storagenode/retain: don't quit on error

It has been noted in the forum that, during a Retain operation, when a
piece can't be deleted, the process never completes. The error is
written to the log, but the completion line "Moved pieces to trash
during retain" never is.

This `return` line is the reason. We should instead continue the loop.

Change-Id: I0f51d34aba0e81ad60a75802069b42dc135ad907
Refs: #6482

storagenode/retain/retain.go

@@ -343,7 +343,7 @@ func (s *Service) retainPieces(ctx context.Context, req Request) (err error) {
 					zap.Stringer("Satellite ID", satelliteID),
 					zap.Stringer("Piece ID", pieceID),
 					zap.Error(err))
-				return nil
+				continue
 			}
 		}
 		numDeleted++