In the Supervisor cluster, the system pod imageregistry-operator may not be affected by this error if the existing imageregistry-operator pod is still running.
However, when viewing and describing the corresponding PKGi, it is found to be in ReconcileFailed state with the below error message:
kubectl get pkgi -A
vmware-system-imageregistry packageinstall.packaging.carvel.dev/imageregistry-operator vmimages.vmware.com #.#.#+vmware.# Reconcile failed: Error (see .status.usefulErrorMessage for details)
kubectl describe pkgi -n vmware-system-imageregistry imageregistry-operator
Fetching image:
GET
https://localhost:5000/v2/vmware/imageregistry-operator/package/manifests/sha256:<omitted>:
MANIFEST_UNKNOWN: manifest unknown;
map[Name:vmware/imageregistry-operator/package
Revision:sha256:<omitted>]
Viewing the logs for the docker-registry system pod shows errors similar to the following referencing the above missing manifest:
kubectl logs -n kube-system <docker-registry-pod>
level=error msg="response completed with error" err.code="manifest unknown" err.detail="unknown manifest name=vmware/imageregistry-operator/package revision=sha256:<omitted>" err.message="manifest unknown" go.version="go1.22.5 X:boringcrypto" http.request.host="127.0.0.1:5002" http.request.id=<omitted> http.request.method=GET http.request.remoteaddr="<localhost>:<port>" http.request.uri="/v2/vmware/imageregistry-operator/package/manifests/sha256:<omitted>" http.request.useragent="go-containerregistry/v0.16.1" http.response.contenttype="application/json; charset=utf-8" http.response.duration=1.074192ms http.response.status=404 http.response.written=213 vars.name="vmware/imageregistry-operator/package" vars.reference="sha256:<omitted>"
vSphere Supervisor
The manifest has gone missing from the Supervisor cluster.
Reach out to VMware by Broadcom Technical Support to validate and perform steps to resolve this issue.