MySQL cluster, failed data node automatically restarted with --initial (no replies)

We have a MySQL cluster with two data nodes, and yesterday the datanode one failed for about 20 minutes, and after it was back on-line, it was restarted like nbdb --initial, so, all the data files were deleted and re_created from the data node 2, but, before this synchronization was done, data node failed, do, only partial data were populated into data node1, we lost several tables.

My question is, is this normal behavior of datanode's reboot from failure? Should all the data files be deleted after restarting? Where can I change the reboot command so that it will start without the --initial option?

thanks