Showing posts with label jpa. Show all posts
Showing posts with label jpa. Show all posts

Tuesday, November 6, 2018

Building Enterprise Java applications, the Spring way

I think it is fair to say that Java EE has gained pretty bad reputation among Java developers. Despite the fact that it has certainly improved on all fronts over the years, even changed home to Eclipse Foundation to become Jakarta EE, its bitter taste is still quite strong. On the other side we have Spring Framework (or to reflect the reality better, a full-fledged Spring Platform): brilliant, lightweight, fast, innovative and hyper-productive Java EE replacement. So why to bother with Java EE?

We are going to answer this question by showing how easy it is to build modern Java applications using most of Java EE specs. And the key ingredient to succeed here is Eclipse Microprofile: enterprise Java in the age of microservices.

The application we are going to build is RESTful web API to manage people, as simple as that. The standard way to build RESTful web services in Java is by using JAX-RS 2.1 (JSR-370). Consequently, CDI 2.0 (JSR-365) is going to take care of dependency injection whereas JPA 2.0 (JSR-317) is going to cover the data access layer. And certainly, Bean Validation 2.0 (JSR-380) is helping us to deal with input verification.

The only non-Java EE specification we would be relying on is OpenAPI v3.0 which helps to provide the usable description of our RESTful web APIs. With that, let us get started with the PersonEntity domain model (omitting getters and setters as not very relevant details):

@Entity
@Table(name = "people")
public class PersonEntity {
    @Id @Column(length = 256) 
    private String email;

    @Column(nullable = false, length = 256, name = "first_name")
    private String firstName;

    @Column(nullable = false, length = 256, name = "last_name")
    private String lastName;

    @Version
    private Long version;
}

It just has the absolute minimum set of properties. The JPA repository is pretty straightforward and implements typical set of CRUD methods.

@ApplicationScoped
@EntityManagerConfig(qualifier = PeopleDb.class)
public class PeopleJpaRepository implements PeopleRepository {
    @Inject @PeopleDb private EntityManager em;

    @Override
    @Transactional(readOnly = true)
    public Optional<PersonEntity> findByEmail(String email) {
        final CriteriaBuilder cb = em.getCriteriaBuilder();
    
        final CriteriaQuery<PersonEntity> query = cb.createQuery(PersonEntity.class);
        final Root<PersonEntity> root = query.from(PersonEntity.class);
        query.where(cb.equal(root.get(PersonEntity_.email), email));
        
        try {
            final PersonEntity entity = em.createQuery(query).getSingleResult();
            return Optional.of(entity);
        } catch (final NoResultException ex) {
            return Optional.empty();
        }
    }

    @Override
    @Transactional
    public PersonEntity saveOrUpdate(String email, String firstName, String lastName) {
        final PersonEntity entity = new PersonEntity(email, firstName, lastName);
        em.persist(entity);
        return entity;
    }

    @Override
    @Transactional(readOnly = true)
    public Collection<PersonEntity> findAll() {
        final CriteriaBuilder cb = em.getCriteriaBuilder();
        final CriteriaQuery<PersonEntity> query = cb.createQuery(PersonEntity.class);
        query.from(PersonEntity.class);
        return em.createQuery(query).getResultList();
    }

    @Override
    @Transactional
    public Optional<PersonEntity> deleteByEmail(String email) {
        return findByEmail(email)
            .map(entity -> {
                em.remove(entity);
                return entity;
            });
    }
}

The transaction management (namely, the @Transactional annotation) needs some explanation. In the typical Java EE application, the container runtime is responsible for managing the transactions. Since we don't want to onboard the application container but stay lean, we could have used EntityManager to start / commit / rollback transactions. It would certainly work out but pollute the code with the boilerplate. Arguably, the better option is to use Apache DeltaSpike CDI extensions for declarative transaction management (this is where @Transactional and @EntityManagerConfig annotations are coming from). The snippet below illustrates how it is being integrated.

@ApplicationScoped
public class PersistenceConfig {
    @PersistenceUnit(unitName = "peopledb")
    private EntityManagerFactory entityManagerFactory;

    @Produces @PeopleDb @TransactionScoped
    public EntityManager create() {
        return this.entityManagerFactory.createEntityManager();
    }

    public void dispose(@Disposes @PeopleDb EntityManager entityManager) {
        if (entityManager.isOpen()) {
            entityManager.close();
        }
    }
}

Awesome, the hardest part is already behind! The Person data transfer object and the service layer are coming next.

public class Person {
    @NotNull private String email;
    @NotNull private String firstName;
    @NotNull private String lastName;
}

Honestly, for the sake of keeping the example application as small as possible we could skip the service layer altogether and go to the repository directly. But this is, in general, not a very good practice so let us introduce PeopleServiceImpl anyway.

@ApplicationScoped
public class PeopleServiceImpl implements PeopleService {
    @Inject private PeopleRepository repository;

    @Override
    public Optional<Person> findByEmail(String email) {
        return repository
            .findByEmail(email)
            .map(this::toPerson);
    }

    @Override
    public Person add(Person person) {
        return toPerson(repository.saveOrUpdate(person.getEmail(), person.getFirstName(), person.getLastName()));
    }

    @Override
    public Collection<Person> getAll() {
        return repository
            .findAll()
            .stream()
            .map(this::toPerson)
            .collect(Collectors.toList());
    }

    @Override
    public Optional<Person> remove(String email) {
        return repository
            .deleteByEmail(email)
            .map(this::toPerson);
    }
    
    private Person toPerson(PersonEntity entity) {
        return new Person(entity.getEmail(), entity.getFirstName(), entity.getLastName());
    }
}

The only part left is the definition of the JAX-RS application and resources.

@Dependent
@ApplicationPath("api")
@OpenAPIDefinition(
    info = @Info(
        title = "People Management Web APIs", 
        version = "1.0.0", 
        license = @License(
            name = "Apache License", 
            url = "https://www.apache.org/licenses/LICENSE-2.0"
        )
    )
)
public class PeopleApplication extends Application {
}

Not much to say, as simple as it could possibly be. The JAX-RS resource implementation is a bit more interesting though (the OpenAPI annotations are taking most of the place).

