Deploy an Example Stack

Video Demonstration

A video demonstration of this example is available here.

Version

This example was written using Docker Desktop with Kubernetes enabled on the Mac platform. The version used for this guide was 4.11.1(84025), which includes Docker Engine v20.10.17 and Kubernetes v1.24.2. The ingress-nginx controller version was 1.3.0.

Latest Docker Desktop Warning

Docker Desktop 4.12 introduced Kubernetes v1.25. This release dropped some deprecated functionality that impacts the ingress controller used in this example. Until the Helm charts for the controller are updated to accommodate the removed APIs, this example will not work on version 4.12 or later.

Kubernetes Services Kubernetes versus Server-Deployed Applications

If you are new to Kubernetes-based deployments, there is a distinct difference when running under Kubernetes compared to running applications on servers. In a server model, many applications typically run on the same server, and you can access any of them using the same host. For example, many on-premise deployments of PingFederate also include the PingDataConsole, hosted on the same server.

Under Kubernetes, however, each application that requires external access is associated with a service. A service is a fixed endpoint in the cluster that routes traffic to a given application. So, in this example, there are distinct service endpoints for PingFederate, PingDataConsole, and the other products.

In this demo, these service endpoints are load balanced using the Nginx ingress controller. By adding entries to the /etc/hosts file, you can access them using typical URL entries.

The Ping Identity Helm Getting Started page has instructions on getting your environment configured for using the Ping Helm charts.

For more examples, see Helm Chart Example Configurations.

For more information on Helm with Ping products, see Ping Identity DevOps Helm Charts.

What You Will Do

After using Git to clone the pingidentity-devops-getting-started repository, you will use Helm to deploy a sample stack to a Kubernetes cluster.

Prerequisites

• Register for the Ping DevOps program and install/configure pingctl with your User and Key
• Install Git
• Follow the instructions on the helm Getting Started page up through updating to the latest charts to ensure you have the latest version of our charts
• Access to a Kubernetes cluster. You can enable Kubernetes in Docker Desktop for a simple cluster, which was the cluster used for this guide (on the Mac platform).

Clone the getting-started repository

1. Clone the pingidentity-devops-getting-started repository to your local ${PING_IDENTITY_DEVOPS_HOME} directory.

   The ${PING_IDENTITY_DEVOPS_HOME} environment variable was set by running pingctl config.

   cd "${PING_IDENTITY_DEVOPS_HOME}"
   git clone \
     https://github.com/pingidentity/pingidentity-devops-getting-started.git
   

Deploy the example stack

1. Deploy the example stack of our product containers.

   Initial Deployment

   For this guide, avoid making changes to the everything.yaml file to ensure a successful first-time deployment.

   1. Create a namespace for running the stack in your Kubernetes cluster.

      # Create the namespace
      kubectl create ns pinghelm
      
      # Set the kubectl context to the namespace
      kubectl config set-context --current --namespace=pinghelm
      
      # Confirm
      kubectl config view --minify | grep namespace:
      

   2. Deploy the ingress controller to Docker Desktop:

      helm upgrade --install ingress-nginx ingress-nginx --repo https://kubernetes.github.io/ingress-nginx --namespace ingress-nginx --create-namespace
      

   3. To wait for the Nginx ingress to reach a healthy state, run the following command. You can also observe the pod status using k9s or by running kubectl get pods --namespace ingress-nginx. You should see one controller pod running when the ingress controller is ready. This command should exit after no more than 90 seconds or so, depending on the speed of your computer:

      kubectl wait --namespace ingress-nginx \
        --for=condition=ready pod \
        --selector=app.kubernetes.io/component=controller \
        --timeout=90s
      

   4. Create a secret in the namespace you will be using to run the example (pinghelm) using the pingctl utility. This secret will obtain an evaluation license based on your Ping DevOps username and key:

      pingctl k8s generate devops-secret | kubectl apply -f -
      

   5. This example will use the Helm release name demo and DNS domain suffix *ping-local.com for accessing applications. Add all expected hosts to /etc/hosts:

      echo '127.0.0.1 demo-pingaccess-admin.ping-local.com demo-pingaccess-engine.ping-local.com demo-pingauthorize.ping-local.com demo-pingauthorizepap.ping-local.com demo-pingdataconsole.ping-local.com demo-pingdelegator.ping-local.com demo-pingdirectory.ping-local.com demo-pingdatagovernance.ping-local.com demo-pingdatagovernancepap.ping-local.com demo-pingfederate-admin.ping-local.com demo-pingfederate-engine.ping-local.com demo-pingcentral.ping-local.com' | sudo tee -a /etc/hosts > /dev/null
      

   6. To install the chart, go to your local "${PING_IDENTITY_DEVOPS_HOME}"/pingidentity-devops-getting-started/30-helm directory and run the command shown here. In this example, the release (deployment into Kubernetes by Helm) is called demo, forming the prefix for all objects created. The ingress-demo.yaml file configures the ingresses for the products to use the ping-local domain:

      helm upgrade --install demo pingidentity/ping-devops -f everything.yaml -f ingress-demo.yaml
      

      The product Docker images are automatically pulled if they have not previously been pulled from Docker Hub.

      Sample output:

      NAME: demo
      LAST DEPLOYED: Fri Sep  2 10:36:51 2022
      NAMESPACE: pinghelm
      STATUS: deployed
      REVISION: 1
      TEST SUITE: None
      NOTES:
      #-------------------------------------------------------------------------------------
      # Ping DevOps
      #
      # Description: Ping Identity helm charts - 08/05/22
      #-------------------------------------------------------------------------------------
      #
      #           Product          tag   typ  #  cpu R/L   mem R/L  Ing
      #    --------------------- ------- --- -- --------- --------- ---
      #    global                2207              0/0       0/0     √
      #
      #  √ pingaccess-admin      2207               /         /
      #  √ pingaccess-engine     2207               /         /
      #  √ pingauthorize         2207               /         /
      #    pingauthorizepap
      #    pingcentral
      #  √ pingdataconsole       2207               /         /
      #    pingdatagovernance
      #    pingdatagovernancepap
      #    pingdatasync
      #    pingdelegator
      #  √ pingdirectory         2207               /         /
      #    pingdirectoryproxy
      #  √ pingfederate-admin    2207               /         /
      #  √ pingfederate-engine   2207               /         /
      #    pingintelligence
      #
      #    ldap-sdk-tools
      #    pd-replication-timing
      #    pingtoolkit
      #    apache-jmeter
      #
      #-------------------------------------------------------------------------------------
      # To see values info, simply set one of the following on your helm install/upgrade
      #
      #    --set help.values=all         # Provides all (i.e. .Values, .Release, .Chart, ...) yaml
      #    --set help.values=global      # Provides global values
      #    --set help.values={ image }   # Provides image values merged with global         
      

      As you can see, PingAccess Admin and Engine, PingData Console, PingDirectory, PingAuthorize, and the PingFederate Admin and Engine are deployed from the provided everything.yaml values file.

      It will take several minutes for all components to become operational.

   7. To display the status of the deployed components, you can use k9s or issue the corresponding commands shown here:

      • Display the services (endpoints for connecting) by running kubectl get service --selector=app.kubernetes.io/instance=demo

