OCPBUGS#19789: 4.14.2 OLM bug fix update for OLM Operator groups

release_notes/ocp-4-14-release-notes.adoc

@@ -3308,6 +3308,30 @@ $ oc adm release info 4.14.2 --pullspecs
 
 * Previously, when MetalLB Operator's controller restarted while having an IP assigned and unassigned load balancer services, it restarted with an empty internal state, which can break workloads. With this release, MetalLB's controller is modified to first process the services that already have an assigned IP. (link:https://issues.redhat.com/browse/OCPBUGS-16267[*OCPBUGS-16267*])
 
+* Previously, if you created an Operator group with same name as a cluster role used by {product-title} or Kubernetes resources, Operator Lifecycle Manager (OLM) would overwrite the cluster roles. With this fix, if you create an Operator group that conflicts with a cluster role used by {product-title} or Kubernetes, OLM generates a unique cluster role name using the following syntax:
++
+.Naming syntax
+[source,text]
+----
+olm.og.<operator_group_name>.<admin_edit_or_view>-<hash_value>
+----
++
+OLM generates unique names only for Operator groups that conflict with a set of defined cluster roles used by {product-title} and Kubernetes. You must ensure that the name of your Operator group does not conflict with other cluster roles that exist on the cluster.
++
+OLM generates unique names for Operator groups that conflict with the following cluster roles:
+
+** `aggregate-olm`
+** `alert-routing`
+** `cluster`
+** `cluster-monitoring`
+** `monitoring`
+** `monitoring-rules`
+** `packagemanifests-v1`
+** `registry`
+** `storage`
++
+(link:https://issues.redhat.com/browse/OCPBUGS-19789[*OCPBUGS-19789*])
+
 [id="ocp-4-14-2-known-issue"]
 ==== Known issue