@ApplicationScoped
@Path( "/people" ) 
@Tag(name = "people")
public class PeopleResource {
    @Inject private PeopleService service;
    
    @Produces(MediaType.APPLICATION_JSON)
    @GET
    @Operation(
        description = "List all people", 
        responses = {
            @ApiResponse(
                content = @Content(array = @ArraySchema(schema = @Schema(implementation = Person.class))),
                responseCode = "200"
            )
        }
    )
    public Collection<Person> getPeople() {
        return service.getAll();
    }

    @Produces(MediaType.APPLICATION_JSON)
    @Path("/{email}")
    @GET
    @Operation(
        description = "Find person by e-mail", 
        responses = {
            @ApiResponse(
                content = @Content(schema = @Schema(implementation = Person.class)), 
                responseCode = "200"
            ),
            @ApiResponse(
                responseCode = "404", 
                description = "Person with such e-mail doesn't exists"
            )
        }
    )
    public Person findPerson(@Parameter(description = "E-Mail address to lookup for", required = true) @PathParam("email") final String email) {
        return service
            .findByEmail(email)
            .orElseThrow(() -> new NotFoundException("Person with such e-mail doesn't exists"));
    }

    @Consumes(MediaType.APPLICATION_JSON)
    @Produces(MediaType.APPLICATION_JSON)
    @POST
    @Operation(
        description = "Create new person",
        requestBody = @RequestBody(
            content = @Content(schema = @Schema(implementation = Person.class)),
        ), 
        responses = {
            @ApiResponse(
                 content = @Content(schema = @Schema(implementation = Person.class)),
                 headers = @Header(name = "Location"),
                 responseCode = "201"
            ),
            @ApiResponse(
                responseCode = "409", 
                description = "Person with such e-mail already exists"
            )
        }
    )
    public Response addPerson(@Context final UriInfo uriInfo,
            @Parameter(description = "Person", required = true) @Valid Person payload) {

        final Person person = service.add(payload);
        return Response
             .created(uriInfo.getRequestUriBuilder().path(person.getEmail()).build())
             .entity(person)
             .build();
    }
    
    @Path("/{email}")
    @DELETE
    @Operation(
        description = "Delete existing person",
        responses = {
            @ApiResponse(
                responseCode = "204",
                description = "Person has been deleted"
            ),
            @ApiResponse(
                responseCode = "404", 
                description = "Person with such e-mail doesn't exists"
            )
        }
    )
    public Response deletePerson(@Parameter(description = "E-Mail address to lookup for", required = true ) @PathParam("email") final String email) {
        return service
            .remove(email)
            .map(r -> Response.noContent().build())
            .orElseThrow(() -> new NotFoundException("Person with such e-mail doesn't exists"));
    }
}

And with that, we are done! But how could we assemble and wire all these pieces together? Here is the time for Microprofile to enter the stage. There are many implementations to chose from, the one we are going to use in this post is Project Hammock. The only thing we have to do is to specify the CDI 2.0, JAX-RS 2.1 and JPA 2.0 implementations we would like to use, which translates to Weld, Apache CXF, and OpenJPA respectively (expressed through the Project Hammock dependencies). Let us take a look on the Apache Maven pom.xml file.

<properties>
    <deltaspike.version>1.8.1</deltaspike.version>
    <hammock.version>2.1</hammock.version>
</properties>

<dependencies>
    <dependency>
        <groupId>org.apache.deltaspike.modules</groupId>
        <artifactId>deltaspike-jpa-module-api</artifactId>
        <version>${deltaspike.version}</version>
        <scope>compile</scope>
    </dependency>

    <dependency>
        <groupId>org.apache.deltaspike.modules</groupId>
        <artifactId>deltaspike-jpa-module-impl</artifactId>
        <version>${deltaspike.version}</version>
        <scope>runtime</scope>
    </dependency>

    <dependency>
        <groupId>ws.ament.hammock</groupId>
        <artifactId>dist-microprofile</artifactId>
        <version>${hammock.version}</version>
    </dependency>

    <dependency>
        <groupId>ws.ament.hammock</groupId>
        <artifactId>jpa-openjpa</artifactId>
        <version>${hammock.version}</version>
    </dependency>

    <dependency>
        <groupId>ws.ament.hammock</groupId>
        <artifactId>util-beanvalidation</artifactId>
        <version>${hammock.version}</version>
    </dependency>

    <dependency>
        <groupId>ws.ament.hammock</groupId>
        <artifactId>util-flyway</artifactId>
        <version>${hammock.version}</version>
    </dependency>

    <dependency>
        <groupId>ws.ament.hammock</groupId>
        <artifactId>swagger</artifactId>
        <version>${hammock.version}</version>
    </dependency>
</dependencies>

Without further ado, let us build and run the application right away (if you are curious what relational datastore the application is using, it is H2 with the database configured in-memory).

> mvn clean package
> java -jar target/eclipse-microprofile-hammock-0.0.1-SNAPSHOT-capsule.jar 

The best way to ensure that our people management RESTful web APIs are fully functional is to send a couple of requests to it:

>  curl -X POST http://localhost:10900/api/people -H "Content-Type: application\json" \
     -d '{"email": "a@b.com", "firstName": "John", "lastName": "Smith"}'

HTTP/1.1 201 Created
Location: http://localhost:10900/api/people/a@b.com
Content-Type: application/json

{
    "firstName":"John","
    "lastName":"Smith",
    "email":"a@b.com"
}

What about making sure the Bean Validation is working fine? To trigger that, let us send the partially prepared request.

>  curl  --X POST http://localhost:10900/api/people -H "Content-Type: application\json" \
     -d '{"firstName": "John", "lastName": "Smith"}'

HTTP/1.1 400 Bad Request
Content-Length: 0

The OpenAPI specification and pre-bundled Swagger UI distribution are also available at http://localhost:10900/index.html?url=http://localhost:10900/api/openapi.json.

So far so good but fairly speaking we have not talked about testing our application at all. How hard it would be to come up with the integration test for, let say, the scenario of adding a person? It turns out that the frameworks around testing Java EE applications have improved a lot. In particular, it is exceptionally easy to accomplish with Arquillian test framework (along with beloved JUnit and REST Assured). One real example is worth thousand words.

@RunWith(Arquillian.class)
@EnableRandomWebServerPort
public class PeopleApiTest {
    @ArquillianResource private URI uri;
    
    @Deployment
    public static JavaArchive createArchive() {
        return ShrinkWrap
            .create(JavaArchive.class)
            .addClasses(PeopleResource.class, PeopleApplication.class)
            .addClasses(PeopleServiceImpl.class, PeopleJpaRepository.class, PersistenceConfig.class)
            .addPackages(true, "org.apache.deltaspike");
    }
            
    @Test
    public void shouldAddNewPerson() throws Exception {
        final Person person = new Person("a@b.com", "John", "Smith");
        
        given()
            .contentType(ContentType.JSON)
            .body(person)
            .post(uri + "/api/people")
            .then()
            .assertThat()
            .statusCode(201)
            .body("email", equalTo("a@b.com"))
            .body("firstName", equalTo("John"))
            .body("lastName", equalTo("Smith"));
    }
}

Amazing, isn't it? It is actually a lot of fun to develop modern Java EE applications, someone may say, the Spring way! And in fact, the parallels with Spring are not coincidental since it was inspiring, is inspiring and undoubtedly is going to continue inspire a lot of innovations in the Java EE ecosystem.

