Tanzu Kubernetes Service

This sample demonstrates how to use the WebLogic Kubernetes Operator (hereafter “the operator”) to set up a WebLogic Server (WLS) cluster on the Tanzu Kubernetes Grid (TKG). After performing the sample steps, your WLS domain with a Model in Image domain source type runs on a TKG Kubernetes cluster instance. After the domain has been provisioned, you can monitor it using the WebLogic Server Administration console.

TKG is a managed Kubernetes Service that lets you quickly deploy and manage Kubernetes clusters. To learn more, see the Tanzu Kubernetes Grid (TKG) overview page.

Contents

Prerequisites

This sample assumes the following prerequisite environment setup:

  • WebLogic Kubernetes Operator: This document was tested with version v3.1.0.
  • Operating System: GNU/Linux.
  • Git; use git --version to test if git works. This document was tested with version 2.17.1.
  • TKG CLI; use tkg version to test if TKG works. This document was tested with version v1.1.3.
  • kubectl; use kubectl version to test if kubectl works. This document was tested with version v1.18.6.
  • Helm version 3.1 or later; use helm version to check the helm version. This document was tested with version v3.2.1.
Create a Tanzu Kubernetes cluster

Create the Kubernetes cluster using the TKG CLI. See the Tanzu documentation to set up your Kubernetes cluster. After your Kubernetes cluster is up and running, run the following commands to make sure kubectl can access the Kubernetes cluster:

$ kubectl get nodes -o wide
NAME                                    STATUS     ROLES    AGE     VERSION            INTERNAL-IP       EXTERNAL-IP       OS-IMAGE                 KERNEL-VERSION   CONTAINER-RUNTIME
k8s-cluster-101-control-plane-8nj7t     NotReady   master   2d20h   v1.18.6+vmware.1   192.168.100.147   192.168.100.147   VMware Photon OS/Linux   4.19.132-1.ph3   containerd://1.3.4
k8s-cluster-101-md-0-577b7dc766-552hn   Ready      <none>   2d20h   v1.18.6+vmware.1   192.168.100.148   192.168.100.148   VMware Photon OS/Linux   4.19.132-1.ph3   containerd://1.3.4
k8s-cluster-101-md-0-577b7dc766-m8wrc   Ready      <none>   2d20h   v1.18.6+vmware.1   192.168.100.149   192.168.100.149   VMware Photon OS/Linux   4.19.132-1.ph3   containerd://1.3.4
k8s-cluster-101-md-0-577b7dc766-p2gkz   Ready      <none>   2d20h   v1.18.6+vmware.1   192.168.100.150   192.168.100.150   VMware Photon OS/Linux   4.19.132-1.ph3   containerd://1.3.4
Oracle Container Registry

You will need an Oracle Container Registry account. The following steps will direct you to accept the Oracle Standard Terms and Restrictions to pull the WebLogic Server images. Make note of your Oracle Account password and email. This sample pertains to 12.2.1.4, but other versions may work as well.

Install WebLogic Kubernetes Operator

The WebLogic Kubernetes Operator is an adapter to integrate WebLogic Server and Kubernetes, allowing Kubernetes to serve as a container infrastructure hosting WLS instances. The operator runs as a Kubernetes Pod and stands ready to perform actions related to running WLS on Kubernetes.

Clone the WebLogic Kubernetes Operator repository to your machine. We will use several scripts in this repository to create a WebLogic domain. Kubernetes Operators use Helm to manage Kubernetes applications. The operator’s Helm chart is located in the kubernetes/charts/weblogic-operator directory. Install the operator by running the following commands.

Clone the repository.

$ git clone --branch v3.3.2 https://github.com/oracle/weblogic-kubernetes-operator.git
$ cd weblogic-kubernetes-operator

Grant the Helm service account the cluster-admin role.

$ cat <<EOF | kubectl apply -f -
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: helm-user-cluster-admin-role
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-admin
subjects:
- kind: ServiceAccount
  name: default
  namespace: kube-system
EOF

Create a namespace and service account for the operator.

$ kubectl create namespace sample-weblogic-operator-ns
namespace/sample-weblogic-operator-ns created
$ kubectl create serviceaccount -n sample-weblogic-operator-ns sample-weblogic-operator-sa
serviceaccount/sample-weblogic-operator-sa created

Install the operator.

$ helm install weblogic-operator kubernetes/charts/weblogic-operator \
  --namespace sample-weblogic-operator-ns \
  --set serviceAccount=sample-weblogic-operator-sa \
  --set "enableClusterRoleBinding=true" \
  --set "domainNamespaceSelectionStrategy=LabelSelector" \
  --set "domainNamespaceLabelSelector=weblogic-operator\=enabled" \
  --wait
NAME: weblogic-operator
LAST DEPLOYED: Tue Nov 17 09:33:58 2020
NAMESPACE: sample-weblogic-operator-ns
STATUS: deployed
REVISION: 1
TEST SUITE: None

Verify the operator with the following commands; the status will be running.

$ helm list -A
NAME                        NAMESPACE                     REVISION   UPDATED                                 STATUS       CHART                   APP VERSION
sample-weblogic-operator    sample-weblogic-operator-ns   1          2020-11-17 09:33:58.584239273 -0700 PDT deployed     weblogic-operator-3.1
$ kubectl get pods -n sample-weblogic-operator-ns
NAME                                 READY   STATUS    RESTARTS   AGE
weblogic-operator-775b668c8f-nwwnn   1/1     Running   0          32s

Create an image

Image creation prerequisites
  1. The JAVA_HOME environment variable must be set and must reference a valid JDK 8 or 11 installation.
  2. Copy the sample to a new directory; for example, use the directory /tmp/mii-sample.
$ mkdir /tmp/mii-sample
$ cp -r /root/weblogic-kubernetes-operator/kubernetes/samples/scripts/create-weblogic-domain/model-in-image/* /tmp/mii-sample

Note: We will refer to this working copy of the sample as /tmp/mii-sample; however, you can use a different location.

Download the latest WebLogic Deploying Tooling (WDT) and WebLogic Image Tool (WIT) installer ZIP files to your /tmp/mii-sample/model-images directory. Both WDT and WIT are required to create your Model in Image container images.

$ cd /tmp/mii-sample/model-images
$ curl -m 120 -fL https://github.com/oracle/weblogic-deploy-tooling/releases/latest/download/weblogic-deploy.zip \
  -o /tmp/mii-sample/model-images/weblogic-deploy.zip
$ curl -m 120 -fL https://github.com/oracle/weblogic-image-tool/releases/latest/download/imagetool.zip \
  -o /tmp/mii-sample/model-images/imagetool.zip

To set up the WebLogic Image Tool, run the following commands:

$ cd /tmp/mii-sample/model-images
$ unzip imagetool.zip
$ ./imagetool/bin/imagetool.sh cache addInstaller \
  --type wdt \
  --version latest \
  --path /tmp/mii-sample/model-images/weblogic-deploy.zip

These steps will install WIT to the /tmp/mii-sample/model-images/imagetool directory, plus put a wdt_latest entry in the tool’s cache which points to the WDT ZIP file installer. You will use WIT later in the sample for creating model images.

Image creation - Introduction

The goal of image creation is to demonstrate using the WebLogic Image Tool to create an image named model-in-image:WLS-v1 from files that you will stage to /tmp/mii-sample/model-images/model-in-image:WLS-v1/. The staged files will contain a web application in a WDT archive, and WDT model configuration for a WebLogic Administration Server called admin-server and a WebLogic cluster called cluster-1.

Overall, a Model in Image image must contain a WebLogic Server installation and a WebLogic Deploy Tooling installation in its /u01/wdt/weblogic-deploy directory. In addition, if you have WDT model archive files, then the image must also contain these files in its /u01/wdt/models directory. Finally, an image optionally may also contain your WDT model YAML file and properties files in the same /u01/wdt/models directory. If you do not specify a WDT model YAML file in your /u01/wdt/models directory, then the model YAML file must be supplied dynamically using a Kubernetes ConfigMap that is referenced by your Domain spec.model.configMap field. We provide an example of using a model ConfigMap later in this sample.

The following sections contain the steps for creating the image model-in-image:WLS-v1.

Understanding your first archive

The sample includes a predefined archive directory in /tmp/mii-sample/archives/archive-v1 that you will use to create an archive ZIP file for the image.

The archive top directory, named wlsdeploy, contains a directory named applications, which includes an ‘exploded’ sample JSP web application in the directory, myapp-v1. Three useful aspects to remember about WDT archives are:

  • A model image can contain multiple WDT archives.
  • WDT archives can contain multiple applications, libraries, and other components.
  • WDT archives have a well defined directory structure, which always has wlsdeploy as the top directory.

The application displays important details about the WebLogic Server instance that it’s running on: namely its domain name, cluster name, and server name, as well as the names of any data sources that are targeted to the server.

Staging a ZIP file of the archive

When you create the image, you will use the files in the staging directory, /tmp/mii-sample/model-in-image__WLS-v1. In preparation, you need it to contain a ZIP file of the WDT application archive.

Run the following commands to create your application archive ZIP file and put it in the expected directory:

# Delete existing archive.zip in case we have an old leftover version
$ rm -f /tmp/mii-sample/model-images/model-in-image__WLS-v1/archive.zip
# Move to the directory which contains the source files for our archive
$ cd /tmp/mii-sample/archives/archive-v1
# Zip the archive to the location will later use when we run the WebLogic Image Tool
$ zip -r /tmp/mii-sample/model-images/model-in-image__WLS-v1/archive.zip wlsdeploy
Staging model files

In this step, you explore the staged WDT model YAML file and properties in the /tmp/mii-sample/model-in-image__WLS-v1 directory. The model in this directory references the web application in your archive, configures a WebLogic Server Administration Server, and configures a WebLogic cluster. It consists of only two files, model.10.properties, a file with a single property, and, model.10.yaml, a YAML file with your WebLogic configuration model.10.yaml.

CLUSTER_SIZE=5

Here is the WLS model.10.yaml:

domainInfo:
    AdminUserName: '@@SECRET:__weblogic-credentials__:username@@'
    AdminPassword: '@@SECRET:__weblogic-credentials__:password@@'
    ServerStartMode: 'prod'

topology:
    Name: '@@ENV:CUSTOM_DOMAIN_NAME@@'
    AdminServerName: 'admin-server'
    Cluster:
        'cluster-1':
            DynamicServers:
                ServerTemplate:  'cluster-1-template'
                ServerNamePrefix: 'managed-server'
                DynamicClusterSize: '@@PROP:CLUSTER_SIZE@@'
                MaxDynamicClusterSize: '@@PROP:CLUSTER_SIZE@@'
                MinDynamicClusterSize: '0'
                CalculatedListenPorts: false
    Server:
        'admin-server':
            ListenPort: 7001
    ServerTemplate:
        'cluster-1-template':
            Cluster: 'cluster-1'
            ListenPort: 8001

appDeployments:
    Application:
        myapp:
            SourcePath: 'wlsdeploy/applications/myapp-v1'
            ModuleType: ear
            Target: 'cluster-1'

The model files:

  • Define a WebLogic domain with:

    • Cluster cluster-1
    • Administration Server admin-server
    • A cluster-1 targeted EAR application that’s located in the WDT archive ZIP file at wlsdeploy/applications/myapp-v1
  • Leverage macros to inject external values:

    • The property file CLUSTER_SIZE property is referenced in the model YAML file DynamicClusterSize and MaxDynamicClusterSize fields using a PROP macro.
    • The model file domain name is injected using a custom environment variable named CUSTOM_DOMAIN_NAME using an ENV macro.
      • You set this environment variable later in this sample using an env field in its Domain.
      • This conveniently provides a simple way to deploy multiple differently named domains using the same model image.
    • The model file administrator user name and password are set using a weblogic-credentials secret macro reference to the WebLogic credential secret.
      • This secret is in turn referenced using the webLogicCredentialsSecret field in the Domain.
      • The weblogic-credentials is a reserved name that always dereferences to the owning Domain actual WebLogic credentials secret name.

