Avoid patching hypershift operator deployment

Let's try killing the hypershift operator to force it to
reload the ICSP/IDMS configuration on its startup.

Signed-off-by: Simone Tiraboschi <stirabos@redhat.com>

ci-operator/step-registry/hypershift/mce/kubevirt/create/disconnected/hypershift-mce-kubevirt-create-disconnected-chain.yaml

@@ -100,9 +100,9 @@ chain:
         PULL_SECRET_PATH="/tmp/.dockerconfigjson"
         TRUST="--additional-trust-bundle /etc/pki/ca-trust/source/anchors/registry.2.crt "
 
-        # TODO: can we avoid this?
-        ROVERRIDES=\$(oc get imagecontentsourcepolicy image-policy -o=json | jq -r '.spec.repositoryDigestMirrors[] | select (.source | test("registry.ci.openshift.org/ocp/") ) | .source+"="+.mirrors[0]')
-        oc patch deploy -n hypershift operator --type=json -p="[{\"op\": \"add\", \"path\": \"/spec/template/spec/containers/0/args/-\", \"value\": \"--registry-overrides=\${ROVERRIDES}\"}]"
+        # TODO: can we avoid this? is this a bug?
+        # kill the hypershift operator pod to force it reloading ICSP/IDMS from the management cluster
+        oc delete pod -n hypershift -l name=operator
         sleep 30
         ###