How the future is looking like? I think, by all means bright, both for Jakarta EE and Eclipse Microprofile. The latter just approached the version 2.0 with tons of new specifications included, oriented to address the needs of the microservice architectures. It is awesome to witness these transformations happening.

The complete source of the project is available on Github.

Saturday, February 28, 2015

A fresh look on accessing database on JVM platform: Slick from Typesafe

In today's post we are going to open our mind, step away from traditional Java EE / Java SE JPA-based stack (which I think is great) and take a refreshing look on how to access database in your Java applications using the new kid on the block: Slick 2.1 from Typesafe. So if JPA is so great, why bother? Well, sometimes you need to do very simple things and there is no need to bring the complete, well modeled persistence layer for that. In here Slick shines.

In the essence, Slick is database access library, not an ORM. Though it is written in Scala, the examples we are going to look at do not require any particular knowledge of this excellent language (although, it is just Scala that made Slick possible to exist). Our relational database schema will have only two tables, customers and addresses, linked by one-to-many relationships. For simplicity, H2 has been picked as an in-memory database engine.

The first question which comes up is defining database tables (schema) and, naturally, database specific DDLs are the standard way of doing that. Can we do something about it and try another approach? If you are using Slick 2.1, the answer is yes, absolutely. Let us just describe our tables as Scala classes:

// The 'customers' relation table definition
class Customers( tag: Tag ) extends Table[ Customer ]( tag, "customers" ) {
  def id = column[ Int ]( "id", O.PrimaryKey, O.AutoInc )

  def email = column[ String ]( "email", O.Length( 512, true ), O.NotNull )
  def firstName = column[ String ]( "first_name", O.Length( 256, true ), O.Nullable )
  def lastName = column[ String ]( "last_name", O.Length( 256, true ), O.Nullable )

  // Unique index for customer's email
  def emailIndex = index( "idx_email", email, unique = true )
}

Very easy and straightforward, resembling a lot typical CREATE TABLE construct. The addresses table is going to be defined the same way and reference users table by foreign key.

// The 'customers' relation table definition
class Addresses( tag: Tag ) extends Table[ Address ]( tag, "addresses" ) {
  def id = column[ Int ]( "id", O.PrimaryKey, O.AutoInc )

  def street = column[ String ]( "street", O.Length( 100, true ), O.NotNull )
  def city = column[ String ]( "city", O.Length( 50, true ), O.NotNull )
  def country = column[ String ]( "country", O.Length( 50, true ), O.NotNull )

  // Foreign key to 'customers' table
  def customerId = column[Int]( "customer_id", O.NotNull )
  def customer = foreignKey( "customer_fk", customerId, Customers )( _.id )
}

Great, leaving off some details, that is it: we have defined two database tables in pure Scala. But details are important and we are going to look closely on following two declarations: Table[ Customer ] and Table[ Address ]. Essentially, each table could be represented as a tuple with as many elements as column it has defined. For example, customers is a tuple of (Int, String, String, String), while addresses table is a tuple of (Int, String, String, String, Int). Tuples in Scala are great but they are not very convenient to work with. Luckily, Slick allows to use case classes instead of tuples by providing so called Lifted Embedding technique. Here are our Customer and Address case classes:

case class Customer( id: Option[Int] = None, email: String, 
  firstName: Option[ String ] = None, lastName: Option[ String ] = None)

case class Address( id: Option[Int] = None,  street: String, city: String, 
  country: String, customer: Customer )

The last but not least question is how Slick is going to convert from tuples to case classes and vice-versa? It would be awesome to have such conversion out-of-the box but at this stage Slick needs a bit of help. Using Slick terminology, we are going to shape * table projection (which corresponds to SELECT * FROM ... SQL construct). Let see how it looks like for customers:

// Converts from Customer domain instance to table model and vice-versa
def * = ( id.?, email, firstName.?, lastName.? ).shaped <> ( 
  Customer.tupled, Customer.unapply )

For addresses table, shaping looks a little bit more verbose due to the fact that Slick does not have a way to refer to Customer case class instance by foreign key. Still, it is pretty straightforward, we just construct temporary Customer from its identifier.

// Converts from Customer domain instance to table model and vice-versa
def * = ( id.?, street, city, country, customerId ).shaped <> ( 
  tuple => {
    Address.apply(
      id = tuple._1,
      street = tuple._2,
      city = tuple._3,
      country = tuple._4,
      customer = Customer( Some( tuple._5 ), "" )
    )
  }, {
    (address: Address) =>
      Some { (
        address.id,
        address.street,
        address.city,
        address.country,
        address.customer.id getOrElse 0 
      )
    }
  }
)

Now, when all details have been explained, how can we materialize our Scala table definitions into the real database tables? Thankfully to Slick, it is a easy as that:

implicit lazy val DB = Database.forURL( "jdbc:h2:mem:test", driver = "org.h2.Driver" )
  
DB withSession { implicit session =>
  ( Customers.ddl ++ Addresses.ddl ).create
}

Slick has many ways to query and update data in database. The most beautiful and powerful one is just using pure functional constructs of the Scala language. The easiest way of doing that is by defining companion object and implement typical CRUD operations in it. For example, here is the method which inserts new customer record into customers table:

object Customers extends TableQuery[ Customers ]( new Customers( _ ) ) {
  def create( customer: Customer )( implicit db: Database ): Customer = 
    db.withSession { implicit session =>
      val id = this.autoIncrement.insert( customer )
      customer.copy( id = Some( id ) )
    } 
}

And it could be used like this:

Customers.create( Customer( None, "tom@b.com",  Some( "Tom" ), Some( "Tommyknocker" ) ) )
Customers.create( Customer( None, "bob@b.com",  Some( "Bob" ), Some( "Bobbyknocker" ) ) )

Similarly, the family of find functions could be implemented using regular Scala for comprehension:

def findByEmail( email: String )( implicit db: Database ) : Option[ Customer ] = 
  db.withSession { implicit session =>
    ( for { customer <- this if ( customer.email === email.toLowerCase ) } 
      yield customer ) firstOption
  }
   
def findAll( implicit db: Database ): Seq[ Customer ] = 
  db.withSession { implicit session =>      
    ( for { customer <- this } yield customer ) list
  }

And here are usage examples:

val customers = Customers.findAll
val customer = Customers.findByEmail( "bob@b.com" )

Updates and deletes are a bit different though very simple as well, let us take a look on those:

def update( customer: Customer )( implicit db: Database ): Boolean = 
  db.withSession { implicit session =>
    val query = for { c <- this if ( c.id === customer.id ) } 
      yield (c.email, c.firstName.?, c.lastName.?)
    query.update(customer.email, customer.firstName, customer.lastName) > 0
  }
  
def remove( customer: Customer )( implicit db: Database ) : Boolean = 
  db.withSession { implicit session =>
    ( for { c <- this if ( c.id === customer.id ) } yield c ).delete > 0
  }

Let us see those two methods in action:

Customers.findByEmail( "bob@b.com" ) map { customer =>
  Customers.update( customer.copy( firstName = Some( "Tommy" ) ) )
  Customers.remove( customer )
}

Looks very neat. I am personally still learning Slick however I am pretty excited about it. It helps me to have things done much faster, enjoying the beauty of Scala language and functional programming. No doubts, the upcoming version 3.0 is going to bring even more interesting features, I am looking forward to it.