      NAME                            TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)                   AGE
      demo-pingaccess-admin           ClusterIP   10.106.227.103   <none>        9090/TCP,9000/TCP         15m
      demo-pingaccess-admin-cluster   ClusterIP   None             <none>        <none>                    15m
      demo-pingaccess-engine          ClusterIP   10.108.102.245   <none>        3000/TCP                  15m
      demo-pingauthorize              ClusterIP   10.110.95.132    <none>        443/TCP                   15m
      demo-pingauthorize-cluster      ClusterIP   None             <none>        1636/TCP                  15m
      demo-pingdataconsole            ClusterIP   10.97.81.22      <none>        8443/TCP                  15m
      demo-pingdirectory              ClusterIP   10.102.91.214    <none>        443/TCP,389/TCP,636/TCP   15m
      demo-pingdirectory-cluster      ClusterIP   None             <none>        1636/TCP                  15m
      demo-pingfederate-admin         ClusterIP   10.99.145.24     <none>        9999/TCP                  15m
      demo-pingfederate-cluster       ClusterIP   None             <none>        7600/TCP,7700/TCP         15m
      demo-pingfederate-engine        ClusterIP   10.108.240.203   <none>        9031/TCP                  15m
      

      • To view the pods, run kubectl get pods --selector=app.kubernetes.io/instance=demo - you will need to run this at intervals until all pods have started ( Running status):

      NAME                                        READY   STATUS    RESTARTS   AGE
      demo-pingaccess-admin-0                     1/1     Running   0          7m29s
      demo-pingaccess-engine-cf9987bb5-npspz      1/1     Running   0          7m52s
      demo-pingauthorize-8bdfd4fd8-j82zg          1/1     Running   0          7m43s
      demo-pingdataconsole-7c875985d4-mfjbq       1/1     Running   0          17m
      demo-pingdirectory-0                        1/1     Running   0          7m38s
      demo-pingfederate-admin-5786787dfd-5b5s5    1/1     Running   0          7m36s
      demo-pingfederate-engine-5ff6546f4f-7jfnt   1/1     Running   0          7m31s
      

      • To see the ingresses you will use to access the product, run kubectl get ingress. If the ingress controller is configured properly, you should see localhost as the address as shown here:

      NAME                       CLASS    HOSTS                                     ADDRESS     PORTS     AGE
      demo-pingaccess-admin      <none>   demo-pingaccess-admin.ping-local.com      localhost   80, 443   5m23s
      demo-pingaccess-engine     <none>   demo-pingaccess-engine.ping-local.com     localhost   80, 443   5m23s
      demo-pingauthorize         <none>   demo-pingauthorize.ping-local.com         localhost   80, 443   5m23s
      demo-pingdataconsole       <none>   demo-pingdataconsole.ping-local.com       localhost   80, 443   5m23s
      demo-pingdirectory         <none>   demo-pingdirectory.ping-local.com         localhost   80, 443   5m23s
      demo-pingfederate-admin    <none>   demo-pingfederate-admin.ping-local.com    localhost   80, 443   5m23s
      demo-pingfederate-engine   <none>   demo-pingfederate-engine.ping-local.com   localhost   80, 443   5m23s
      

      Address must be localhost

      If the ingress controller is working properly, the ingress definitions will all report the ADDRESS column as localhost as shown above. If you do not see this entry, then you will not be able to access the services later. This problem is due to a known error with Docker Desktop and the embedded virtual machine (VM) used on the Mac and Windows platform in combination with the ingress controller. To correct the problem, uninstall the chart as instructed at the bottom of this page and restart Docker Desktop. Afterward, you can re-run the helm command to install the Ping products as instructed above. The issue appears to be related to a stale networking configuration under the covers of Docker Desktop.

      • To see everything tied to the helm release run kubectl get all --selector=app.kubernetes.io/instance=demo:

      NAME                                            READY   STATUS    RESTARTS   AGE
      pod/demo-pingaccess-admin-0                     1/1     Running   0          107m
      pod/demo-pingaccess-engine-cf9987bb5-npspz      1/1     Running   0          107m
      pod/demo-pingauthorize-8bdfd4fd8-j82zg          1/1     Running   0          107m
      pod/demo-pingdataconsole-7c875985d4-mfjbq       1/1     Running   0          116m
      pod/demo-pingdirectory-0                        1/1     Running   0          107m
      pod/demo-pingfederate-admin-5786787dfd-5b5s5    1/1     Running   0          107m
      pod/demo-pingfederate-engine-5ff6546f4f-7jfnt   1/1     Running   0          107m
      
