Mason Sharp at StormDB reported the following issue, which should be fixed.
I added one single data node slave out of 4 masters and tried to restart pgxc_ctl. I get a core dump though.
ERROR: Number of elements in datanode slave definitions are different datanodeNames and datanodeSlaveServers. Check your configuration
ERROR: Number of elements in datanode slave definitions are different datanodeNames and datanodeSlaveDirs. Check your configuration
ERROR: Number of elements in datanode slave definitions are different datanodeNames and datanodeArchLogDirs. Check your configuration
ERROR: Found fundamental configuration error.
Segmentation fault (core dumped)
So, I think if you create one data node slave, you will have to create one for all, otherwise a core dump will occur.