This post is just an introduction into world of Slick, a lot of implementation details and use cases have been left aside to keep it short and simple. However Slick's documentation is pretty good and please do not hesitate to consult it.

The complete project is available on GitHub.

Saturday, August 9, 2014

OSGi: the gateway into micro-services architecture

The terms "modularity" and "microservices architecture" pop up quite often these days in context of building scalable, reliable distributed systems. Java platform itself is known to be weak with regards to modularity (Java 9 is going to address that by delivering project Jigsaw), giving a chance to frameworks like OSGi and JBoss Modules to emerge.

When I first heard about OSGi back in 2007, I was truly excited about all those advantages Java applications might benefit of by being built on top of it. But very quickly the frustration took place instead of excitement: no tooling support, very limited set of compatible libraries and frameworks, quite unstable and hard to troubleshoot runtime. Clearly, it was not ready to be used by average Java developer and as such, I had to put it on the shelf. With years, OSGi has matured a lot and gained a widespread community support.

The curious reader may ask: what are the benefits of using modules and OSGi in particular? To name just a few problems it helps to solve:

  • explicit (and versioned) dependency management: modules declare what they need (and optionally the version ranges)
  • small footprint: modules are not packaged with all their dependencies
  • easy release: modules can be developed and released independently
  • hot redeploy: individual modules may be redeployed without affecting others

In today's post we are going to take a 10000 feet view on a state of the art in building modular Java applications using OSGi. Leaving aside discussions how good or bad OSGi is, we are going to build an example application consisting of following modules:

  • data access module
  • business services module
  • REST services module
Apache OpenJPA 2.3.0 / JPA 2.0 for data access (unfortunately, JPA 2.1 is not yet supported by OSGi implementation of our choice), Apache CXF 3.0.1 / JAX-RS 2.0 for REST layer are two main building blocks of the application. I found Christian Schneider's blog, Liquid Reality, to be invaluable source of information about OSGi (as well as many other topics).

In OSGi world, the modules are called bundles. Bundles manifest their dependencies (import packages) and the packages they expose (export packages) so other bundles are able to use them. Apache Maven supports this packaging model as well. The bundles are managed by OSGi runtime, or container, which in our case is going to be Apache Karaf 3.0.1 (actually, it is the single thing we need to download and unpack).

Let me stop talking and better show some code. We are going to start from the top (REST) and go all the way to the bottom (data access) as it would be easier to follow. Our PeopleRestService is a typical example of JAX-RS 2.0 service implementation:

package com.example.jaxrs;

import java.util.Collection;

import javax.ws.rs.DELETE;
import javax.ws.rs.DefaultValue;
import javax.ws.rs.FormParam;
import javax.ws.rs.GET;
import javax.ws.rs.POST;
import javax.ws.rs.PUT;
import javax.ws.rs.Path;
import javax.ws.rs.PathParam;
import javax.ws.rs.Produces;
import javax.ws.rs.QueryParam;
import javax.ws.rs.core.Context;
import javax.ws.rs.core.MediaType;
import javax.ws.rs.core.Response;
import javax.ws.rs.core.UriInfo;

import com.example.data.model.Person;
import com.example.services.PeopleService;

@Path( "/people" )
public class PeopleRestService {
    private PeopleService peopleService;
        
    @Produces( { MediaType.APPLICATION_JSON } )
    @GET
    public Collection< Person > getPeople( 
            @QueryParam( "page") @DefaultValue( "1" ) final int page ) {
        return peopleService.getPeople( page, 5 );
    }

    @Produces( { MediaType.APPLICATION_JSON } )
    @Path( "/{email}" )
    @GET
    public Person getPerson( @PathParam( "email" ) final String email ) {
        return peopleService.getByEmail( email );
    }

    @Produces( { MediaType.APPLICATION_JSON  } )
    @POST
    public Response addPerson( @Context final UriInfo uriInfo,
            @FormParam( "email" ) final String email, 
            @FormParam( "firstName" ) final String firstName, 
            @FormParam( "lastName" ) final String lastName ) {
        
        peopleService.addPerson( email, firstName, lastName );
        return Response.created( uriInfo
            .getRequestUriBuilder()
            .path( email )
            .build() ).build();
    }
    
    @Produces( { MediaType.APPLICATION_JSON  } )
    @Path( "/{email}" )
    @PUT
    public Person updatePerson( @PathParam( "email" ) final String email, 
            @FormParam( "firstName" ) final String firstName, 
            @FormParam( "lastName" )  final String lastName ) {
        
        final Person person = peopleService.getByEmail( email );
        
        if( firstName != null ) {
            person.setFirstName( firstName );
        }
        
        if( lastName != null ) {
            person.setLastName( lastName );
        }

        return person;              
    }
    
    @Path( "/{email}" )
    @DELETE
    public Response deletePerson( @PathParam( "email" ) final String email ) {
        peopleService.removePerson( email );
        return Response.ok().build();
    }
    
    public void setPeopleService( final PeopleService peopleService ) {
        this.peopleService = peopleService;
    }
}

As we can see, there is nothing here telling us about OSGi. The only dependency is the PeopleService which somehow should be injected into the PeopleRestService. How? Typically, OSGi applications use blueprint as the dependency injection framework, very similar to old buddy, XML based Spring configuration. It should be packaged along with application inside OSGI-INF/blueprint folder. Here is a blueprint example for our REST module, built on top of Apache CXF 3.0.1:

<blueprint xmlns="http://www.osgi.org/xmlns/blueprint/v1.0.0"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xmlns:jaxrs="http://cxf.apache.org/blueprint/jaxrs"
    xmlns:cxf="http://cxf.apache.org/blueprint/core"
    xsi:schemaLocation="
        http://www.osgi.org/xmlns/blueprint/v1.0.0 
        http://www.osgi.org/xmlns/blueprint/v1.0.0/blueprint.xsd
        http://cxf.apache.org/blueprint/jaxws 
        http://cxf.apache.org/schemas/blueprint/jaxws.xsd
        http://cxf.apache.org/blueprint/jaxrs 
        http://cxf.apache.org/schemas/blueprint/jaxrs.xsd
        http://cxf.apache.org/blueprint/core 
        http://cxf.apache.org/schemas/blueprint/core.xsd">