      NAME                                    TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)                   AGE
      service/demo-pingaccess-admin           ClusterIP   10.96.234.72     <none>        9090/TCP,9000/TCP         116m
      service/demo-pingaccess-admin-cluster   ClusterIP   None             <none>        <none>                    116m
      service/demo-pingaccess-engine          ClusterIP   10.106.190.217   <none>        3000/TCP                  116m
      service/demo-pingauthorize              ClusterIP   10.104.246.123   <none>        443/TCP                   116m
      service/demo-pingauthorize-cluster      ClusterIP   None             <none>        1636/TCP                  116m
      service/demo-pingdataconsole            ClusterIP   10.96.166.28     <none>        8443/TCP                  116m
      service/demo-pingdirectory              ClusterIP   10.107.42.8      <none>        443/TCP,389/TCP,636/TCP   116m
      service/demo-pingdirectory-cluster      ClusterIP   None             <none>        1636/TCP                  116m
      service/demo-pingfederate-admin         ClusterIP   10.105.26.94     <none>        9999/TCP                  116m
      service/demo-pingfederate-cluster       ClusterIP   None             <none>        7600/TCP,7700/TCP         116m
      service/demo-pingfederate-engine        ClusterIP   10.99.223.48     <none>        9031/TCP                  116m
      
      NAME                                       READY   UP-TO-DATE   AVAILABLE   AGE
      deployment.apps/demo-pingaccess-engine     1/1     1            1           116m
      deployment.apps/demo-pingauthorize         1/1     1            1           116m
      deployment.apps/demo-pingdataconsole       1/1     1            1           116m
      deployment.apps/demo-pingfederate-admin    1/1     1            1           116m
      deployment.apps/demo-pingfederate-engine   1/1     1            1           116m
      
      NAME                                                  DESIRED   CURRENT   READY   AGE
      replicaset.apps/demo-pingaccess-engine-cf9987bb5      1         1         1       116m
      replicaset.apps/demo-pingauthorize-8bdfd4fd8          1         1         1       116m
      replicaset.apps/demo-pingdataconsole-7c875985d4       1         1         1       116m
      replicaset.apps/demo-pingfederate-admin-5786787dfd    1         1         1       116m
      replicaset.apps/demo-pingfederate-engine-5ff6546f4f   1         1         1       116m
      
      NAME                                     READY   AGE
      statefulset.apps/demo-pingaccess-admin   1/1     116m
      statefulset.apps/demo-pingdirectory      1/1     116m
      

      • To view logs, look at the logs for the deployment of the product in question. For example:

       kubectl logs -f deployment/demo-pingfederate-admin
      

2. These are the URLs and credentials to sign on to the management consoles for the products.

   Certificates

   This example uses self-signed certificates that will have to be accepted in your browser or added to your keystore.

   With the ingresses in place, you can access the products at these URLs:

   Product	Connection Details
   PingFederate	URL: https://demo-pingfederate-admin.ping-local.com/pingfederate/app Username: administrator Password: 2FederateM0re
   PingDirectory	URL: https://demo-pingdataconsole.ping-local.com/console Server: ldaps://demo-pingdirectory-cluster:1636 Username: administrator Password: 2FederateM0re
   PingAccess	URL: https://demo-pingaccess-admin.ping-local.com/ Username: administrator Password: 2FederateM0re
   PingAuthorize	URL: https://demo-pingdataconsole.ping-local.com/console Server: ldaps://demo-pingauthorize-cluster:1636 Username: administrator Password: 2FederateM0re

3. When you are finished, you can remove the demonstration components by running the uninstall command for helm:

   helm uninstall demo
   

Next Steps

Now that you have deployed a set of our product images using the provided chart, you can move on to deployments using configurations that more closely reflect use cases to be explored. Refer to the helm examples) page for other typical deployments.

Container logging

Maintaining logs in a containerized model is different from the typical server-deployed application. See this page for additional details.