Uploaded image for project: 'Red Hat OpenShift AI Engineering'
  1. Red Hat OpenShift AI Engineering
  2. RHOAIENG-288

Recommended image version label for workbench is showed for two versions

XMLWordPrintable

    • 2
    • False
    • Hide

      None

      Show
      None
    • False
    • Release Notes
    • Known Issue
    • Done
    • 2
    • RHOAI IDE 2.5 Extended, RHOAI IDE 2.6

      Description of problem:

      For workbench notebook images, there is a Recommended label showed for the latest version of the notebook image. Unfortunately with the update from 1.33 to 2.4, this label is showed for two versions - the new 2023b and also for version 2023a, which was tagged with this label in 1.33 release. You can see it in this picture double-recommended.png.

      This issue is present only after the upgrade operation (1.33 to 2.3 ATM) and isn't present with the fresh installation.

      Root cause of the issue is right in the relevant imagestreams definitions, as I can see following for one of the image:

         - name: '1.2'
            annotations:
              opendatahub.io/notebook-python-dependencies: >-
                [{"name":"JupyterLab","version": "3.2"}, {"name":
                "Notebook","version": "6.4"}]
              opendatahub.io/notebook-software: '[{"name":"Python","version":"v3.8"}]'
              openshift.io/imported-from: quay.io/modh/odh-minimal-notebook-container
            from:
              kind: DockerImage
              name: >-
                quay.io/modh/odh-minimal-notebook-container@sha256:39068767eebdf3a127fe8857fbdaca0832cdfef69eed6ec3ff6ed1858029420f
            generation: 4
            importPolicy: {}
            referencePolicy:
              type: Source
          - name: '2023.1'
            annotations:
              opendatahub.io/default-image: 'true'
              opendatahub.io/notebook-python-dependencies: >-
                [{"name":"JupyterLab","version": "3.5"}, {"name":
                "Notebook","version": "6.5"}]
              opendatahub.io/notebook-software: '[{"name":"Python","version":"v3.9"}]'
              opendatahub.io/workbench-image-recommended: 'true'
              openshift.io/imported-from: quay.io/modh/odh-minimal-notebook-container
            from:
              kind: DockerImage
              name: >-
                quay.io/modh/odh-minimal-notebook-container@sha256:54a5af13057df3445f917494a8cc12a0bed081c67204e00c9ed26daeaf9d0079
            generation: 4
            importPolicy: {}
            referencePolicy:
              type: Source
          - name: '2023.2'
            annotations:
              opendatahub.io/default-image: 'true'
              opendatahub.io/notebook-python-dependencies: >-
                [{"name":"JupyterLab","version": "3.6"}, {"name":
                "Notebook","version": "6.5"}]
              opendatahub.io/notebook-software: '[{"name":"Python","version":"v3.9"}]'
              opendatahub.io/workbench-image-recommended: 'true'
              openshift.io/imported-from: quay.io/modh/odh-minimal-notebook-container
            from:
              kind: DockerImage
              name: >-
                quay.io/modh/odh-minimal-notebook-container@sha256:a942ab984181b1ea0bffec92473995bbb1488d65a4d1b64717385f49bfd0adf4
            generation: 4
            importPolicy: {}
            referencePolicy:
              type: Source

      Prerequisites (if any, like setup, operators/versions):

      Steps to Reproduce

      1. install RHODS 1.33
      2. upgrade to 2.4
      3. see the imagestreams definitions for workbench images or simply check the UI in JupyterHub or in DataScience project during the workbench creation

      Actual results:

      Double recommended label for two versions of the workbench image.

      Expected results:

      Only one recommended label for each workbench image.

      Reproducibility (Always/Intermittent/Only Once):

      As mentioned - always with the upgrade from 1.33 to 2.4

      Build Details:

      Workaround:

      Manually delete the tag from the imagestreams definitions.

      Additional info:

      As discussed, this is highly probably caused by the fact, that the imagestreams are patched only with the new data during the upgrade process, which results in keeping the recommended label for the originally recommended version and adding a new version with the same recommended label.

            hnalla Harshad Reddy Nalla
            jstourac@redhat.com Jan Stourac
            Jan Stourac
            RHOAI IDE
            Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

              Created:
              Updated:
              Resolved: