Copyright UTS Faculty of Information Technology 2002 EJB2 EJB-3. Use Design Patterns 1 to re-use well known programming techniques.

Transcription

1 Advanced Java Programming (J2EE) Enterprise Java Beans (EJB) Part 2 Entity & Message Beans Chris Wong chw@it.uts.edu.au (based on prior class notes & Sun J2EE tutorial) Enterprise Java Beans Last lesson we looked at: Introduction to EJB EJB basics EJB Architecture Session Beans EJB Clients EJB Development Process This lesson will cover Entity Beans Message Driven Beans EJB issues Copyright UTS 2002 EJB2 EJB-1 Copyright UTS 2002 EJB2 EJB-2 -Interlude- Earlier we discussed why we use EJB s. EJB containers provide much of the plumbing that traditionally programmers had to develop and consider themselves. We still need good programming practices to use EJB s well. Use Design Patterns 1 to re-use well known programming techniques. Good references: 1 Gamma, E., R. Helm, R. Johnson, and J. Vlissides. Design Patterns: Elements of Reusable Object-Oriented Software. Addison-Wesley Marinescu, Floyd EJB design patterns Wiley 2002 Copyright UTS 2002 EJB2 EJB-3 The Session Façade pattern This is the most common design pattern used with EJB s You partition business logic to minimise dependencies between client and server This optimises network traffic and executes as one transaction. Idea is to wrap an entity EJB with a session EJB Only one interface between client and server Completely hides internal implementation from client Copyright UTS 2002 EJB2 EJB-4 Traditional client/server Session Façade Marinescu, Floyd EJB design patterns Wiley 2002 Fig 1.2 Copyright UTS 2002 EJB2 EJB-5 Marinescu, Floyd EJB design patterns Wiley 2002 Fig 1.2 Copyright UTS 2002 EJB2 EJB-6

2 Client (e.g. servlet) Session & Entity Beans EJB container BankTeller - deposit() - withdraw() BankManager - openaccount() Session Beans ("actions") Account - acctnum = getbalance() - setbalance() Account - acctnum = getbalance() - setbalance() Account - acctnum = getbalance() - setbalance() Entity Beans ("data") DBMS Enterprise Java Beans Last lesson we looked at: Introduction to EJB EJB basics EJB Architecture Session Beans EJB Clients EJB Development Process This lesson will cover Entity Beans Message Driven Beans EJB issues Copyright UTS 2002 EJB2 EJB-7 Copyright UTS 2002 EJB2 EJB-8 Entity Beans Entity beans are EJBs that are created to encapsulate some data contained by the system (such as a row in the database or, more generally, to an entry that exists in persistent storage) Entity Beans In addition to the usual EJB implementation, home interface and remote interface, Entity beans also have special primary key classes defined for them that relate to the primary keys of an associated entity stored in the database Entity beans can be thought of as the nouns of a particular problem domain that they are implemented to solve The data can be created, searched for (found) or removed by clients The data can also be retrieved and updated by clients Entity beans are very different from Session beans. Entity beans: can be used concurrently by several clients are long-lived they are intended to exist beyond the lifetime of a client will survive server crashes directly represent data in a database Copyright UTS 2002 EJB2 EJB-9 Copyright UTS 2002 EJB2 EJB-10 Entity Beans scalability Because clients could potentially hold many entity references as a result of a find operation, many effective EJB designs will only provide access to entity beans via session beans and limit how many entity bean handles are returned to the client Scalability of the system can become compromised if too many remote entity bean references are handed out to clients Articles describing various patterns for using entity beans are listed in the references Copyright UTS 2002 EJB2 EJB-11 Entity Beans container's role Entity beans assume that more than one client will use them concurrently. Interactions with the bean are moderated by the container so as to guarantee the integrity of the underlying data The container may accomplish this by either: queuing client requests one at a time -- or -- creating an instance of the bean for each client and relying on the underlying database to deal with synchronisation issues Remember, session beans do not support concurrent access. A stateful bean is an extension of the client and stateless beans don t maintain state Entity beans (unlike Session beans) exist until they are explicitly destroyed: There is no time out period associated with Entity beans They are guaranteed to survive orderly server restarts They are guaranteed to survive server crashes Copyright UTS 2002 EJB2 EJB-12

