Some users have reported a problem during installations and upgrades where the CoreDNS image can not be retrieved. The error may resemble the following:
FATA[0000] pulling image: rpc error: code = NotFound desc = failed to pull and unpack image "k8s.gcr.io/coredns:v1.9.3": failed to resolve reference "k8s.gcr.io/coredns:v1.9.3": k8s.gcr.io/coredns:v1.9.3: not found
Normally, kubeadm manages the image tags for the assets it needs to retrieve. However, if you encounter this issue, you can instead find the image at k8s.gcr.io/coredns/coredns:v1.9.3.
The workaround, if you encounter this issue, is to manually pull the image on each node in your cluster and re-tag it:
crictl pull k8s.gcr.io/coredns/coredns:v1.9.3
ctr --namespace=k8s.io image tag k8s.gcr.io/coredns/coredns:v1.9.3 k8s.gcr.io/coredns:v1.9.3