While documenting legacy distributed search, for the purpose of an example, I wanted to start 2 instances on the same machine in standalone mode with a core each and the same config set.
Here's what I did to start the 2 nodes:
So far so good. Now, create_core doesn't accept a port number and so it pseudo-randomly picks a node to create the core i.e. I can't create a core using scripts on both nodes smoothly unless we support "-p <port number>" with that call (and may be collection too?).
FYI, I also tried :
but this failed as -e overrides -s. I don't really remember why we did that, but perhaps we can consider not overriding -s, even when -e is specified i.e. copy whatever is required and use -s.