    <cxf:bus id="bus">
        <cxf:features>
            <cxf:logging/>
        </cxf:features>       
    </cxf:bus>

    <jaxrs:server address="/api" id="api">
        <jaxrs:serviceBeans>
            <ref component-id="peopleRestService"/>
        </jaxrs:serviceBeans>
        <jaxrs:providers>
            <bean class="com.fasterxml.jackson.jaxrs.json.JacksonJsonProvider" />
        </jaxrs:providers>
    </jaxrs:server>
    
    <!-- Implementation of the rest service -->
    <bean id="peopleRestService" class="com.example.jaxrs.PeopleRestService">
        <property name="peopleService" ref="peopleService"/>
    </bean>         
    
    <reference id="peopleService" interface="com.example.services.PeopleService" />
</blueprint>

Very small and simple: basically the configuration just states that in order for the module to work, the reference to the com.example.services.PeopleService should be provided (effectively, by OSGi container). To see how it is going to happen, let us take a look on another module which exposes services. It contains only one interface PeopleService:

package com.example.services;

import java.util.Collection;

import com.example.data.model.Person;

public interface PeopleService {
    Collection< Person > getPeople( int page, int pageSize );
    Person getByEmail( final String email );
    Person addPerson( final String email, final String firstName, final String lastName );
    void removePerson( final String email );
}

And also provides its implementation as PeopleServiceImpl class:

package com.example.services.impl;

import java.util.Collection;

import org.osgi.service.log.LogService;

import com.example.data.PeopleDao;
import com.example.data.model.Person;
import com.example.services.PeopleService;

public class PeopleServiceImpl implements PeopleService {
    private PeopleDao peopleDao;
    private LogService logService;
 
    @Override
    public Collection< Person > getPeople( final int page, final int pageSize ) {        
        logService.log( LogService.LOG_INFO, "Getting all people" );
        return peopleDao.findAll( page, pageSize );
    }

    @Override
    public Person getByEmail( final String email ) {
        logService.log( LogService.LOG_INFO, 
            "Looking for a person with e-mail: " + email );
        return peopleDao.find( email );        
    }

    @Override
    public Person addPerson( final String email, final String firstName, 
            final String lastName ) {
        logService.log( LogService.LOG_INFO, 
            "Adding new person with e-mail: " + email );
        return peopleDao.save( new Person( email, firstName, lastName ) );
    }

    @Override
    public void removePerson( final String email ) {
        logService.log( LogService.LOG_INFO, 
            "Removing a person with e-mail: " + email );
        peopleDao.delete( email );
    }
    
    public void setPeopleDao( final PeopleDao peopleDao ) {
        this.peopleDao = peopleDao;
    }
    
    public void setLogService( final LogService logService ) {
        this.logService = logService;
    }
}

And this time again, very small and clean implementation with two injectable dependencies, org.osgi.service.log.LogService and com.example.data.PeopleDao. Its blueprint configuration, located inside OSGI-INF/blueprint folder, looks quite compact as well:

<blueprint xmlns="http://www.osgi.org/xmlns/blueprint/v1.0.0"  
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"         
    xsi:schemaLocation="
        http://www.osgi.org/xmlns/blueprint/v1.0.0 
        http://www.osgi.org/xmlns/blueprint/v1.0.0/blueprint.xsd">
        
    <service ref="peopleService" interface="com.example.services.PeopleService" />        
    <bean id="peopleService" class="com.example.services.impl.PeopleServiceImpl">
        <property name="peopleDao" ref="peopleDao" />    
        <property name="logService" ref="logService" />
    </bean>
    
    <reference id="peopleDao" interface="com.example.data.PeopleDao" />
    <reference id="logService" interface="org.osgi.service.log.LogService" />
</blueprint>

The references to PeopleDao and LogService are expected to be provided by OSGi container at runtime. Hovewer, PeopleService implementation is exposed as service and OSGi container will be able to inject it into PeopleRestService once its bundle is being activated.

The last piece of the puzzle, data access module, is a bit more complicated: it contains persistence configuration (META-INF/persistence.xml) and basically depends on JPA 2.0 capabilities of the OSGi container. The persistence.xml is quite basic:

<persistence xmlns="http://java.sun.com/xml/ns/persistence"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    version="2.0">
 
    <persistence-unit name="peopleDb" transaction-type="JTA">
        <jta-data-source>
        osgi:service/javax.sql.DataSource/(osgi.jndi.service.name=peopleDb)
        </jta-data-source>       
        <class>com.example.data.model.Person</class>
        
        <properties>
            <property name="openjpa.jdbc.SynchronizeMappings" 
                value="buildSchema"/>         
        </properties>        
    </persistence-unit>
</persistence>

Similarly to the service module, there is an interface PeopleDao exposed:

package com.example.data;

import java.util.Collection;

import com.example.data.model.Person;

public interface PeopleDao {
    Person save( final Person person );
    Person find( final String email );
    Collection< Person > findAll( final int page, final int pageSize );
    void delete( final String email ); 
}

With its implementation PeopleDaoImpl:

package com.example.data.impl;

import java.util.Collection;

import javax.persistence.EntityManager;
import javax.persistence.criteria.CriteriaBuilder;
import javax.persistence.criteria.CriteriaQuery;

import com.example.data.PeopleDao;
import com.example.data.model.Person;

public class PeopleDaoImpl implements PeopleDao {
    private EntityManager entityManager;
 
    @Override
    public Person save( final Person person ) {
        entityManager.persist( person );
        return person;
    }
 
    @Override
    public Person find( final String email ) {
        return entityManager.find( Person.class, email );
    }
 
    public void setEntityManager( final EntityManager entityManager ) {
        this.entityManager = entityManager;
    }
 
    @Override
    public Collection< Person > findAll( final int page, final int pageSize ) {
        final CriteriaBuilder cb = entityManager.getCriteriaBuilder();
     
        final CriteriaQuery< Person > query = cb.createQuery( Person.class );
        query.from( Person.class );
     
        return entityManager
            .createQuery( query )
            .setFirstResult(( page - 1 ) * pageSize )
            .setMaxResults( pageSize ) 
            .getResultList();
    }
 
    @Override
    public void delete( final String email ) {
        entityManager.remove( find( email ) );
    }
}

Please notice, although we are performing data manipulations, there is no mention of transactions as well as there are no explicit calls to entity manager's transactions API. We are going to use the declarative approach to transactions as blueprint configuration supports that (the location is unchanged, OSGI-INF/blueprint folder):

<blueprint xmlns="http://www.osgi.org/xmlns/blueprint/v1.0.0"  
    xmlns:jpa="http://aries.apache.org/xmlns/jpa/v1.1.0"
    xmlns:tx="http://aries.apache.org/xmlns/transactions/v1.0.0" 
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"         
    xsi:schemaLocation="
        http://www.osgi.org/xmlns/blueprint/v1.0.0 
        http://www.osgi.org/xmlns/blueprint/v1.0.0/blueprint.xsd">
    
    <service ref="peopleDao" interface="com.example.data.PeopleDao" />
    <bean id="peopleDao" class="com.example.data.impl.PeopleDaoImpl">
     <jpa:context unitname="peopleDb" property="entityManager" />
     <tx:transaction method="*" value="Required"/>
    </bean>
    
    <bean id="dataSource" class="org.hsqldb.jdbc.JDBCDataSource">
       <property name="url" value="jdbc:hsqldb:mem:peopleDb"/>
    </bean>
    
    <service ref="dataSource" interface="javax.sql.DataSource"> 
        <service-properties> 
            <entry key="osgi.jndi.service.name" value="peopleDb" /> 
        </service-properties> 
    </service>     
</blueprint>

One thing to keep in mind: the application doesn't need to create JPA 2.1's entity manager: the OSGi runtime is able do that and inject it everywhere it is required, driven by jpa:context declarations. Consequently, tx:transaction instructs the runtime to wrap the selected service methods inside transaction.

Now, when the last service PeopleDao is exposed, we are ready to deploy our modules with Apache Karaf 3.0.1. It is quite easy to do in three steps:

  • run the Apache Karaf 3.0.1 container
    bin/karaf (or bin\karaf.bat on Windows)
    
  • execute following commands from the Apache Karaf 3.0.1 shell:
    feature:repo-add cxf 3.0.1
    feature:install http cxf jpa openjpa transaction jndi jdbc
    install -s mvn:org.hsqldb/hsqldb/2.3.2
    install -s mvn:com.fasterxml.jackson.core/jackson-core/2.4.0
    install -s mvn:com.fasterxml.jackson.core/jackson-annotations/2.4.0
    install -s mvn:com.fasterxml.jackson.core/jackson-databind/2.4.0
    install -s mvn:com.fasterxml.jackson.jaxrs/jackson-jaxrs-base/2.4.0
    install -s mvn:com.fasterxml.jackson.jaxrs/jackson-jaxrs-json-provider/2.4.0
    
  • build our modules and copy them into Apache Karaf 3.0.1's deploy folder (while container is still running):
    mvn clean package
    cp module*/target/*jar apache-karaf-3.0.1/deploy/
    
When you run the list command in the Apache Karaf 3.0.1 shell, you should see the list of all activated bundles (modules), similar to this one:

Where module-service, module-jax-rs and module-data correspond to the ones we are being developed. By default, all our Apache CXF 3.0.1 services will be available at base URL http://:8181/cxf/api/. It is easy to check by executing cxf:list-endpoints -f command in the Apache Karaf 3.0.1 shell.

Let us make sure our REST layer works as expected by sending couple of HTTP requests. Let us create new person:

curl http://localhost:8181/cxf/api/people -iX POST -d "firstName=Tom&lastName=Knocker&email=a@b.com"

HTTP/1.1 201 Created
Content-Length: 0
Date: Sat, 09 Aug 2014 15:26:17 GMT
Location: http://localhost:8181/cxf/api/people/a@b.com
Server: Jetty(8.1.14.v20131031)

And verify that person has been created successfully:

curl -i http://localhost:8181/cxf/api/people

HTTP/1.1 200 OK
Content-Type: application/json
Date: Sat, 09 Aug 2014 15:28:20 GMT
Transfer-Encoding: chunked
Server: Jetty(8.1.14.v20131031)

[{"email":"a@b.com","firstName":"Tom","lastName":"Knocker"}]

Would be nice to check if database has the person populated as well. With Apache Karaf 3.0.1 shell it is very simple to do by executing just two commands: jdbc:datasources and jdbc:query peopleDb "select * from people".

Awesome! I hope this quite introductory blog post opens yet another piece of interesting technology you may use for developing robust, scalable, modular and manageable software. We have not touched many, many things but these are here for you to discover. The complete source code is available on GitHub.

Note to Hibernate 4.2.x / 4.3.x users: unfortunately, in the current release of Apache Karaf 3.0.1 the Hibernate 4.3.x does work properly at all (as JPA 2.1 is not yet supported) and, however I have managed to run with Hibernate 4.2.x, the container often refused to resolve the JPA-related dependencies.

Saturday, April 20, 2013

Fault Injection with Byteman and JUnit: do even more to ensure robustness of your applications

The time when our applications lived in isolation have passed long-long ago. Nowadays applications are a very complicated beasts talking to each other using myriads of APIs and protocols, storing data in traditional or NoSQL databases, sending messages and events over the wire ...

How often did you think about what will happen if, for example, a database goes down when your application is actively querying it? Or some API endpoint suddenly starts to refuse connection? Wouldn't it be nice to have such accidents covered as part of your test suite? That's what fault injection and Byteman framework are about.

As an example, we will build a realistic, full-blown Spring application which uses Hibernate/JPA to access MySQL database and manages customers. As part of application's JUnit integration test suite, we will include three kind of test cases:

  • store / find a customer
  • store customer and try to query database when it's down (fault simulation)
  • store customer and a database query times out (fault simulation)

There are only two preconditions for application to run on your local development box:

  • MySQL server is installed and has customers database
  • Oracle JDK is installed and JAVA_HOME environment variable points to it
That's being said, we are ready to go.

First, let's describe our domain model which consists from single class Customer with id and single property name. It looks as simple as that:

package com.example.spring.domain;

import java.io.Serializable;

import javax.persistence.Column;
import javax.persistence.Entity;
import javax.persistence.GeneratedValue;
import javax.persistence.Id;
import javax.persistence.Table;

@Entity
@Table( name = "customers" )
public class Customer implements Serializable{
    private static final long serialVersionUID = 1L;

    @Id
    @GeneratedValue
    @Column(name = "id", unique = true, nullable = false)
    private long id;
 
    @Column(name = "name", nullable = false)
    private String name;
  
    public Customer() {
    }
 
    public Customer( final String name ) {
        this.name = name;
    }

    public long getId() {
        return this.id;
    }

    protected void setId( final long id ) {
        this.id = id;
    }
 
    public String getName() {
        return this.name;
    }

    public void setName( final String name ) {
        this.name = name;
    }
}

For simplicity, the servicing layer is mixed with data access layer and calls database directly. Here is our CustomerService implementation:

package com.example.spring.services;

import javax.persistence.EntityManager;
import javax.persistence.PersistenceContext;

import org.springframework.stereotype.Service;
import org.springframework.transaction.annotation.Transactional;

import com.example.spring.domain.Customer;

@Service
public class CustomerService {
    @PersistenceContext private EntityManager entityManager;
 
    @Transactional( readOnly = true )
    public Customer find( long id ) {
        return this.entityManager.find( Customer.class, id );
    }

    @Transactional( readOnly = false )
    public Customer create( final String name ) {
        final Customer customer = new Customer( name );
        this.entityManager.persist(customer);
        return customer;
    }
 
    @Transactional( readOnly = false )
    public void deleteAll() {
        this.entityManager.createQuery( "delete from Customer" ).executeUpdate();
    }
}

And lastly, the Spring application context which defines data source and transaction manager. A small note here: as we won't introduce data access layer (@Repository) classes, in order for Spring to perform exception translation properly we define PersistenceExceptionTranslationPostProcessor instance to post-process service classes (@Service). Everything else should be very familiar.

package com.example.spring.config;

import java.util.Properties;

import javax.sql.DataSource;

import org.hibernate.dialect.MySQL5InnoDBDialect;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.ComponentScan;
import org.springframework.context.annotation.Configuration;
import org.springframework.dao.annotation.PersistenceExceptionTranslationPostProcessor;
import org.springframework.jdbc.datasource.DriverManagerDataSource;
import org.springframework.orm.jpa.JpaTransactionManager;
import org.springframework.orm.jpa.LocalContainerEntityManagerFactoryBean;
import org.springframework.orm.jpa.vendor.Database;
import org.springframework.orm.jpa.vendor.HibernateJpaVendorAdapter;
import org.springframework.stereotype.Service;
import org.springframework.transaction.PlatformTransactionManager;
import org.springframework.transaction.annotation.EnableTransactionManagement;

import com.example.spring.services.CustomerService;

@EnableTransactionManagement
@Configuration
@ComponentScan( basePackageClasses = CustomerService.class )
public class AppConfig {
    @Bean
    public PersistenceExceptionTranslationPostProcessor exceptionTranslationPostProcessor() {
        final PersistenceExceptionTranslationPostProcessor processor = new PersistenceExceptionTranslationPostProcessor();
        processor.setRepositoryAnnotationType( Service.class );
        return processor;
    }

    @Bean
    public HibernateJpaVendorAdapter hibernateJpaVendorAdapter() {
        final HibernateJpaVendorAdapter adapter = new HibernateJpaVendorAdapter();

        adapter.setDatabase( Database.MYSQL );
        adapter.setShowSql( false );

        return adapter;
    }

    @Bean
    public LocalContainerEntityManagerFactoryBean entityManager() throws Throwable {
        final LocalContainerEntityManagerFactoryBean entityManager = new LocalContainerEntityManagerFactoryBean();
     
        entityManager.setPersistenceUnitName( "customers" );
        entityManager.setDataSource( dataSource() );
        entityManager.setJpaVendorAdapter( hibernateJpaVendorAdapter() );

        final Properties properties = new Properties();
        properties.setProperty("hibernate.dialect", MySQL5InnoDBDialect.class.getName());
        properties.setProperty("hibernate.hbm2ddl.auto", "create-drop" );
        entityManager.setJpaProperties( properties );

        return entityManager;
    }

    @Bean
    public DataSource dataSource() {
        final DriverManagerDataSource dataSource = new DriverManagerDataSource();

        dataSource.setDriverClassName( com.mysql.jdbc.Driver.class.getName() );
        dataSource.setUrl( "jdbc:mysql://localhost/customers?enableQueryTimeouts=true" );
        dataSource.setUsername( "root" );
        dataSource.setPassword( "" );

        return dataSource;
    }

    @Bean
    public PlatformTransactionManager transactionManager() throws Throwable {
        return new JpaTransactionManager( this.entityManager().getObject() );
    }
}

Now let's add a simple JUnit test case to verify our Spring application actually works as expected. Before doing that, the database customers should be created:

> mysql -u root
mysql> create database customers;
Query OK, 1 row affected (0.00 sec)

And here is a CustomerServiceTestCase which for now has single test to create the customer and verify it's actually has been created.

package com.example.spring;

import static org.hamcrest.CoreMatchers.notNullValue;
import static org.junit.Assert.assertThat;

import javax.inject.Inject;

import org.junit.After;
import org.junit.Test;
import org.junit.runner.RunWith;
import org.springframework.test.context.ContextConfiguration;
import org.springframework.test.context.junit4.SpringJUnit4ClassRunner;
import org.springframework.test.context.support.AnnotationConfigContextLoader;

import com.example.spring.config.AppConfig;
import com.example.spring.domain.Customer;
import com.example.spring.services.CustomerService;

@RunWith( SpringJUnit4ClassRunner.class )
@ContextConfiguration(loader = AnnotationConfigContextLoader.class, classes = { AppConfig.class } )
public class CustomerServiceTestCase {
    @Inject private CustomerService customerService; 
 
    @After
    public void tearDown() {
        customerService.deleteAll();
    }

    @Test
    public void testCreateCustomerAndVerifyItHasBeenCreated() throws Exception {
        Customer customer = customerService.create( "Customer A" );
        assertThat( customerService.find( customer.getId() ), notNullValue() );
    }
}

That looks simple and straightforward. Now, let's think about scenario when customer creation succeeded but find fails because of query timeout. To do that, we need a help from Byteman.

In short, Byteman is bytecode manipulation framework. It's a Java agent implementation which runs with JVM (or attaches to it) and modifies running application bytecode as such changing its behavior. Byteman has a very good documentation and own rich set of rule definitions to perform mostly everything developer can come up with. Also, it has pretty good integration with JUnit framework. On that subject, Byteman tests are supposed to be run with @RunWith( BMUnitRunner.class ), but we already using @RunWith( SpringJUnit4ClassRunner.class ) and JUnit doesn't allow multiple test runners to be specified. Looks like a problem unless you are familiar with JUnit @Rule mechanics. It turns out that converting BMUnitRunner to JUnit rule is quite easy task:

package com.example.spring;

import org.jboss.byteman.contrib.bmunit.BMUnitRunner;
import org.junit.rules.MethodRule;
import org.junit.runners.model.FrameworkMethod;
import org.junit.runners.model.InitializationError;
import org.junit.runners.model.Statement;

public class BytemanRule extends BMUnitRunner implements MethodRule {
    public static BytemanRule create( Class< ? > klass ) {
        try {
            return new BytemanRule( klass ); 
        } catch( InitializationError ex ) { 
            throw new RuntimeException( ex ); 
        }
    }
 
    private BytemanRule( Class klass ) throws InitializationError {
        super( klass );
    }
 
    @Override
    public Statement apply( final Statement statement, final FrameworkMethod method, final Object target ) {
        Statement result = addMethodMultiRuleLoader( statement, method ); 
  
        if( result == statement ) {
            result = addMethodSingleRuleLoader( statement, method );
        }

        return result;
    }
}

And JUnit @Rule injection is as simple as that:

