import java.io.IOException;

import java.util.concurrent.TimeUnit;

import java.util.logging.Logger;

import java.util.concurrent.ExecutorService;

import java.util.concurrent.Executors;

/**

* Server that manages startup/shutdown of a {@code Greeter} server.

private static final Logger logger = Logger.getLogger(HelloWorldServer.class.getName());

private Server server;

private void start() throws IOException {

/* The port on which the server should run */

int port = 50051;

/*

* By default gRPC uses a global, shared Executor.newCachedThreadPool() for gRPC callbacks into

* your application. This is convenient, but can cause an excessive number of threads to be

* created if there are many RPCs. It is often better to limit the number of threads your

* application uses for processing and let RPCs queue when the CPU is saturated.

* The appropriate number of threads varies heavily between applications.

* Async application code generally does not need more threads than CPU cores.

*/

ExecutorService executor = Executors.newFixedThreadPool(2);

server = Grpc.newServerBuilderForPort(port, InsecureServerCredentials.create())

.executor(executor)

.addService(new GreeterImpl())

.build()

.start();

try {

HelloWorldServer.this.stop();

} catch (InterruptedException e) {

if (server != null) {

server.shutdownNow();

}

e.printStackTrace(System.err);

} finally {

executor.shutdown();

}

System.err.println("*** server shut down");

}