Uploaded image for project: 'OpenShift Logging'
  1. OpenShift Logging
  2. LOG-2679

Elasticsearch operator selects invalid image for Kibana oauth-proxy if no internal registry is available

XMLWordPrintable

    • False
    • None
    • False
    • NEW
    • VERIFIED
    • Hide
      Before this update, the Elasticsearch Operator always used the internal image registry for the OAuth-Proxy which caused issued when running in a cluster without an internal image registry. With this update, the public image reference is used as a fallback when no internal registry is available.
      Show
      Before this update, the Elasticsearch Operator always used the internal image registry for the OAuth-Proxy which caused issued when running in a cluster without an internal image registry. With this update, the public image reference is used as a fallback when no internal registry is available.
    • Logging (LogExp) - Sprint 219, Logging (LogExp) - Sprint 220

      If there isn't an image-registry on the cluster, the dockerImageRepository is null in the imagestream/oauth-proxy [1]. The elasticsearch-operator can not provide a correct image to the deployment/kibana. the kabana-proxy image is given only the sha256 value[2]. that made the kibana pod into the status InvalidImageName[3]

      https://github.com/openshift/elasticsearch-operator/blob/master/internal/kibana/reconciler.go#L454

      [1]

       

       "status": {
        "dockerImageRepository": "",
        "tags": [
            {
                "items": [
                    
{                         "created": "2022-06-02T01:00:39Z",                         "dockerImageReference": "quay.io/openshift-release-dev/ocp-v4.0-art-dev@sha256:42ec80258937241130f3215d0395cf63e8e834dde96f5105a60cf9c0eaf5d046",                         "generation": 2,                         "image": "sha256:42ec80258937241130f3215d0395cf63e8e834dde96f5105a60cf9c0eaf5d046"                     }
                ],
                "tag": "v4.4"
            }
        ]
    }
 

       

      [2] "image": "@sha256:42ec80258937241130f3215d0395cf63e8e834dde96f5105a60cf9c0eaf5d046"

      [3]  kibana-5867bcc6ff-2rmvv                        1/2     InvalidImageName   0          83m

       

      Step to Reproduce:

      1. Deploy Cluster without embedded docker-registry. (docker-registry is not enabled on vsphere and baremetal cluster)
      2. Check the imagestream oauth-proxy. The status.dockerImageRepository is "".
      3. Deploy clusterlogging intance. Check the kibana pod status

      Actual Result:

      #oc describe pod kibana-5867bcc6ff-2rmvv

       

        Normal   Created                 84m                    kubelet            Created container kibana
  Normal   Started                 84m                    kubelet            Started container kibana
  Warning  Failed                  82m (x11 over 84m)     kubelet            Error: InvalidImageName
  Warning  InspectFailed           3m57s (x374 over 84m)  kubelet            Failed to apply default image tag "@sha256:42ec80258937241130f3215d0395cf63e8e834dde96f5105a60cf9c0eaf5d046": couldn't parse image reference "@sha256:42ec80258937241130f3215d0395cf63e8e834dde96f5105a60cf9c0eaf5d046": invalid reference format

       

      Expected result:

      The kibana pods can be started.

              rojacob@redhat.com Robert Jacob
              rhn-support-anli Anping Li
              Giriyamma Karagere Ramaswamy Giriyamma Karagere Ramaswamy (Inactive)
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

                Created:
                Updated:
                Resolved: