diff --git a/enterprise/server/raft/store/store.go b/enterprise/server/raft/store/store.go
index 40d7b5a4f72..0a879ac34b8 100644
--- a/enterprise/server/raft/store/store.go
+++ b/enterprise/server/raft/store/store.go
@@ -1117,12 +1117,36 @@ func (s *Store) removeAndStopReplica(ctx context.Context, rd *rfpb.RangeDescript
 	return nil
 }
 
+func (s *Store) syncRemoveData(ctx context.Context, rangeID, replicaID uint64) error {
+	err := client.RunNodehostFn(ctx, func(ctx context.Context) error {
+		err := s.nodeHost.SyncRemoveData(ctx, rangeID, replicaID)
+		// If the shard is not stopped, we want to retry SyncRemoveData call.
+		if err == dragonboat.ErrShardNotStopped {
+			err = dragonboat.ErrTimeout
+		}
+		return err
+	})
+	if err != nil {
+		return status.InternalErrorf("failed to remove data of c%dn%d from raft: %s", rangeID, replicaID, err)
+	}
+	return nil
+}
+
 // RemoveData tries to remove all data associated with the specified node (shard, replica). It waits for the node (shard, replica) to be fully offloaded or the context is cancelled. This method should only be used after the node is deleted from its Raft cluster.
 func (s *Store) RemoveData(ctx context.Context, req *rfpb.RemoveDataRequest) (*rfpb.RemoveDataResponse, error) {
+	if req.GetRangeId() != 0 {
+		if err := s.syncRemoveData(ctx, req.GetRangeId(), req.GetReplicaId()); err != nil {
+			return nil, err
+		}
+		return &rfpb.RemoveDataResponse{}, nil
+	}
 	rd := req.GetRange()
+	if rd == nil {
+		return nil, status.InvalidArgumentErrorf("need to specify either range_id or range")
+	}
 	remoteRD, err := s.Sender().LookupRangeDescriptor(ctx, rd.GetStart(), true /*skip Cache */)
 	if err != nil {
-		return nil, status.InternalErrorf("failed to look up range descriptor")
+		return nil, status.InternalErrorf("failed to look up range descriptor: %s", err)
 	}
 
 	if rd.GetRangeId() != remoteRD.GetRangeId() {
@@ -1144,21 +1168,11 @@ func (s *Store) RemoveData(ctx context.Context, req *rfpb.RemoveDataRequest) (*r
 		// This should not happen because we don't allow a range to be split while there are removals in progress.
 		if !bytes.Equal(remoteRD.GetStart(), rd.GetStart()) || !bytes.Equal(remoteRD.GetEnd(), rd.GetEnd()) {
 			err := status.InternalErrorf("range descriptor's range changed from [%q, %q) (gen: %d) to [%q, %q) (gen: %d) while there are replicas in process of removal", rd.GetStart(), rd.GetEnd(), rd.GetGeneration(), remoteRD.GetStart(), remoteRD.GetEnd(), remoteRD.GetGeneration())
-			s.log.Errorf("%s", err)
 			return nil, err
 		}
 	}
-
-	err = client.RunNodehostFn(ctx, func(ctx context.Context) error {
-		err := s.nodeHost.SyncRemoveData(ctx, rd.GetRangeId(), req.GetReplicaId())
-		// If the shard is not stopped, we want to retry SyncRemoveData call.
-		if err == dragonboat.ErrShardNotStopped {
-			err = dragonboat.ErrTimeout
-		}
-		return err
-	})
-	if err != nil {
-		return nil, status.InternalErrorf("failed to remove data of c%dn%d from raft: %s", rd.GetRangeId(), req.GetReplicaId(), err)
+	if err := s.syncRemoveData(ctx, req.GetRangeId(), req.GetReplicaId()); err != nil {
+		return nil, err
 	}
 
 	if shouldDeleteRange {
@@ -1167,7 +1181,7 @@ func (s *Store) RemoveData(ctx context.Context, req *rfpb.RemoveDataRequest) (*r
 			return nil, err
 		}
 		defer db.Close()
-		if err := db.DeleteRange(rd.GetStart(), rd.GetEnd(), pebble.NoSync); err != nil {
+		if err = db.DeleteRange(remoteRD.GetStart(), remoteRD.GetEnd(), pebble.NoSync); err != nil {
 			return nil, status.InternalErrorf("failed to delete data of c%dn%d from pebble: %s", req.GetRange().GetRangeId(), req.GetReplicaId(), err)
 		}
 
@@ -1178,7 +1192,7 @@ func (s *Store) RemoveData(ctx context.Context, req *rfpb.RemoveDataRequest) (*r
 		}
 	}
 
-	//update range descriptor: remove the replica descriptor from the removed list.
+	// update range descriptor: remove the replica descriptor from the removed list.
 	if markedForRemoval {
 		rd, err = s.removeReplicaFromRangeDescriptor(ctx, remoteRD.GetRangeId(), req.GetReplicaId(), remoteRD)
 		if err != nil {