3 Entity Beans persistence Entity beans represent data in an underlying database. Their lifetime matches the lifetime of the underlying data Entity beans can be removed in one of two ways: A client explicitly invokes the remove() method of the entity bean, or The data for the entity is deleted from the underlying database Entity beans typically map to data in an underlying database Usually to a single row in a table. The instance variables in the Entity bean map directly to the fields in the table. Entity Beans BMP vs. CMP Bean Managed Persistence (BMP) bean developer takes control of bean s persistence eg: use JDBC and SQL directly Container Managed Persistence (CMP) container manages the process of saving and restoring the state of beans. Containers generate the code based on the information in the deployment descriptor for the bean. 2 Types of Entity beans Bean Managed Persistence (BMP) Container Managed Persistence (CMP) Copyright UTS 2002 EJB2 EJB-13 Copyright UTS 2002 EJB2 EJB-14 Entity Beans primary keys Each Entity bean instance has a primary key. A primary key is: A value (or combination of values, called compound primary key) that uniquely identifies the entity Implemented as a Java class NOT a primitive (e.g., java.lang.integer vs int) You can write your own class to represent a primary key Must be serializable ie: implement the java.io.serializable Must have a no argument constructor May have a constructor with the compound key values Must implement a hashcode() and equals() method Easiest way is to use code generator to do this eg: Eclipse -> Source -> generate constructor from fields Source -> generate hashcode() and equals() Entity Beans primary keys public class CustomerPK implements Serializable { public int custid; public String surname; public CustomerPK(int custid, String surname) { ; // constructor this.custid = custid; this.surname = surname; public int hashcode() { final int PRIME = 31; int result = 1; result = PRIME * result + custid; result = PRIME * result + ((surname == null)? 0 : surname.hashcode()); return result; public boolean equals(object obj) { if (custid!= other.custid) return false; else if (!surname.equals(other.surname)) return false; return true; Copyright UTS 2002 EJB2 EJB-15 Copyright UTS 2002 EJB2 EJB-16 Entity Beans lifecycle Does not exist newinstance() unsetentitycontext() setentitycontext(ec) ejbcreate(args) Pooled ejbfind<method>() ejbpostcreate(args) ejbremove() ejbactivate() ejbpassivate() Ready ejbstore() ejbload() Entity Beans lifecycle An entity bean can have one of three states: Does-not -exist state not yet been instantiated. This state provides a beginning and an end for the life cycle of a beans instance Pooled state bean has been instantiated by the container but not associated with an EJB object Ready state it has been associated with an EJB object and is ready to respond to business method invocations Business method(s) Copyright UTS 2002 EJB2 EJB-17 Copyright UTS 2002 EJB2 EJB-18

