Description of problem:
Upgrade 4.7.10 to 4.8-fc3 succeeds but downgrade back to 4.7.10 fails.


Version-Release number of selected component (if applicable):
4.7.10


How reproducible:


Steps to Reproduce:
1 Start with 4.7.10. 
2 Upgrade to 4.8-fc3.
3 Downgrade back to 4.7.10 - fails.


Actual results:
Only etcd is at 4.7.10 and rest stays at 4.8-fc3 with this error:

./oc describe clusterversion -o yaml
info: An upgrade is in progress. Unable to apply 4.7.10: a required extension is not available to update
    - lastTransitionTime: "2021-05-17T21:01:37Z"
      message: 'Could not update flowschema "openshift-etcd-operator" (73 of 669):
        the server does not recognize this resource, check extension API servers'
      reason: UpdatePayloadResourceTypeMissing
      status: "True"
      type: Failing



Expected results:
Downgrade back to original version should work (has worked in the past).


Additional info:
must-gather says
ClusterID: 6c1c42fb-cb6b-4f25-9154-cddf0370b0d1
ClusterVersion: Updating to "4.7.10" from "4.8.0-fc.3" for 2 days: Working towards 4.7.10: 72 of 669 done (10% complete)
ClusterOperators:
	All healthy and stable

must-gather is too big to attach. Have it on google drive and can share