Is it possible to start with a 2-node Ray Core and Serve cluster on a laptop (during development) and scale it to a cluster of servers for production?
To be clear @Henry_Thornton do you want two node on the same laptop? If so, take a look at ray.cluster_utils.Cluster: ray.cluster_utils — Ray v1.8.0
Yes, I want 2-nodes on the same laptop. Will look at ray.cluster_utils. Thank-you
On a single machine (eg. multi-core cpu laptop or a multi-core server), is the minimum Ray Core + Ray Serve system a Head-node or a Head-node + one Worker-node?
You just need a head node as minimal. In fact this is what the ray.init() is doing if you run it inside a process with connecting to ray cluster, it will start a local head node.