4 Entity Beans implementation Implementation of Entity beans is similar to that of session beans create a home interface, a remote interface and the bean implementation class The developer must also create the primary key class used to find instances of an existing bean. This primary key class is used as a parameter to the ejbfindbyprimarykey() method Must be implemented by all entity beans Entity Beans methods Entity beans are created via the ejbcreate() method this corresponds to an INSERT on the database Existing entity beans are loaded into the container via the ejbfindbyprimarykey() method (and any other ejbfindxxxx() methods that may be defined) these correspond to SELECT on the database Entity beans are removed via the ejbremove() method this corresponds to a DELETE on the database All entity bean implementation classes must implement the javax.ejb.entitybean interface (which extends javax.ejb.enterprisebean) The methods that need to be implemented depend on the persistence model you are implementing BMP or CMP. When using CMP the container will implement most of the methods for you The container will call the ejbload() and ejbstore() methods when it needs to synchronise the state of the bean with the underlying database these correspond to SELECT and UPDATE See the EJB specification and your text book for a complete reference of all the methods that exist on entity beans and the relevant configuration parameters that are defined in the deployment descriptor Copyright UTS 2002 EJB2 EJB-19 Copyright UTS 2002 EJB2 EJB-20 CMP vs. BMP CMP Major change in EJB 2.0 (not compatible with EJB 1.1); however, most containers support both versions Don t have to code the database access calls in the entity bean this makes bean portable across many database servers Many consider CMP is the default choice BMP Given all the benefits of CMP, why someone bother with BMP? gives you more flexibility can better accommodate a complex or unusual set of data EJB 1.1 only supported BMP Cons more work is required ties the bean to a specific database type and structure Copyright UTS 2002 EJB2 EJB-21 Coding differences between Persistent Types Difference CMP BMP Class definition Abstract Not abstract Database access calls Generated by tools Coded by developer Persistent state Access methods for persistent and relation fields findbyprimarykey method Customized finder methods Represented by virtual persistent fields Required Coded as instance variables None Handled by container Coded by developer Handled by container Coded by developer but the developer must define the EJB QL Select methods Handled by container None Return value of Should be null Must be the primary ejbcreate key Copyright UTS 2002 EJB2 EJB-22 EJB 2.0 CMP Entity Bean (1) package bank; import javax.ejb.*; public class abstract CustomerBean implements EntityBean { SessionContext sessioncontext; // "Standard" session bean methods public void setsessioncontext (SessionContext sc) { this.sessioncontext = sc; public void ejbremove() { public void ejbactivate() { public void ejbpassivate() { // continued on next page EJB 2.0 CMP Entity Bean (2) // continued from previous page // Argument list must match create() method in home interface public Integer ejbcreate(integer id) { setid(id); // call the usual setter method return null; // Now the business logic methods (as usual) // Abstract get/set methods public abstract void setid(integer id); public abstract Integer getid(); public abstract void setname (String name); public abstract String getname (); Copyright UTS 2002 EJB2 EJB-23 Copyright UTS 2002 EJB2 EJB-24

5 EJB 2.0 CMP Entity Bean (3) Entity EJB s have additional parameters in the ejb-jar.xml Deployment Descriptor (DD) <ejb-jar> <enterprise-beans> <entity> // usual EJB stuff <ejb-name> Customer </ejb-name> <home> bank.customerbeanhome </home> <remote> bank.customerbean </remote> <ejb-class> bank.customerbean </ejb-class> // set as CMP <persistence-type> Container </persistence-type> // define primary key as Integer type <prim-key-class> java.lang.integer </prim-key-class> <reentrant> False </reentrant> // important! We are using CMP 2.0 <cmp-version> 2.x </cmp-version> EJB 2.0 CMP Entity Bean (4) // continued from prior page <abstract-schema-name>customerbean</abstract-schema-name> // define the fields these correspond to a column in the table <cmp-field> <field-name> Id </field-name> </cmp-field> <cmp-field> <field-name> Name </field-name> </cmp-field> // define the primary key the container generates findprimarykey() method! <primkey-field> Id </primkey-field> </entity> </enterprise-beans> & so on Copyright UTS 2002 EJB2 EJB-25 Copyright UTS 2002 EJB2 EJB-26 EJB 2.0 CMP Entity Bean (5) You will also need 2 extra deployment descriptors weblogic-ejb-jar.xml for weblogic specific parameters weblogic-cmp-rdbms-jar.xml how weblogic maps the datasource name and table column names to the CMP bean. See lab for details on this. Copyright UTS 2002 EJB2 EJB-27 Entity Beans Summary Entity beans represent persistent data Two types of entity beans: CMP and BMP Entity beans must have a primary key class defined BMP must provide database statements for: ejbcreate(), ejbremove(), ejbfindbyprimarykey(), ejbfindxxx() etc. More deployment descriptor (ejb-jar.xml) tags: CMP fields (only for CMP EJBs) Primary key class Abstract schema name (only for EJB 2.0) Still need extra deployment descriptor for CMP entity beans (depending on the Application Server) to map bean to database table. weblogic-cmp-rdbms-jar.xml for WebLogic. Copyright UTS 2002 EJB2 EJB-28 EJB Exception Types Note: javax.ejb.ejbobject and javax.ejb.ejbhome are subclasses of java.rmi.remote. ie: your Home & Implementation will throw the java.rmi.remoteexception from each of their methods (and you will need to catch them in your client!) EJB Exception Types (cont) There are other exceptions thrown by EJBs eg: CreateException, DuplicateKeyException FinderException, ObjectNotFoundException RemoveException EJBException, NoSuchEntityException EJB exceptions are depicted in the following figure: Perrone et al Building Java Enterprise Systems with J2EE Sams 2000 fig 36.2 Copyright UTS 2002 EJB2 EJB-29 Copyright UTS 2002 EJB2 EJB-30

6 EJB Exception Types (cont) EJBException thrown by an EJB when an application specific method cannot be completed NoSuchEntityException thrown by an EJB when an application specific method cannot be completed because a particular Entity bean does not exist CreateException thrown when an EJB cannot be created (invalid arguments) DuplicateKeyException a subclass of CreateException, thrown when a particular Entity bean cannot be created because objects with the same key already exist RemoveException thrown when a particular EJB cannot be removed FinderException thrown by all query methods (find or select) when an application error (business-logic) occurs ObjectNotFoundException thrown when a singular EJB cannot be found Copyright UTS 2002 EJB2 EJB-31 Local Client API Only remote interfaces in EJB 1.1 EJB 2.0 introduced local component interfaces Local Interface Defines the bean s business methods that can be used by other beans co-located in the same EJB container Allows beans to interact without the network overhead Objects passed by reference rather than copied Local home interface Defines the bean s life cycle methods that can be used by other beans co-located in the same EJB container Similar to Remote Client API but less complicated Copyright UTS 2002 EJB2 EJB-32 package bank; import java.rmi.*; import javax.ejb.*; Local Interface package bank; import java.rmi.*; import javax.ejb.*; Local Home Interface interface LocalBankTeller extends EJBLocalObject { String sayhello(); void deposit(string acctnum, Float amount); void withdraw(string acctnum, Float amount); Notes: Business logic methods for clients in the same container Note that none of the methods throw RemoteException Identical to remote interface except for a couple of differences Copyright UTS 2002 EJB2 EJB-33 interface LocalBankTellerHome extends EJBLocalHome { public BankTeller create () throws CreateException; Notes: Note that none of the methods throw RemoteException Copyright UTS 2002 EJB2 EJB-34 <?xml version="1.0"?> DD for Local Client <!DOCTYPE ejb-jar PUBLIC '-//Sun Microsystems, Inc.//DTD Enterprise JavaBeans 2.0//EN' ' <ejb-jar> <enterprise-beans> <session> <ejb-name>banktellerejb</ejb-name> <home>bank.banktellerhome</home> <remote>bank.bankteller</remote> <local-home>bank.localbanktellerhome</local-home> <local>bank.localbankteller</local> <ejb-class>bank.banktellerbean</ejb-class> <session-type>stateless</session-type> <transaction-type>container</transaction-type> </session> </enterprise-beans> </ejb-jar> Copyright UTS 2002 EJB2 EJB-35 Local Client API (cont d) When to use Local Client API When EJBs accessing each other from within the same container Are local component interfaces necessary? Some argue that local interfaces added more complexity Most EJB1.1 vendors were already optimizing the Remote Client API for co-located beans You are locked into a single JVM However, local interfaces are here to stay Container Managed Relationships (EJB 2.0, CMP) need local interfaces Not covered in this subject Copyright UTS 2002 EJB2 EJB-36

