I want to set up solr cloud with data split across 2 machines. For now, I need no replication, load balancing, or fault tolerance. Is there a simple way of achieving this? Most of the tutorials end up talking a lot about external zookeeper dependencies, which I think aren't needed for the barebones configuration I mentioned, and it has been hard to use those to create what I want.
If you do not need any fault tolerance, you can just start two Solr cloud instances and point them to the embedded Zookeeper. You'll need three nodes for Zookeeper to be able to do establish a quorum on failure anyway.
The embedded zookeeper runs on port <solrport + 1000>, and you'd start the nodes with -z host1:port,host2:port. In this case you could just point the latter instance to the first one, since you don't need any fault tolerance.
This is the same configuration as given in Getting Started with Solr Cloud.