Description of problem:
There are three errors in CRD elasticsearch.status.  The status in conditions, nodes are wrong. the podName should be "" once the pod is deleted (replicas=0 in my test).

Version-Release number of selected component (if applicable):
docker.io/openshift/origin-elasticsearch-operator:v4.0
imageID: docker.io/openshift/origin-elasticsearch-operator@sha256:01ef458b402117811b2aee172f3ce104c700bbadabc7e36147561ef355ee1e84


How reproducible:
always

Steps to Reproduce:
1. Deploy elasticsearch by operator
2. Check the status.conditions
3. Change the elasticsearch master replicas=0
4. check the status.nodes

status:
  clusterHealth: ""
  conditions:
  - lastTransitionTime: 2018-12-20T06:49:05Z
    status: "False"
    type: ScalingUp
  - lastTransitionTime: 2018-12-20T06:49:05Z
    message: Config Map is up to date
    reason: ConfigChange
    status: "False"
    type: UpdatingSettings
  - lastTransitionTime: 2018-12-20T05:33:43Z
    status: "False"
    type: ScalingDown
  - lastTransitionTime: 2018-12-20T05:33:43Z
    status: "False"
    type: Restarting
  nodes:
  - deploymentName: elasticsearch-clientdatamaster-0-1
    podName: elasticsearch-clientdatamaster-0-1-84d764899d-bh7jl
    replicaSetName: elasticsearch-clientdatamaster-0-1-84d764899d
    roles:
    - client
    - data
    - master
    status: Running
    upgradeStatus:
      underUpgrade: "False"
  - deploymentName: elasticsearch-data-1-1
    podName: elasticsearch-data-1-1-77ffddbf7b-zdd76
    replicaSetName: elasticsearch-data-1-1-77ffddbf7b
    roles:
    - data
    status: Running
    upgradeStatus:
      underUpgrade: "False"
  - podName: elasticsearch-client-2-1-0
    roles:
    - client
    statefulSetName: elasticsearch-client-2-1
    status: Running
    upgradeStatus:
      underUpgrade: "False"
  pods:
    client:
      failed: []
      notReady:
      - elasticsearch-client-1-1-0
      - elasticsearch-client-2-1-0
      ready: []
    data:
      failed: []
      notReady:
      - elasticsearch-data-1-1-77ffddbf7b-zdd76
      ready: []
    master:
      failed: []
      notReady: []
      ready: []
  shardAllocationEnabled: "True"

Expected results:
1) in .status.conditions, the actions ScalingUp/ConfigChange/ScalingDown succeed, the status true.
2) in .status.nodes.deploymentName=elasticsearch-clientdatamaster-0-1.  As the replicas=0, the status should be not running. The podName should be "".