7 EJB 2.0 CMP: EJB QL No standard language prior to EJB 2.0 EJB QL is a language similar to SQL-99 EJB QL statements are specified in the deployment descriptor and the container will generate SQL EJB QL is an attempt to map Java objects and rows/columns Use EJB QL to: Implement finders Implement select methods Navigate relationships EJB 2.0 CMP: EJB QL (2) Find methods are always declared in home interface public interface CustomerHome extends EJBHome { // findbyprimarykey() is automatically generated by container public Customer findbyprimarykey(integer pk) throws FinderException, RemoteException; // findbyname() needs a query-method declared in the DD public Customer findbyname(string firstname, lastname) throws FinderException, RemoteException; See for a good tutorial on EJB-QL Copyright UTS 2002 EJB2 EJB-37 Copyright UTS 2002 EJB2 EJB-38 EJB 2.0 CMP: EJB QL (3) Query declaration in the bean s DD <query> <query-method> <method-name>findbyname</method-name> <method-params> <method-param>java.lang.string</method-params> <method-param>java.lang.string</method-params> </method-params> </query-method> <ejb-ql> <![CDATA[ SELECT OBJECT(c) FROM Customer c WHERE c.lastname =?1 AND c.firstname =?2 ]]> </ejb-ql> </query> Enterprise Java Beans Last lesson we looked at: Introduction to EJB EJB basics EJB Architecture Session Beans EJB Clients EJB Development Process This lesson will cover Entity Beans Message Driven Beans EJB issues // cmp name not shown Copyright UTS 2002 EJB2 EJB-39 Copyright UTS 2002 EJB2 EJB-40 Message Driven Beans Message-driven beans (MDBs) were introduced as part of EJB 2.0 MDBs represent the integration of EJBs with JMS (the Java Message Service) they handle incoming JMS messages JMS is an API for enterprise messaging by Sun Microsystems. JMS is not a messaging system itself; it s an abstraction of the interfaces and classes needed by clients when communicating with messaging systems. MDBs act as a standard JMS message consumer. The Message-driven bean is a stateless component that is invoked by the EJB container as a result of receiving messages from a JMS Queue or Topic The MDB then performs business logic based on the message contents Message Driven Beans Unlike session or entity beans, MDBs have no home or remote interface, and therefore cannot be directly accessed by internal or external clients Clients interact with MDBs only indirectly, by sending a JMS message to a JMS Queue or Topic The EJB container automatically creates and removes MDB instances as needed to process incoming messages The goal of the MDB model is to ensure that developing an EJB that is asynchronously invoked to handle the processing of incoming JMS messages is as easy as developing the same functionality in any other JMS MessageListener Copyright UTS 2002 EJB2 EJB-41 Copyright UTS 2002 EJB2 EJB-42

8 Message Driven Beans Queue or Topic is defined in the deployment descriptor message selector is also defined in DD - restrictive MDB must implement javax.jms.messagelistener and javax.ejb.messagedrivenbean onmessage() must not throw JMS or application exceptions; instead they must be caught or handled. At least, they could be re-thrown as EJBException MessageDrivenBean specifies two methods: ejbremove() is invoked just before the bean is removed setmessagedrivencontext(messagedrivenbean ctx) sets the context for the bean ejbcreate() needs to be defined as well MDB Life Cycle newinstance() setmessagecontext(mdc) ejbcreate() onmessage() Does not exist Ready pool ejbremove() Copyright UTS 2002 EJB2 EJB-43 Copyright UTS 2002 EJB2 EJB-44 Example Message-driven bean import javax.ejb.*; import javax.jms.*; public class MessageTraderBean implements MessageDrivenBean, MessageListener { public void ejbcreate () throws CreateException { public void ejbremove() { public void setmessagedrivencontext(messagedrivencontext ctx) { public void onmessage(message msg) { try { // process message catch (EJBException ex) { Copyright UTS 2002 EJB2 EJB-45 Example Message-driven bean ejb-jar.xml <!DOCTYPE ejb-jar PUBLIC "-//Sun Microsystems, Inc.//DTD Enterprise JavaBeans 2.0//EN" " <ejb-jar> <enterprise-beans> <message-driven> <ejb-name>examplemessagedriven</ejb-name> <ejb-class>examples.ejb20.message.messagetraderbean</ejb-class> <transaction-type>container</transaction-type> <message-driven-destination> <jms-destination-type>javax.jms.topic</jms-destination-type> </message-driven-destination> <security-identity> <run-as-specified-identity> <role-name>foo</role-name> </run-as-specified-identity> </security-identity> </message-driven> </enterprise-beans> </ejb-jar> Copyright UTS 2002 EJB2 EJB-46 Enterprise Java Beans Last lesson we looked at: Introduction to EJB EJB basics EJB Architecture Session Beans EJB Clients EJB Development Process This lesson will cover Entity Beans Message Driven Beans EJB issues EJB Clustering Issues Highly scalable, mission critical enterprise applications will inevitably require more than a single application server to deliver services with minimal downtime and maximum performance In these environments, clustering is used to guarantee availability and scale to meet the demands placed on the system A cluster is a group of application servers that transparently run your J2EE application as if it was a single entity Scaling is accomplished by adding extra machines to the cluster. Multiple machines provide redundancy that increases availability Copyright UTS 2002 EJB2 EJB-47 Copyright UTS 2002 EJB2 EJB-48