A Model in Image image can contain multiple properties files, archive ZIP files, and YAML files but in this sample you use just one of each. For a complete discussion of Model in Images model file naming conventions, file loading order, and macro syntax, see Model files files in the Model in Image user documentation.

Creating the image with WIT

At this point, you have staged all of the files needed for the image model-in-image:WLS-v1; they include:

  • /tmp/mii-sample/model-images/weblogic-deploy.zip
  • /tmp/mii-sample/model-images/model-in-image__WLS-v1/model.10.yaml
  • /tmp/mii-sample/model-images/model-in-image__WLS-v1/model.10.properties
  • /tmp/mii-sample/model-images/model-in-image__WLS-v1/archive.zip

If you don’t see the weblogic-deploy.zip file, then you missed a step in the prerequisites.

Now, you use the Image Tool to create an image named model-in-image:WLS-v1 that’s layered on a base WebLogic image. You’ve already set up this tool during the prerequisite steps.

Run the following commands to create the model image and verify that it worked:

$ cd /tmp/mii-sample/model-images
$ ./imagetool/bin/imagetool.sh update \
  --tag model-in-image:WLS-v1 \
  --fromImage container-registry.oracle.com/middleware/weblogic:12.2.1.4 \
  --wdtModel      ./model-in-image__WLS-v1/model.10.yaml \
  --wdtVariables  ./model-in-image__WLS-v1/model.10.properties \
  --wdtArchive    ./model-in-image__WLS-v1/archive.zip \
  --wdtModelOnly \
  --wdtDomainType WLS \
  --chown oracle:root

If you don’t see the imagetool directory, then you missed a step in the prerequisites.

This command runs the WebLogic Image Tool in its Model in Image mode, and does the following:

  • Builds the final image as a layer on the container-registry.oracle.com/middleware/weblogic:12.2.1.4 base image.
  • Copies the WDT ZIP file that’s referenced in the WIT cache into the image.
    • Note that you cached WDT in WIT using the keyword latest when you set up the cache during the sample prerequisites steps.
    • This lets WIT implicitly assume it’s the desired WDT version and removes the need to pass a -wdtVersion flag.
  • Copies the specified WDT model, properties, and application archives to image location /u01/wdt/models.

When the command succeeds, it should end with output like the following:

[INFO   ] Build successful. Build time=36s. Image tag=model-in-image:WLS-v1

Also, if you run the docker images command, then you will see an image named model-in-image:WLS-v1.

Note: If you have Kubernetes cluster worker nodes that are remote to your local machine, then you need to put the image in a location that these nodes can access. See Ensuring your Kubernetes cluster can access images.

Create WebLogic domain

In this section, you will deploy the new image to namespace sample-domain1-ns, including the following steps:

  • Create a namespace for the WebLogic domain.
  • Upgrade the operator to manage the WebLogic domain namespace.
  • Create a Secret containing your WebLogic administrator user name and password.
  • Create a Secret containing your Model in Image runtime encryption password:
    • All Model in Image domains must supply a runtime encryption Secret with a password value.
    • It is used to encrypt configuration that is passed around internally by the operator.
    • The value must be kept private but can be arbitrary; you can optionally supply a different secret value every time you restart the domain.
  • Deploy a Domain YAML file that references the new image.
  • Wait for the domain’s Pods to start and reach their ready state.
Namespace

Create a namespace that can host one or more domains:

$ kubectl create namespace sample-domain1-ns
## label the domain namespace so that the operator can autodetect and create WebLogic Server pods.
$ kubectl label namespace sample-domain1-ns weblogic-operator=enabled
Secrets

First, create the secrets needed by the WLS type model domain. In this case, you have two secrets.

Run the following kubectl commands to deploy the required secrets:

$ kubectl -n sample-domain1-ns create secret generic \
  sample-domain1-weblogic-credentials \
   --from-literal=username=weblogic --from-literal=password=welcome1
$ kubectl -n sample-domain1-ns label  secret \
  sample-domain1-weblogic-credentials \
  weblogic.domainUID=sample-domain1
$ kubectl -n sample-domain1-ns create secret generic \
  sample-domain1-runtime-encryption-secret \
   --from-literal=password=my_runtime_password
$ kubectl -n sample-domain1-ns label  secret \
  sample-domain1-runtime-encryption-secret \
  weblogic.domainUID=sample-domain1

Some important details about these secrets:

  • The WebLogic credentials secret:

    • It is required and must contain username and password fields.
    • It must be referenced by the spec.webLogicCredentialsSecret field in your Domain.
    • It also must be referenced by macros in the domainInfo.AdminUserName and domainInfo.AdminPassWord fields in your model YAML file.
  • The Model WDT runtime secret:

    • This is a special secret required by Model in Image.
    • It must contain a password field.
    • It must be referenced using the spec.model.runtimeEncryptionSecret field in its Domain.
    • It must remain the same for as long as the domain is deployed to Kubernetes but can be changed between deployments.
    • It is used to encrypt data as it’s internally passed using log files from the domain’s introspector job and on to its WebLogic Server pods.
  • Deleting and recreating the secrets:

    • You delete a secret before creating it, otherwise the create command will fail if the secret already exists.
    • This allows you to change the secret when using the kubectl create secret command.
  • You name and label secrets using their associated domain UID for two reasons:

    • To make it obvious which secrets belong to which domains.
    • To make it easier to clean up a domain. Typical cleanup scripts use the weblogic.domainUID label as a convenience for finding all resources associated with a domain.
Domain resource

Now, you create a Domain YAML file. A Domain is the key resource that tells the operator how to deploy a WebLogic domain.

Copy the following to a file called /tmp/mii-sample/mii-initial.yaml or similar, or use the file /tmp/mii-sample/domain-resources/WLS/mii-initial-d1-WLS-v1.yaml that is included in the sample source.

Click here to view the WLS Domain YAML file.

Note: Before you deploy the domain custom resource, determine if you have Kubernetes cluster worker nodes that are remote to your local machine. If so, you need to put the Domain’s image in a location that these nodes can access and you may also need to modify your Domain YAML file to reference the new location. See Ensuring your Kubernetes cluster can access images.

Run the following command to create the domain custom resource:

$ kubectl apply -f /tmp/mii-sample/domain-resources/WLS/mii-initial-d1-WLS-v1.yaml

Note: If you are choosing not to use the predefined Domain YAML file and instead created your own Domain YAML file earlier, then substitute your custom file name in the above command. Previously, we suggested naming it /tmp/mii-sample/mii-initial.yaml.

Verify the WebLogic Server pods are all running:

$ kubectl get all -n sample-domain1-ns
NAME                                 READY   STATUS    RESTARTS   AGE
pod/sample-domain1-admin-server      1/1     Running   0          41m
pod/sample-domain1-managed-server1   1/1     Running   0          40m
pod/sample-domain1-managed-server2   1/1     Running   0          40m

NAME                                       TYPE        CLUSTER-IP     EXTERNAL-IP   PORT(S)    AGE
service/sample-domain1-admin-server        ClusterIP   None           <none>        7001/TCP   41m
service/sample-domain1-cluster-cluster-1   ClusterIP   100.66.99.27   <none>        8001/TCP   40m
service/sample-domain1-managed-server1     ClusterIP   None           <none>        8001/TCP   40m
service/sample-domain1-managed-server2     ClusterIP   None           <none>        8001/TCP   40m

Invoke the web application

Create a load balancer to access the WebLogic Server Administration Console and applications deployed in the cluster. Tanzu supports the MetalLB load balancer and NGINX ingress for routing.

Install the MetalLB load balancer by running following commands:

## create namespace metallb-system
$ kubectl create ns metallb-system
## deploy MetalLB load balancer
$ kubectl apply -f https://raw.githubusercontent.com/google/metallb/v0.9.2/manifests/metallb.yaml -n metallb-system
## create secret
$ kubectl create secret generic -n metallb-system memberlist --from-literal=secretkey="$(openssl rand -base64 128)"
$ cat metallb-configmap.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  namespace: metallb-system
  name: config
data:
  config: |
    address-pools:
    - name: default
      protocol: layer2
      addresses:
      - 192.168.100.50-192.168.100.65    
$ kubectl apply -f metallb-configmap.yaml
configmap/config created
$ kubectl get all -n metallb-system
NAME                              READY   STATUS    RESTARTS   AGE
pod/controller-684f5d9b49-jkzfk   1/1     Running   0          2m14s
pod/speaker-b457r                 1/1     Running   0          2m14s
pod/speaker-bzmmj                 1/1     Running   0          2m14s
pod/speaker-gphh5                 1/1     Running   0          2m14s
pod/speaker-lktgc                 1/1     Running   0          2m14s

NAME                     DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR                 AGE
daemonset.apps/speaker   4         4         4       4            4           beta.kubernetes.io/os=linux   2m14s

NAME                         READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/controller   1/1     1            1           2m14s

NAME                                    DESIRED   CURRENT   READY   AGE
replicaset.apps/controller-684f5d9b49   1         1         1       2m14s

Install NGINX.

$ helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx --force-update
$ helm install ingress-nginx ingress-nginx/ingress-nginx

Create ingress for accessing the application deployed in the cluster and to access the Administration console.

$ cat ingress.yaml
apiVersion: networking.k8s.io/v1beta1
kind: Ingress
metadata:
  name: sample-nginx-ingress-pathrouting
  namespace: sample-domain1-ns
  annotations:
    kubernetes.io/ingress.class: nginx
spec:
  rules:
  - host:
    http:
      paths:
      - path: /
        backend:
          serviceName: sample-domain1-cluster-cluster-1
          servicePort: 8001
      - path: /console
        backend:
          serviceName: sample-domain1-admin-server
          servicePort: 7001
$ kubectl apply -f ingress.yaml

Verify ingress is running.

$ kubectl get ingresses -n sample-domain1-ns
NAME                               CLASS    HOSTS   ADDRESS          PORTS   AGE
sample-nginx-ingress-pathrouting   <none>   *       192.168.100.50   80      7m18s

Access the Administration Console using the load balancer IP address, http://192.168.100.50/console

Access the sample application.

# Access the sample application using the load balancer IP (192.168.100.50)
$ curl http://192.168.100.50/myapp_war/index.jsp
<html><body><pre>
*****************************************************************

Hello World! This is version 'v1' of the mii-sample JSP web-app.

Welcome to WebLogic Server 'managed-server1'!

 domain UID  = 'sample-domain1'
 domain name = 'domain1'

Found 1 local cluster runtime:
  Cluster 'cluster-1'

Found 0 local data sources:

*****************************************************************
</pre></body></html>
$ curl http://192.168.100.50/myapp_war/index.jsp
<html><body><pre>
*****************************************************************

Hello World! This is version 'v1' of the mii-sample JSP web-app.

Welcome to WebLogic Server 'managed-server2'!

 domain UID  = 'sample-domain1'
 domain name = 'domain1'

Found 1 local cluster runtime:
  Cluster 'cluster-1'

Found 0 local data sources:

*****************************************************************
</pre></body></html>