JBoss EAP on OpenShift

• JBoss EAP on OpenShift
  • Binary Build and Deployment
  • Configure Standalone Cluster

Binary Build and Deployment

• Clone verify cluster app

• Pacakge application (WAR) and copy at your local directory

  mvn clean package
  mkdir -p deployments
  cp target/verify-cluster.war deployments
  

• Create build config with EAP image stream and use local WAR file

  • Initial environment variables

    APP_NAME=verify-cluster
    PROJECT=$(oc project -q)
    IMAGE_STREAM=jboss-eap73-openshift:7.3
    REPLICAS=2
    VERSION=1.0.0
    BINARY_PATH=deployments/verify-cluster.war
    

  • Build EAP container image with deployed application

    • Create build config

      oc new-build --binary=true \
      --name=${APP_NAME} -l app=${APP_NAME} --image-stream=${IMAGE_STREAM}
      

    • Start build

      oc start-build ${APP_NAME} \
      --from-file=${BINARY_PATH} \
      --follow
      

    • Tag image with version

      oc tag ${APP_NAME}:latest ${APP_NAME}:${VERSION}
      

    • Check build config

      oc get buildconfig/verify-cluster
      

      output

      NAME             TYPE     FROM     LATEST
      verify-cluster   Source   Binary   1
      

• Create Deoloyment

  • Start build and deploy with oc new-app

    oc new-app ${APP_NAME}:${VERSION} \
    --labels=app=${APP_NAME},deploymentconfig=${APP_NAME},version=${VERSION},app.kubernetes.io/name=jboss
    

  • Check pods

    oc get pods -l app=$APP_NAME
    

    Output

    NAME                              READY   STATUS    RESTARTS   AGE
    verify-cluster-565fbb6f86-jt5xl   1/1     Running   0          4m42s
    

  • Check on Development Console

    Pods

    

    Resources

    

• Create route

  oc expose svc/${APP_NAME}
  echo "URL: http://$(oc get route ${APP_NAME} -n ${PROJECT} -o jsonpath='{.spec.host}')/verify-cluster"
  

• Test access verify-cluster and press Increment. Notice pod name and number of hits will be increased

  

Configure Standalone Cluster

• Service Account default needs view role to run KUBE_PING

  oc adm policy add-cluster-role-to-user \
  view system:serviceaccount:$PROJECT:default -n $PROJECT
  

• Pause rollout

  oc rollout pause deployment  ${APP_NAME}
  

• Configure JGROUP

  oc set env deployment ${APP_NAME} \
  JGROUPS_PING_PROTOCOL=kubernetes.KUBE_PING \
  KUBERNETES_NAMESPACE=${PROJECT} \
  KUBERNETES_LABELS=app=${APP_NAME},version=${VERSION}
  oc set env deployment ${APP_NAME} DISABLE_EMBEDDED_JMS_BROKER=true
  

• Scale replicas

  oc scale deployment ${APP_NAME} --replicas=${REPLICAS}
  

• Resume rollout and check pods

  oc rollout resume deployment  ${APP_NAME}
  oc get pods -l app=$APP_NAME,version=$VERSION
  

  Output

  NAME                              READY   STATUS        RESTARTS   AGE
  verify-cluster-75584cb799-rcf5t   1/1     Terminating   0          52s
  verify-cluster-789f5dd9d-qmg9s    1/1     Running       0          34s
  verify-cluster-789f5dd9d-s4q8z    1/1     Running       0          38s
  

  Check log

  oc logs  $(oc get pods -l app=$APP_NAME,version=$VERSION | tail -n 1 | awk '{print $1}')
  

• Test that sessions is replicated

  • Access verify-cluster app
  • Increment counter

  • Delete pod that response your request

    oc delete pods <pod name>
    

  • Refresh page again and check that counter still continue.

• binary_build.sh can be used to automated all steps.