Uploaded image for project: 'JBoss Enterprise Application Platform'
  1. JBoss Enterprise Application Platform
  2. JBEAP-24439

EAP Operator scaling down shutdowns two pods instead of one

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Major Major
    • OP-3.0.0.GA
    • OP-2.4.0.GA
    • OpenShift, Operator
    • None
    • False
    • None
    • False
    • Hide

      1. Install the EAP Operator in a cluster.
      2. Build the image, for example by using helm charts:

      $ cat <<EOF > /tmp/helm-chart-build.yaml
image:
  tag: latest
build:
  enabled: true
  mode: s2i
  uri: 'https://github.com/jboss-developer/jboss-eap-quickstarts.git'
  ref: xp-4.0.x
  contextDir: microprofile-config
  output:
    kind: ImageStreamTag
  env:
    - name: MAVEN_ARGS_APPEND
      value: '-Dcom.redhat.xpaas.repo.jbossorg'
  triggers: {}
  s2i:
    version: latest
    arch: amd64
    jdk: '11'
    amd64:
      jdk11:
        builderImage: registry.redhat.io/jboss-eap-7/eap-xp4-openjdk11-openshift-rhel8
        runtimeImage: registry.redhat.io/jboss-eap-7/eap-xp4-openjdk11-runtime-openshift-rhel8
deploy:
  enabled: false
EOF

$ helm install microprofile-config-app \
    -f /tmp/helm-chart-build.yaml \
    jboss-eap/eap74

      3. Deploy the application with persistent storage.

      cat <<EOF | oc create -f -
apiVersion: wildfly.org/v1alpha1
kind: WildFlyServer
metadata:
  name: microprofile-config-operator-app
spec:
 applicationImage: 'microprofile-config-app:latest'
 replicas: 3
 storage:
    volumeClaimTemplate:
      spec:
        resources:
          requests:
            storage: 10Mi
EOF

      4. Scale down the replicas from 3 to 2

      Show
      1. Install the EAP Operator in a cluster. 2. Build the image, for example by using helm charts: $ cat <<EOF > /tmp/helm-chart-build.yaml image: tag: latest build: enabled: true mode: s2i uri: 'https: //github.com/jboss-developer/jboss-eap-quickstarts.git' ref: xp-4.0.x contextDir: microprofile-config output: kind: ImageStreamTag env: - name: MAVEN_ARGS_APPEND value: '-Dcom.redhat.xpaas.repo.jbossorg' triggers: {} s2i: version: latest arch: amd64 jdk: '11' amd64: jdk11: builderImage: registry.redhat.io/jboss-eap-7/eap-xp4-openjdk11-openshift-rhel8 runtimeImage: registry.redhat.io/jboss-eap-7/eap-xp4-openjdk11-runtime-openshift-rhel8 deploy: enabled: false EOF $ helm install microprofile-config-app \ -f /tmp/helm-chart-build.yaml \ jboss-eap/eap74 3. Deploy the application with persistent storage. cat <<EOF | oc create -f - apiVersion: wildfly.org/v1alpha1 kind: WildFlyServer metadata: name: microprofile-config- operator -app spec: applicationImage: 'microprofile-config-app:latest' replicas: 3 storage: volumeClaimTemplate: spec: resources: requests: storage: 10Mi EOF 4. Scale down the replicas from 3 to 2

      Reviewing the scaling down process done by the EAP Operator when the TX recovery facility is enabled, I've found that when the CRD is scaled down in one replica, two pods are stopped instead of one:

      This is the pod event sequence shown after scaling down from 3 replicas to 2:

      $ oc get pods -w
NAME                                              READY   STATUS      RESTARTS   AGE
eap-operator-8557b68cb-wg2r7                      1/1     Running     0          25m
microprofile-config-app-2-build                   0/1     Completed   0          16m
microprofile-config-app-build-artifacts-1-build   0/1     Completed   0          19m
microprofile-config-operator-app-0                1/1     Running     0          3m50s
microprofile-config-operator-app-1                1/1     Running     0          4m12s
microprofile-config-operator-app-2                1/1     Running     0          4m33s


microprofile-config-operator-app-2                1/1     Running     0          5m27s
microprofile-config-operator-app-2                1/1     Running     0          5m57s
microprofile-config-operator-app-2                1/1     Running     0          6m6s
microprofile-config-operator-app-2                1/1     Terminating   0          6m20s
microprofile-config-operator-app-1                1/1     Terminating   0          5m59s
microprofile-config-operator-app-1                0/1     Terminating   0          6m5s
microprofile-config-operator-app-1                0/1     Terminating   0          6m5s
microprofile-config-operator-app-1                0/1     Terminating   0          6m5s
microprofile-config-operator-app-2                0/1     Terminating   0          6m26s
microprofile-config-operator-app-1                0/1     Pending       0          0s
microprofile-config-operator-app-2                0/1     Terminating   0          6m26s
microprofile-config-operator-app-1                0/1     Pending       0          0s
microprofile-config-operator-app-2                0/1     Terminating   0          6m26s
microprofile-config-operator-app-1                0/1     ContainerCreating   0          0s
microprofile-config-operator-app-1                0/1     ContainerCreating   0          2s
microprofile-config-operator-app-1                0/1     Running             0          3s
microprofile-config-operator-app-1                1/1     Running             0          20s

      In the above sequence, microprofile-config-operator-app-1 and microprofile-config-operator-app-2 were stopped, however, I scaled down only one pod.

      I've checked this using EAP 7.4 images, EAP 8 images are not at this moment supported by the Operator with the transaction recover enabled.

              yborgess1@redhat.com Yeray Borges Santana
              yborgess1@redhat.com Yeray Borges Santana
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

                Created:
                Updated:
                Resolved: