Bug 2008356

Description of problem (please be detailed as possible and provide log
snippests):
Upgrading osus operator from v4.6.0 to v4.9.0 failed. After changing the source of subscription, no upgrade happen as expected. Checking v4.9.0 csv to find that there is not upgrade related info in v4.9.0 csv which can be used for path detect.

# cat manifests/update-service-operator.clusterserviceversion.yaml|grep -E "replace|skip"|wc -l
0

# yaml2json manifests/update-service-operator.clusterserviceversion.yaml|jq .metadata
{
  "annotations": {
    "alm-examples": "[\n  {\n    \"apiVersion\": \"updateservice.operator.openshift.io/v1\",\n    \"kind\": \"UpdateService\",\n    \"metadata\": {\n      \"name\": \"sample\"\n    },\n    \"spec\": {\n      \"replicas\": 1\n    }\n  }\n]",
    "capabilities": "Basic Install",
    "description": "Creates and maintains an OpenShift Update Service instance",
    "operatorframework.io/suggested-namespace": "openshift-update-service",
    "operators.operatorframework.io/builder": "operator-sdk-v1.9.0+git",
    "operators.operatorframework.io/project_layout": "go.kubebuilder.io/v3"
  },
  "name": "update-service-operator.v4.9.0",
  "namespace": "placeholder"
}

Before triggering upgrade:
# ./oc get sub cincinnati-operator -n openshift-update-service -ojson|jq .spec
{
  "channel": "v1",
  "installPlanApproval": "Manual",
  "name": "cincinnati-operator",
  "source": "redhat-operators",
  "sourceNamespace": "openshift-marketplace",
  "startingCSV": "update-service-operator.v4.6.0"
}
# ./oc get sub cincinnati-operator -n openshift-update-service -ojson|jq .status.installedCSV
"update-service-operator.v4.6.0"
# ./oc get sub cincinnati-operator -n openshift-update-service -ojson|jq .status.currentCSV
"update-service-operator.v4.6.0"
# ./oc get ip
NAME            CSV                              APPROVAL   APPROVED
install-dx8f2   update-service-operator.v4.6.0   Manual     true
# ./oc get csv
NAME                             DISPLAY                    VERSION   REPLACES   PHASE
update-service-operator.v4.6.0   OpenShift Update Service   4.6.0                Succeeded

After changing the source of subsription:
# ./oc get sub cincinnati-operator -n openshift-update-service -ojson|jq .spec
{
  "channel": "v1",
  "installPlanApproval": "Manual",
  "name": "cincinnati-operator",
  "source": "qe-app-registry",
  "sourceNamespace": "openshift-marketplace",
  "startingCSV": "update-service-operator.v4.6.0"
}
# ./oc get sub -n openshift-update-service
NAME                  PACKAGE               SOURCE            CHANNEL
cincinnati-operator   cincinnati-operator   qe-app-registry   v1

# ./oc get csv -n openshift-update-service
NAME                             DISPLAY                    VERSION   REPLACES   PHASE
update-service-operator.v4.6.0   OpenShift Update Service   4.6.0                Succeeded

# ./oc get ip -n openshift-update-service
NAME            CSV                              APPROVAL   APPROVED
install-dx8f2   update-service-operator.v4.6.0   Manual     true

Confirmed with OLM qe that we should have replace/skipRange info defined in v4.9.0 csv so that OLM can detect the upgrade path from it.


Version of all relevant components (if applicable):
cincinnati-operator-bundle-container-v4.9.0-1
# ./oc get clusterversion
NAME      VERSION   AVAILABLE   PROGRESSING   SINCE   STATUS
version   4.8.13    True        False         27h     Cluster version is 4.8.13

Does this issue impact your ability to continue to work with the product
(please explain in detail what is the user impact)?
No upgrade test can be done now.

Is there any workaround available to the best of your knowledge?


Rate from 1 - 5 the complexity of the scenario you performed that caused this
bug (1 - very simple, 5 - very complex)?


Can this issue reproducible?
always

Can this issue reproduce from the UI?


If this is a regression, please provide more details to justify this:
no

Steps to Reproduce:
1. Add latest bundle image v4.9.0 to qe pre-release test index image(quay.io/openshift-qe-optional-operators/ocp4-index:latest)
2. Install osus 4.6.0(from default redhat-operators source) on ocp v4.8
3. Edit subscription to change the `source` to qe-app-registry source(this is a catalogsource which pointed to qe pre-release index image)


Actual results:
No upgrade happen.

Expected results:
The osus operator upgrade should work.

Additional info:
https://docs.openshift.com/container-platform/4.8/operators/understanding/olm/olm-workflow.html