VMs, cloned from golden images (which also use dataVolumeTemplates) also fail creation.

VM from golden images:

apiVersion: kubevirt.io/v1alpha3
kind: VirtualMachine
metadata:
  labels:
    app: eeeee
    vm.kubevirt.io/template: rhel8-desktop-tiny-v0.11.3
    vm.kubevirt.io/template.revision: '1'
    vm.kubevirt.io/template.version: v0.12.2
    os.template.kubevirt.io/rhel8.3: 'true'
    flavor.template.kubevirt.io/tiny: 'true'
    workload.template.kubevirt.io/desktop: 'true'
    vm.kubevirt.io/template.namespace: openshift
  name: eeeee
  namespace: openshift-virtualization-os-images
  annotations:
    name.os.template.kubevirt.io/rhel8.3: Red Hat Enterprise Linux 8.0 or higher
spec:
  dataVolumeTemplates:
    - apiVersion: cdi.kubevirt.io/v1alpha1
      kind: DataVolume
      metadata:
        name: eeeee-rootdisk-cj2nh
      spec:
        pvc:
          accessModes:
            - ReadWriteOnce
          resources:
            requests:
              storage: 20Gi
          volumeMode: Filesystem
        source:
          pvc:
            name: rhel8
            namespace: openshift-virtualization-os-images
  running: false
  template:
    metadata:
      labels:
        kubevirt.io/domain: eeeee
        kubevirt.io/size: tiny
        vm.kubevirt.io/name: eeeee
        os.template.kubevirt.io/rhel8.3: 'true'
        flavor.template.kubevirt.io/tiny: 'true'
        workload.template.kubevirt.io/desktop: 'true'
    spec:
      domain:
        cpu:
          cores: 1
          sockets: 1
          threads: 1
        devices:
          disks:
            - disk:
                bus: virtio
              name: cloudinitdisk
            - bootOrder: 1
              disk:
                bus: virtio
              name: rootdisk
          inputs:
            - bus: virtio
              name: tablet
              type: tablet
          interfaces:
            - masquerade: {}
              model: virtio
              name: nic-0
          networkInterfaceMultiqueue: true
          rng: {}
        machine:
          type: pc-q35-rhel8.2.0
        resources:
          requests:
            memory: 1.5Gi
      evictionStrategy: LiveMigrate
      hostname: eeeee
      networks:
        - name: nic-0
          pod: {}
      terminationGracePeriodSeconds: 180
      volumes:
        - cloudInitNoCloud:
            userData: |
              #cloud-config
              user: cloud-user
              password: gbjz-9420-yqkb
              chpasswd:
                expire: false
          name: cloudinitdisk
        - dataVolume:
            name: eeeee-rootdisk-cj2nh
          name: rootdisk

This regression has been caused by https://github.com/kubevirt/kubevirt/pull/4261

I understand it was an API clean-up and we were not supposed to be using it, it was part of the api nonetheless and we were using it. Now we have passed code freeze and we can not do this cleanup on the UI side easily (maybe we could get an exception but I would love to avoid risking regressions by doing this type of changes now). I would say the https://github.com/kubevirt/kubevirt/pull/4261 should be reverted and re-introduced for 2.6 so we can properly adapt in 4.7.

>  I understand it was an API clean-up and we were not supposed to be using it, it was part of the api nonetheless and we were using it. Now we have passed code freeze and we can not do this cleanup on the UI side easily (maybe we could get an exception but I would love to avoid risking regressions by doing this type of changes now). I would say the https://github.com/kubevirt/kubevirt/pull/4261 should be reverted and re-introduced for 2.6 so we can properly adapt in 4.7.

here's the background. The DataVolumeTemplates section exposed two structs that should never have been captured by our api to begin with, the api group/version and status structs. If a user attempts to set either of these structs, they do nothing. They are essentially no-ops, which means we're exposing parts of our API that we don't actually support. That's what PR 4261 addressed by removing the unused fields.

That PR obvious causes issues, beyond the one that's captured in this BZ even. We already had to re-introduce a dummy status struct to maintain compatibility. We can do the same with the api group/version. We'll just need to clearly state that the DataVolumeTemplate group/version fields are no-ops that exist to maintain compatibility. 

So I wouldn't revert pr4261, instead i'd add the group/version field back. There are still changes with pr 4261 around how dataVolume status is documented/supported that we'd like to keep.

https://github.com/kubevirt/kubevirt/pull/4319

(In reply to David Vossel from comment #7)
> https://github.com/kubevirt/kubevirt/pull/4319

Thank you!
I have opened a followup (https://bugzilla.redhat.com/show_bug.cgi?id=1885319) to remove the use of it from the ui code.

*** Bug 1886927 has been marked as a duplicate of this bug. ***

Created attachment 1721038 [details]
review_and_confirm_details

Created attachment 1721039 [details]
successfully created virtual machine

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (OpenShift Virtualization 2.5.0 Images), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHEA-2020:5127