MongoDB Java Server

Fake implementation of the core MongoDB server in Java that can be used for integration tests.

Think of H2/HSQLDB/SQLite but for MongoDB.

The MongoDB Wire Protocol is implemented with Netty. Different backends are possible and can be extended.

In-Memory backend

The in-memory backend is the default backend that is typically used to fake MongoDB for integration tests. It supports most CRUD operations, commands and the aggregation framework. Some features are not yet implemented, such as full-text search or map/reduce.

Add the following Maven dependency to your project:

<dependency>
    <groupId>de.bwaldvogel</groupId>
    <artifactId>mongo-java-server</artifactId>
    <version>1.38.0</version>
</dependency>

Example

public class SimpleTest {

    private MongoCollection<Document> collection;
    private MongoClient client;
    private MongoServer server;

    @Before
    public void setUp() {
        server = new MongoServer(new MemoryBackend());

        // optionally:
        // server.enableSsl(key, keyPassword, certificate);
        // server.enableOplog();

        // bind on a random local port
        InetSocketAddress serverAddress = server.bind();

        client = new MongoClient(new ServerAddress(serverAddress));
        collection = client.getDatabase("testdb").getCollection("testcollection");
    }

    @After
    public void tearDown() {
        client.close();
        server.shutdown();
    }

    @Test
    public void testSimpleInsertQuery() throws Exception {
        assertEquals(0, collection.count());

        // creates the database and collection in memory and insert the object
        Document obj = new Document("_id", 1).append("key", "value");
        collection.insertOne(obj);

        assertEquals(1, collection.count());
        assertEquals(obj, collection.find().first());
    }

}

Example with SpringBoot

@RunWith(SpringRunner.class)
@SpringBootTest(classes={SimpleSpringBootTest.TestConfiguration.class})
public class SimpleSpringBootTest {

    @Autowired private MyRepository repository;

    @Before
    public void setUp() {
        // initialize your repository with some test data
        repository.deleteAll();
        repository.save(...);
    }

    @Test
    public void testMyRepository() {
        // test your repository ...
        ...
    }

    @Configuration
    @EnableMongoTestServer
    @EnableMongoRepositories(basePackageClasses={MyRepository.class})
    protected static class TestConfiguration {
        // test bean definitions ...
        ...
    }
}

@Target(ElementType.TYPE)
@Retention(RetentionPolicy.RUNTIME)
@Import(MongoTestServerConfiguration.class)
public @interface EnableMongoTestServer {

}

public class MongoTestServerConfiguration {
	@Bean
	public MongoTemplate mongoTemplate(MongoDbFactory mongoDbFactory) {
		return new MongoTemplate(mongoDbFactory);
	}

	@Bean
	public MongoDbFactory mongoDbFactory(MongoServer mongoServer) {
		InetSocketAddress serverAddress = mongoServer.getLocalAddress();
		return new SimpleMongoClientDbFactory("mongodb://" + serverAddress.getHostName() + ":" + serverAddress.getPort() + "/test");
	}

	@Bean(destroyMethod = "shutdown")
	public MongoServer mongoServer() {
		MongoServer mongoServer = new MongoServer(new MemoryBackend());
		mongoServer.bind();
		return mongoServer;
	}
}

H2 MVStore backend

The H2 MVStore backend connects the server to a MVStore that can either be in-memory or on-disk.

<dependency>
    <groupId>de.bwaldvogel</groupId>
    <artifactId>mongo-java-server-h2-backend</artifactId>
    <version>1.38.0</version>
</dependency>

Example

public class Application {

    public static void main(String[] args) throws Exception {
        MongoServer server = new MongoServer(new H2Backend("database.mv"));
        server.bind("localhost", 27017);
    }

}

PostgreSQL backend

The PostgreSQL backend is a proof-of-concept implementation that connects the server to a database in a running PostgreSQL 9.5+ instance. Each MongoDB database is mapped to a schema in Postgres and each MongoDB collection is stored as a table.

<dependency>
    <groupId>de.bwaldvogel</groupId>
    <artifactId>mongo-java-server-postgresql-backend</artifactId>
    <version>1.38.0</version>
</dependency>

Example

public class Application {

    public static void main(String[] args) throws Exception {
        DataSource dataSource = new org.postgresql.jdbc3.Jdbc3PoolingDataSource();
        dataSource.setDatabaseName(…);
        dataSource.setUser(…);
        dataSource.setPassword(…);
        MongoServer server = new MongoServer(new PostgresqlBackend(dataSource));
        server.bind("localhost", 27017);
    }

}

Building a "fat" JAR that contains all dependencies

If you want to build a version that is not on Maven Central you can do the following:

1. Build a "fat" JAR that includes all dependencies using "./gradlew shadowJar"
2. Copy build/libs/mongo-java-server-[version]-all.jar to your project, e.g. to the libs directory.
3. Import that folder (e.g. via Gradle using testCompile fileTree(dir: 'libs', include: '*.jar'))

Contributing

Please read the contributing guidelines if you want to contribute code to the project.

If you want to thank the author for this library or want to support the maintenance work, we are happy to receive a donation.

Ideas for other backends

Faulty backend

A faulty backend could randomly fail queries or cause timeouts. This could be used to test the client for error resilience.

Fuzzy backend

Fuzzing the wire protocol could be used to check the robustness of client drivers.

Related Work

• Embedded MongoDB

  • Spins up a real MongoDB instance

• fongo

  • focus on unit testing
  • no wire protocol implementation
  • intercepts the java mongo driver
  • currently used in nosql-unit