9 EJB Clustering Issues (cont) Clustering support isn t defined in the J2EE specification application server vendors are left to implement clustering in whatever manner they see fit Most J2EE application servers implement clustering around their implementation of JNDI. 1. Independent each app server has its own independent JNDI tree 2. Centralised a single, centralised JNDI tree is maintained for all app servers 3. Shared Global each app server shares a global JNDI tree with the cluster. All objects bound into the JNDI tree are broadcast to all servers in the cluster. This is the mechanism used by WebLogic EJB Clustering Issues (cont) You do have to be aware of how clustering is implemented in your application server when deploying a J2EE application to a cluster The relevant deployment properties will be container specific See the WebLogic server documentation for a description of how clustering of EJBs (and web applications) is configured Note that there are some limitations introduced when deploying to a cluster (particularly for stateful session beans) however these are still preferable to running a single application server with no redundancy or failover capability Copyright UTS 2002 EJB2 EJB-49 Copyright UTS 2002 EJB2 EJB-50 EJB Key Features Summary They provide a model for defining server-side components EJB Key Features Summary EJB key features (continued): They provide a model for defining distributed client interfaces to the services provided by these components They provide standard operations and semantics for allowing a container to create, destroy, allocate, persist and activate component instances They provide a standard model for defining a component that maintains a conversational session with a client, with session management handled by the container They provide a standard model for defining a component that encapsulates a data source (such as a database) entry, with objectrelational mapping handled by the container They provide a standard for defining configuration and deployment characteristics of a component, independent of its implementation They provide a standard model for declaratively defining the security attributes of a component They provide a standard model for declaratively defining the transactional attributes of a component They provide a standard component interface contract such that components can run in any EJB-compliant container/server that implements the standard interface contract Copyright UTS 2002 EJB2 EJB-51 Copyright UTS 2002 EJB2 EJB-52 EJB s - Cautions EJB 1.1 only had remote interfaces. This meant that clients had to use the network to connect to EJBs, even on the same physical server!! use Local interface instead. + new EJB-QL and selector methods available. But this means the session bean will run in the same Java VM as the entity bean! Alternatives? Java EE 5 has new persistence mechanism Based on Java 5 annotations Java Persistence API (JPA) EJB 3.0 based on JPA Use EJBs where appropriate EJB containers have high overhead (and cost $$). Sometimes using lightweight Web Containers such as Tomcat + JDBC/stored procedures is more appropriate. Java Data Objects (JDO) provides an alternative mechanism See You can also use open source products such as Hibernate (which has a JPA interface) or commercial products such as Oracle Toplink Copyright UTS 2002 EJB2 EJB-53 Copyright UTS 2002 EJB2 EJB-54

10 EJB 3.0 Just plain old Java Objects! Eg: BankTeller: Eg: Account EJB 3.0 public interface BankTellerRemote { public string sayhello(string BankTeller ) public class BankTeller implements BankTellerRemote { public string sayhello(string name) { return Hello, + name; Note: no XML, no housekeeping code! Copyright UTS (name= Accounts ( name = findbyacct, query= select Object(a) from Accounts where a.acct =?1 ) public class Account implements Serializable // next field is primary (name= Acct ) public Integer getacctnum() (name= Balance ) public float getbalance() { Copyright UTS 2002 EJB2 EJB-56 Summary The past two lessons have introduced the component model for Java enterprise applications Enterprise Java Beans We have touched on a number of aspects of EJB architecture and development You should now understand the different types of EJB, when they should be used, how they are developed and deployed, and how they are used from client applications EJBs are a fairly complex topic, and you will need to do additional study to fully understand the intricacies of EJB design and development Copyright UTS 2002 EJB2 EJB-57