Going reactive with Spring Data

Last weeks' Spring Data Kay M1 is the first release ever that comes with support for reactive data access. Its initial set of supported stores — MongoDB, Apache Cassandra and Redis — all ship reactive drivers already, which made them very natural candidates for such a prototype. Let’s take a more detailed look at the new programming model and the APIs that make up that support.

Reactive Repositories

The repositories programming model is the most high-level abstraction Spring Data users usually deal with. They’re usually comprised of a set of CRUD methods defined in a Spring Data provided interface and domain-specific query methods. Here’s what a reactive Spring Data repository definition would look like:

public interface ReactivePersonRepository
  extends ReactiveCrudRepository<Person, String> {

  Flux<Person> findByLastname(Mono<String> lastname);

  @Query("{ 'firstname': ?0, 'lastname': ?1}")
  Mono<Person> findByFirstnameAndLastname(String firstname, String lastname);
}

As you can see, there’s not too much difference to what you’re used to. However, in contrast to the traditional repository interfaces, a reactive repository uses reactive types as return types and can do so for parameter types, too. The CRUD methods in the newly introduced ReactiveCrudRepository, of course make use of these types, too.

By default, reactive repositories use Project Reactor types but other reactive libraries can also be used. We provide a custom repository base interface (e.g. RxJava2CrudRepository) for those and also automatically adapt the types as needed for query methods, e.g RxJava’s Observable and Single. The rest basically stays the same. Note, however, that the current milestone does not support pagination yet and you of course have to have the necessary reactive libraries on the classpath to activate support for a particular library.

Activating reactive Spring Data

Similarly to what we have in the blocking world, the support for reactive Spring Data is activated through an @Enable… annotation alongside some infrastructure setup:

@EnableReactiveMongoRepositories
public class AppConfig extends AbstractReactiveMongoConfiguration {

  @Bean
  public MongoClient mongoClient() {
    return MongoClients.create();
  }

  @Override
  protected String getDatabaseName() {
    return "reactive";
  }
}

See how we use a different base class for the infrastructure configuration, as we need to make use of the MongoDB async driver.

Read more...

SpringOne Platform 2016 Replay: Leadership Election with Spring Cloud Cluster

Recorded at SpringOne Platform 2016.
Speaker: Dr. David Syer
Slides: http://www.slideshare.net/SpringCentral/leadership-election-with-spring-cloud-cluster

Leader election allows application to work together with other applications to coordinate a cluster leadership via a third party system. A leader can then be used to provide global state or global ordering, generally without sacrificing availability. In this presentation we show how Spring Cloud Cluster provides a simple abstraction for leader election and how it is implemented using zookeeper, hazelcast and etcd.

Read more...