In this hands on, Ignite is used as an embedded library.
It can also be used in standalone mode.
Ignite is designed to work in cluster. Every node can be started like this:
import org.apache.ignite.Ignite;
public class Main {
public static void main(String[] args) {
IgniteConfiguration configuration = new IgniteConfiguration();
Ignition.start(configuration); // that's all folks !
}
}/!\ For the hands on, do not use default configuration settings, otherwise you'll be joining a cluster with all nodes from local network. Instead use configuration from Config.igniteConfiguration().
By default, server nodes participate in all computing and caching operations in the cluster.
In this hands-on server nodes are plain-old Java applications.
Complete TODO in ServerApp.
Start 3 ServerApp instances (ServerApp1, ServerApp2 and ServerApp3).
When server nodes are started you should progressively see in each ServerApp node:
Topology snapshot [ver=1, servers=1, clients=0, ...] Topology snapshot [ver=1, servers=2, clients=0, ...] Topology snapshot [ver=1, servers=3, clients=0, ...]
By default, client nodes do not participate in cluster operations, but rather forward them to server nodes.
In this hands on, each test starts a new client node and sends compute queries to server nodes.
A cluster can perfectly work with server nodes only. Using client nodes avoids restarting all server nodes every time a test is executed.
Play around and execute some compute queries from ClientApp node. Try this for example:
import org.apache.ignite.Ignite; public class Main { public static void main(String[] args) { IgniteConfiguration configuration = Config.igniteConfiguration(); Ignite ignite = Ignition.start(configuration); ignite.compute().broadcast(() -> System.out.println("Hello World")); } }Check each ServerApp console!