Image pull fails with "error rewriting reference: repository name must be canonical"

Issue

• When the OpenShift Container Platform Worker Nodes tries to pull a container image, this fails with the following error message:

  Type     Reason          Age                   From               Message
  ----     ------          ----                  ----               -------
  Normal   Scheduled       20m                   default-scheduler  Successfully assigned openshift-console/console-f867f65d6-rw4mx to ocp-example-cdwpk-master-2
  Normal   AddedInterface  20m                   multus             Add eth0 [10.132.0.87/23]
  Normal   Pulling         18m (x4 over 20m)     kubelet            Pulling image "quay.io/openshift-release-dev/ocp-v4.0-art-dev@sha256:a920f9aea6ac44b409ed427612e5538f36e7bb1a22d7f7014af9a3e6f4c55e04"
  Warning  Failed          18m (x4 over 20m)     kubelet            Failed to pull image "quay.io/openshift-release-dev/ocp-v4.0-art-dev@sha256:a920f9aea6ac44b409ed427612e5538f36e7bb1a22d7f7014af9a3e6f4c55e04": rpc error: code = Unknown desc = error rewriting reference: repository name must be canonical
  Warning  Failed          18m (x4 over 20m)     kubelet            Error: ErrImagePull
  Normal   BackOff         9m54s (x44 over 20m)  kubelet            Back-off pulling image "quay.io/openshift-release-dev/ocp-v4.0-art-dev@sha256:a920f9aea6ac44b409ed427612e5538f36e7bb1a22d7f7014af9a3e6f4c55e04"
  Warning  Failed          4m58s (x66 over 20m)  kubelet            Error: ImagePullBackOff
  

• The error message error rewriting reference: repository name must be canonical is shown and the image is not being pulled.