes services to expose them.

1. Deploy your app to the cluster. Ensure that you add a label to your deployment in the metadata section of your configuration file, such as app: code. This label is needed to identify all pods where your app runs so that the pods are in the Ingress load balancing.

2. For each app deployment that you want to expose, create a Kubernetes ClusterIP service. Your app must be exposed by a Kubernetes service to be included in the Ingress load balancing.

oc expose deploy <app_deployment_name> --name my-app-svc --port <app_port> -n <namespace>

Step 2: Set up TLS termination with TLS certificates and Kubernetes secrets

Your TLS certificate must be stored as a Kubernetes secret in each namespace where your apps exist.

Step 3: Create and configure a private Ingress controller

After you get your domain and TLS certificate ready, you must create a private Ingress controller and configure the controller with your domain.

1. Create a configuration file for a private Ingress controller.

   apiVersion: operator.openshift.io/v1
   kind: IngressController
   metadata:
     name: private-ingress-controller
     namespace: openshift-ingress-operator
   spec:
     #defaultCertificate: If you are using a custom domain, specify the domain certificate
       #name: custom-certs-default
     replicas: 2
     domain: <domain>
     endpointPublishingStrategy:
       loadBalancer:
         scope: Internal
       type: LoadBalancerService
   

2. Create the IngressController resource in the openshift-ingress-operator project of your cluster. When you create the IngressController, a private Ingress controller is automatically created and deployed in the openshift-ingress project based on the IngressController settings that you set in the previous step. Additionally, an Ingress controller service is created to expose the Ingress controller with an IP address (classic clusters) or a VPC hostname (VPC clusters).

   oc create -f private-ingress-controller.yaml -n openshift-ingress-operator
   

3. Run the oc get command and find the IP address or VPC hostname in the EXTERNAL IP field of the router-private-ingress-controller service.

   oc get svc router-private-ingress-controller -n openshift-ingress
   

   Example output for classic clusters.

   NAME                                  TYPE           CLUSTER-IP       EXTERNAL-IP    PORT(S)                      AGE
   router-private-ingress-controller     LoadBalancer   172.21.57.132    10.XX.XX.XX    80/TCP,443/TCP,1940/TCP      3m
   

   Example output for VPC clusters:

   NAME                                  TYPE           CLUSTER-IP       EXTERNAL-IP                             PORT(S)                      AGE
   router-private-ingress-controller     LoadBalancer   172.21.57.132    1234abcd-us-south.lb.appdomain.cloud    80/TCP,443/TCP,1940/TCP      3m
   

4. Register the service's external IP address or VPC hostname with the domain you previously chose.
   • Custom domain: Work with your DNS provider to add the router-private-ingress-controller service's external IP address as an A record (classic clusters) or VPC hostname as a CNAME (VPC clusters) that maps to your custom domain.
   • IBM-provided domain: Create a DNS entry for the router-private-ingress-controller service's VPC hostname. When you run the following command, the subdomain that you specified in the private-ingress-controller.yaml file is automatically generated, and is registered with the router-private-ingress-controller service. A TLS secret for the domain is automatically generated in the project that you specify where your app runs. The secret name follows a truncated format of the subdomain, such as mycluster-a1b2cdef345678g9hi012j3kl4567890-0003.

     ibmcloud oc nlb-dns create vpc-gen2 --cluster <cluster_name_or_ID> --lb-host <VPC_hostname> --secret-namespace <project>
     

Step 4: Create the Ingress resource

Ingress resources define the routing rules that the Ingress controller uses to route traffic to your app service.

1. Define an Ingress resource configuration file that uses the IBM-provided domain or your custom domain to route incoming network traffic to the services that you created earlier.

   apiVersion: networking.k8s.io/v1
   kind: Ingress
   metadata:
     name: myingressresource
   spec:
     tls:
     - hosts:
       - <subdomain>
       secretName: <custom_secret_name>
     rules:
     - host: <subdomain>
       http:
         paths:
         - path: /<app1_path>
           backend:
               service:
                   name: <app1_service>
                   port:
                       number: 80
         - path: /<app2_path>
           backend:
               serivce:
                 name: <app2_service>
                 port:
                   number: 80
   

   tls

   If you want to use TLS, include this TLS section in your resource. Replace <domain> with your subdomain. Do not use * for your host or leave the host property empty to avoid failures during Ingress creation.

   Replace <tls_secret_name> with the secret that you created earlier that holds your TLS certificate and key for a custom domain or the TLS secret that was automatically generated for an IBM-provided subdomain.

   host

   Replace <domain> with your subdomain.

   If your cluster has multiple projects where apps are exposed, one Ingress resource is required per project. You can use the same subdomain in each resource or different subdomains in each resource. For example, if you use a wildcard domain, you can append a wildcard subdomain to the beginning of the domain, such as subdomain1.custom_domain.net.

   Do not use * for your host or leave the host property empty to avoid failures during Ingress creation.

   path

   Replace <app_path> with the path that your app is listening on. The path is appended to the IBM-provided or your custom domain to create a unique route to your app. When you enter this route into a web browser, network traffic is routed to the Ingress controller. The Ingress controller looks up the associated service, and sends network traffic to the service. The service then forwards the traffic to the pods where the app runs. Many apps don't listen on a specific path, but use the root path and a specific port. In this case, define the root path as / and don't specify an individual path for your app.

   For example, to use http://domain/, enter / as the path. For http://domain/app1_path, enter /app1_path as the path.

   serviceName

   Replace <app1_service> and <app2_service>, and so on, with the name of the services you created to expose your apps. If your apps are exposed by services in different projects in the cluster, include only app services that are in the same project. You must create one Ingress resource for each project that contains apps that you want to expose.

   servicePort

   The port that your service listens to. Use the same port that you defined when you created the Kubernetes service for your app.

