Description of problem:
Change resource limits and requests for CEFK in ClusterLogging CR, the elasticsearch, fluentd and kibana pods are not redeployed, the curator cronjob is not updated. 

The elasticsearch CR changed, but the deploy isn't.

$ oc get clusterlogging -o yaml
apiVersion: v1
items:
- apiVersion: logging.openshift.io/v1
  kind: ClusterLogging
  metadata:
    creationTimestamp: 2019-04-03T06:54:15Z
    generation: 1
    name: instance
    namespace: openshift-logging
    resourceVersion: "183763"
    selfLink: /apis/logging.openshift.io/v1/namespaces/openshift-logging/clusterloggings/instance
    uid: 4ba58e72-55dd-11e9-ab3c-0aefbeae65e8
  spec:
    collection:
      logs:
        fluentd:
          resources:
            limits:
              memory: 1Gi
            requests:
              cpu: 200m
              memory: 1Gi
        rsyslog:
          resources: null
        type: fluentd
    curation:
      curator:
        resources:
          limits:
            memory: 200Mi
          requests:
            cpu: 200m
            memory: 200Mi
        schedule: 30 3 * * *
      type: curator
    logStore:
      elasticsearch:
        nodeCount: 2
        redundancyPolicy: SingleRedundancy
        resources:
          limits:
            cpu: "1"
            memory: 4Gi
          requests:
            cpu: 500m
            memory: 1300Mi
        storage:
          size: 10Gi
          storageClassName: gp2
      type: elasticsearch
    managementState: Managed
    visualization:
      kibana:
        proxy:
          resources:
            limits:
              memory: 100Mi
            requests:
              cpu: 100m
              memory: 100Mi
        replicas: 1
        resources:
          limits:
            memory: 1Gi
          requests:
            cpu: 500m
            memory: 1Gi
      type: kibana

$ oc get elasticsearch -o yaml |grep resources -A 6
      resources:
        limits:
          cpu: "1"
          memory: 4Gi
        requests:
          cpu: 500m
          memory: 1300Mi

$ oc get deploy elasticsearch-cdm-ivvgu8p3-1 -o yaml |grep resources -A 6
        resources:
          limits:
            cpu: "1"
            memory: 4Gi
          requests:
            cpu: 200m
            memory: 1Gi

$ oc get deploy kibana -o yaml |grep resources -A 5
        resources:
          limits:
            memory: 736Mi
          requests:
            cpu: 100m
            memory: 736Mi
--
        resources:
          limits:
            memory: 256Mi
          requests:
            cpu: 100m
            memory: 256Mi

$ oc get ds fluentd -o yaml |grep resources -A 5
        resources:
          limits:
            memory: 736Mi
          requests:
            cpu: 100m
            memory: 736Mi

$ oc get cj curator -o yaml |grep resources -A 5
            resources:
              limits:
                memory: 256Mi
              requests:
                cpu: 100m
                memory: 256Mi


Version-Release number of selected component (if applicable):
quay.io/openshift/origin-elasticsearch-operator@sha256:f3e56412389727015e80b01420304f4736c58be20410fe67b9e4e676ba7cfd4a
quay.io/openshift/origin-cluster-logging-operator@sha256:abd72897d24101f17a390be3eb5e3f7abc1f00eb8bb905ccca66515a6c904433


How reproducible:
Always

Steps to Reproduce:
1. Deploy logging use cr: https://raw.githubusercontent.com/QiaolingTang/scripts-and-files/master/cr.yaml
2. change resource limits and requests in clusterlogging CR
3.

Actual results:


Expected results:


Additional info: