attempt delete marker replication after object is replicated (#15857)

Ensure delete marker replication success, especially since the
recent optimizations to heal on HEAD, LIST and GET can force
replication attempts on delete marker before underlying object
version could have synced.

cmd/object-handlers.go

@@ -740,6 +740,15 @@ func (api objectAPIHandlers) headObjectHandler(ctx context.Context, objectAPI Ob
 			}
 			QueueReplicationHeal(ctx, bucket, objInfo)
 		}
+		// do an additional verification whether object exists when opts.DeleteMarker is set by source
+		// cluster as part of delete marker replication
+		if opts.DeleteMarker && opts.ProxyHeaderSet {
+			opts.VersionID = ""
+			goi, gerr := getObjectInfo(ctx, bucket, object, opts)
+			if gerr == nil || goi.VersionID != "" { // object layer returned more info because object is deleted
+				w.Header().Set(xhttp.MinIOTargetReplicationReady, "true")
+			}
+		}
 		writeErrorResponseHeadersOnly(w, toAPIError(ctx, err))
 		return
 	}