2. Create the Ingress resource for your cluster. Ensure that the resource deploys into the same project as the app services that you specified in the resource.

   oc apply -f myingressresource.yaml -n <project>
   

3. Verify that the Ingress resource was created successfully. If messages in the events describe an error in your resource configuration, fix the values in your resource file and reapply the file for the resource.

   oc describe ingress myingressresource
   

Your Ingress resource is created in the same project as your app services, and your apps are registered with the Ingress controller.

Step 5: Access your app from your private network

1. Classic clusters: Before you can access your app, make sure that you can access a DNS service. To use the default external DNS provider, you must configure edge nodes with public access and configure a Virtual Router Appliance.

2. From within your private network, enter the URL of the app service in a web browser.

   https://<domain>/<app1_path>
   

   If you exposed multiple apps, access those apps by changing the path that is appended to the URL.

   https://<domain>/<app2_path>
   

   If you use a wildcard domain, access those apps with their own subdomains.

   http://<subdomain1>.<domain>/<app1_path>
   

   http://<subdomain2>.<domain>/<app1_path>
   

Can't connect to your app through Ingress? Try Troubleshooting Ingress.

Step 1: Deploy apps and create app services

Start by deploying your apps and creating Kubernetes services to expose them.

1. Deploy your app to the cluster. Ensure that you add a label to your deployment in the metadata section of your configuration file, such as app: code. This label is needed to identify all pods where your app runs so that the pods are in the Ingress load balancing.

2. For each app deployment that you want to expose, create a Kubernetes ClusterIP service. Your app must be exposed by a Kubernetes service to be included in the Ingress load balancing.

oc expose deploy <app_deployment_name> --name my-app-svc --port <app_port> -n <namespace>

Step 2: Set up TLS termination with TLS certificates and Kubernetes secrets

Your TLS certificate must be stored as a Kubernetes secret in each namespace where your apps exist.

• To use the IBM-managed Ingress domain, see Setting up TLS secrets for the IBM-provided Ingress subdomain.

• To use a domain that you created yourself, such as a domain registered with an external provider, see Setting up TLS secrets for custom subdomains.

Step 3: Create the Ingress resource

Ingress resources define the routing rules that the Ingress controller uses to route traffic to your app service.

1. Define an Ingress resource configuration file that uses the IBM-provided domain or your custom domain to route incoming network traffic to the services that you created earlier.

   apiVersion: networking.k8s.io/v1 # For 4.5 or earlier, use networking.k8s.io/v1beta1 instead
   kind: Ingress
   metadata:
     name: myingressresource
   spec:
     tls:
     - hosts:
       - <custom_domain>
       secretName: <custom_secret_name>
     rules:
     - host: <domain>
       http:
         paths:
         - path: /<app1_path>
           backend:
               service:
                 name: <app1_service>
                 port:
                   number: 80
         - path: /<app2_path>
           backend:
               service:
                 name: <app2_service>
                 port:
                   number: 80
   

   tls

   If you want to use TLS, include this TLS section in your resource.

   Replace <domain> with your subdomain. Do not use * for your host or leave the host property empty to avoid failures during Ingress creation.

   Replace <tls_secret_name> with the secret that you created earlier that holds your TLS certificate and key for a custom domain or the TLS secret that was automatically generated for an IBM-provided subdomain.

   host

   Replace <domain> with the IBM-provided Ingress subdomain or your custom domain.

   If your cluster has multiple projects where apps are exposed, one Ingress resource is required per project. You can use the same subdomain in each resource or different subdomains in each resource. For example, if you use a wildcard domain, you can append a wildcard subdomain to the beginning of the domain, such as subdomain1.custom_domain.net or subdomain1.mycluster-<hash>-0000.us-south.containers.appdomain.cloud. Do not use* for your host or leave the host property empty to avoid failures during Ingress creation.

   path

   Replace <app_path> with the path that your app is listening on. The path is appended to the IBM-provided or your custom domain to create a unique route to your app. When you enter this route into a web browser, network traffic is routed to the Ingress controller. The Ingress controller looks up the associated service, and sends network traffic to the service. The service then forwards the traffic to the pods where the app runs. Many apps don't listen on a specific path, but use the root path and a specific port. In this case, define the root path as / and don't specify an individual path for your app.

   For example, to use http://domain/, enter / as the path. For http://domain/app1_path, enter /app1_path as the path.

   name

   Replace <app1_service> and <app2_service>, and so on, with the name of the services you created to expose your apps. If your apps are exposed by services in different projects in the cluster, include only app services that are in the same project. You must create one Ingress resource for each project where you have apps that you want to expose.

   port

   The port that your service listens to. Use the same port that you defined when you created the Kubernetes service for your app.

2. Create the Ingress resource for your cluster. Ensure that the resource deploys into the same project as the app services that you specified in the resource.

   oc apply -f myingressresource.yaml -n <project>
   

3. Verify that the Ingress resource was created successfully. If messages in the events describe an error in your resource configuration, fix the values in your resource file and reapply the file for the resource.

   oc describe ingress myingressresource
   

Your Ingress resource is created in the same project as your app services, and your apps are registered with the Ingress controller.

Step 4: Access your app

In a web browser, enter the URL of the app service to access.

https://<domain>/<app1_path>

If you exposed multiple apps, access those apps by changing the path that is appended to the URL.

https://<domain>/<app2_path>

If you use a wildcard domain, access those apps with their own subdomains.

http://<subdomain1>.<domain>/<app1_path>

http://<subdomain2>.<domain>/<app1_path>

Can't connect to your app through Ingress? Try Troubleshooting Ingress.