@Rule public BytemanRule byteman = BytemanRule.create( CustomerServiceTestCase.class );

Easy, right? The scenario we mentioned before could be rephrased a bit: when JDBC statement to select from 'customers' table is executed, we should fail with timeout exception. Here is how it looks like as JUnit test case with additional Byteman annotations:

    @Test( expected = DataAccessException.class )
    @BMRule(
        name = "introduce timeout while accessing MySQL database",
        targetClass = "com.mysql.jdbc.PreparedStatement",
        targetMethod = "executeQuery",
        targetLocation = "AT ENTRY",
        condition = "$0.originalSql.startsWith( \"select\" ) && !flagged( \"timeout\" )",
        action = "flag( \"timeout\" ); throw new com.mysql.jdbc.exceptions.MySQLTimeoutException( \"Statement timed out (simulated)\" )"
    )
    public void testCreateCustomerWhileDatabaseIsTimingOut()  {
        Customer customer = customerService.create( "Customer A" );
        customerService.find( customer.getId() );
    }

We could read it like this: "When someone calls executeQuery method of PreparedStatement class and query starts with 'SELECT' than MySQLTimeoutException will be thrown, and it should happen only once (controlled by timeout flag)". Running this test case prints stacktrace in a console and expects DataAccessException to be thrown:

com.mysql.jdbc.exceptions.MySQLTimeoutException: Statement timed out (simulated)
 at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method) ~[na:1.7.0_21]
 at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:57) ~[na:1.7.0_21]
 at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45) ~[na:1.7.0_21]
 at java.lang.reflect.Constructor.newInstance(Constructor.java:525) ~[na:1.7.0_21]
 at org.jboss.byteman.rule.expression.ThrowExpression.interpret(ThrowExpression.java:231) ~[na:na]
 at org.jboss.byteman.rule.Action.interpret(Action.java:144) ~[na:na]
 at org.jboss.byteman.rule.helper.InterpretedHelper.fire(InterpretedHelper.java:169) ~[na:na]
 at org.jboss.byteman.rule.helper.InterpretedHelper.execute0(InterpretedHelper.java:137) ~[na:na]
 at org.jboss.byteman.rule.helper.InterpretedHelper.execute(InterpretedHelper.java:100) ~[na:na]
 at org.jboss.byteman.rule.Rule.execute(Rule.java:682) ~[na:na]
 at org.jboss.byteman.rule.Rule.execute(Rule.java:651) ~[na:na]
 at com.mysql.jdbc.PreparedStatement.executeQuery(PreparedStatement.java) ~[mysql-connector-java-5.1.24.jar:na]
 at org.hibernate.engine.jdbc.internal.ResultSetReturnImpl.extract(ResultSetReturnImpl.java:56) ~[hibernate-core-4.2.0.Final.jar:4.2.0.Final]
 at org.hibernate.loader.Loader.getResultSet(Loader.java:2031) [hibernate-core-4.2.0.Final.jar:4.2.0.Final]

Looks good, what about another scenario: customer creation succeeded but find fails because the database went down? This one is a bit more complicated but easy to do anyway, let's take a look:

@Test( expected = CannotCreateTransactionException.class )
@BMRules(
    rules = {
        @BMRule(
            name="create countDown for AbstractPlainSocketImpl",
            targetClass = "java.net.AbstractPlainSocketImpl",
            targetMethod = "getOutputStream",
            condition = "$0.port==3306",
            action = "createCountDown( \"connection\", 1 )"
        ),
        @BMRule(
            name = "throw IOException when trying to execute 2nd query to MySQL",
            targetClass = "java.net.AbstractPlainSocketImpl",
            targetMethod = "getOutputStream",
            condition = "$0.port==3306 && countDown( \"connection\" )",
            action = "throw new java.io.IOException( \"Connection refused (simulated)\" )"
        )
    }
)
public void testCreateCustomerAndTryToFindItWhenDatabaseIsDown() {
    Customer customer = customerService.create( "Customer A" );
    customerService.find( customer.getId() );
}

Let me explain what's going on here. We would like to sit on socket level and actually control the communication as close to network as we can, not on JDBC driver level. That's why we instrumenting AbstractPlainSocketImpl. We also know that MySQL's default port is 3306 so we are instrumenting only sockets opened on this port. Another fact, we know that first created socket corresponds to customer creation and we should let it go through. But second one corresponds to find and must fail. The createCountDown named "connection" serves this purposes: the first call goes through (latch doesn't count to zero yet) but second call triggers MySQLTimeoutException exception. Running this test case prints stacktrace in a console and expects CannotCreateTransactionException to be thrown:

Caused by: java.io.IOException: Connection refused (simulated)
 at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method) ~[na:1.7.0_21]
 at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:57) ~[na:1.7.0_21]
 at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45) ~[na:1.7.0_21]
 at java.lang.reflect.Constructor.newInstance(Constructor.java:525) ~[na:1.7.0_21]
 at org.jboss.byteman.rule.expression.ThrowExpression.interpret(ThrowExpression.java:231) ~[na:na]
 at org.jboss.byteman.rule.Action.interpret(Action.java:144) ~[na:na]
 at org.jboss.byteman.rule.helper.InterpretedHelper.fire(InterpretedHelper.java:169) ~[na:na]
 at org.jboss.byteman.rule.helper.InterpretedHelper.execute0(InterpretedHelper.java:137) ~[na:na]
 at org.jboss.byteman.rule.helper.InterpretedHelper.execute(InterpretedHelper.java:100) ~[na:na]
 at org.jboss.byteman.rule.Rule.execute(Rule.java:682) ~[na:na]
 at org.jboss.byteman.rule.Rule.execute(Rule.java:651) ~[na:na]
 at java.net.AbstractPlainSocketImpl.getOutputStream(AbstractPlainSocketImpl.java) ~[na:1.7.0_21]
 at java.net.PlainSocketImpl.getOutputStream(PlainSocketImpl.java:214) ~[na:1.7.0_21]
 at java.net.Socket$3.run(Socket.java:915) ~[na:1.7.0_21]
 at java.net.Socket$3.run(Socket.java:913) ~[na:1.7.0_21]
 at java.security.AccessController.doPrivileged(Native Method) ~[na:1.7.0_21]
 at java.net.Socket.getOutputStream(Socket.java:912) ~[na:1.7.0_21]
 at com.mysql.jdbc.MysqlIO.(MysqlIO.java:330) ~[mysql-connector-java-5.1.24.jar:na]

Great! The possibilities Byteman provides for different fault simulations are enormous. Carefully adding test suites to verify how application reacts to erroneous conditions greatly improves application robustness and resiliency to failures. Bunch of thanks to Byteman guys!

Please find the complete project on GitHub.