Description of problem:

When deploy GPU Operator in disconnected cluster after mirror olm catalog. it will threw the error:

error using catalog community-operators (in namespace openshift-marketplace): failed to list bundles: rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp 172.30.197.6:50051: connect: no route to host"

But we don't deploy GPU operator using community-operator hub
Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:
1. Prune Index Image for GPU Operator
export REGISTRY_AUTH_FILE=~/pull-secret.json
export VERSION=4.9
./opm index prune -f registry.redhat.io/redhat/certified-operator-index:v${VERSION} -p gpu-operator-certified -t quay.io/openshitf-psap-qe/certified-operator-index:v${VERSION}

podman push quay.io/openshift-psap-qe/certified-operator-index:v${VERSION}

2. To Mirror GPU Operator catalog
oc adm catalog mirror --insecure=true --index-filter-by-os='linux/amd64' quay.io/openshift-psap-qe/certified-operator-index:v${VERSION} ${JUMP_HOST}:5000/redhat -a ${REGISTRY_AUTH_FILE}

cd manifests-certified-operator-index-1643091511
oc create -f imageContentSourcePolicy.yaml
oc image mirror -f mapping.txt -a ${REGISTRY_AUTH_FILE} --insecure

3. Create CatalogSource for Certified Operators

1. catalogSource.yaml contains the case, can be extended with displayName, publisher, ...
   oc create f <<EOF
   apiVersion: operators.coreos.com/v1alpha1
   kind: CatalogSource
   metadata:
   name: certified-operator-index
   namespace: openshift-marketplace
   spec:
   image: ${JUMP_HOST}:5000/redhat/openshift-psap-qe-certified-operator-index:v${VERSION}
   sourceType: grpc
   displayName: Openshift PSAP QE Certified
   publisher: Openshift QE
   updateStrategy:
   registryPoll:
   interval: 30m
   EOF

Actual results:

Expected results:

Additional info: