Using Special Techniques | Sun Microsystems 820434310 guide

EJB Performance Tuning

source are going to be involved in a transaction. If a database participates in some distributed transactions, but mostly in local or single database transactions, it is advisable to register two separate JDBC resources and use the appropriate resource in the application.

Configure JDBC Resources as One-Phase Commit Resources

To improve performance of transactions involving multiple resources, the Application Server uses last agent optimization (LAO), which allows the configuration of one of the resources in a distributed transaction as a one-phase commit (1PC) resource. Since the overhead of multiple-resource transactions is much higher for a JDBC resource than a message queue, LAO substantially improves performance of distributed transactions involving one JDBC resource and one or more message queues. To take advantage of LAO, configure a JDBC resource as a 1PC resource. Nothing special needs to be done to configure JMS resources.

In global transactions involving multiple JDBC resources, LAO will still improve performance, however, not as much as for one JDBC resource. In this situation, one of the JDBC resources should be configured as 1PC, and all others should be configured as XA.

Use the Least Expensive Transaction Attribute

Set the following transaction attributes in the EJB deployment descriptor file (ejb-jar.xml). Options are listed from best performance to worst. To improve performance, choose the least expensive attribute that will provide the functionality your application needs:

1.NEVER

2.TX_NOTSUPPORTED

3.TX_MANDATORY

4.TX_SUPPORTS

5.TX_REQUIRED

6.TX_REQUIRESNEW

Using Special Techniques

Special performance-enhancing techniques are discussed in the following sections:

■“Version Consistency” on page 39

■“Request Partitioning” on page 41

Version Consistency

Note – The technique in section applies only to the EJB 2.1 architecture. In the EJB 3.0 architecture, use the Java Persistence API (JPA).

Chapter 2 • Tuning Your Application

39

Image 39

Sun Microsystems 820434310 manual Using Special Techniques, Configure Jdbc Resources as One-Phase Commit Resources

Contents

Sun Microsystems, Inc Network Circle Santa Clara, CA Sun GlassFish Enterprise Server 2.1 Performance Tuning Guide 090304@21990 Contents Tuning the Enterprise Server 100 Tuning for High-Availability Figures Page Tables Page Examples Page Table P-1Books in the Enterprise Server Documentation Set Sun GlassFish Enterprise Server Documentation Set Table P-2Default Paths and File Names Default Paths and File Names Table P-3Typographic Conventions Symbol ConventionsSymbol Conventions Typographic Conventions Documentation, Support, and Training Third-Party Web Site ReferencesSun Welcomes Your Comments Process Overview 1Performance Tuning Roadmap Performance Tuning Roadmap Performance Tuning Sequence Application Architecture Understanding Operational Requirements Client-Side Presentation User Authentication and Authorization Security RequirementsEncryption Hardware Resources Administration General Tuning Concepts 2Factors That Affect Performance Capacity Planning To Determine Capacity User Expectations Further Information Java Programming Guidelines Avoid Serialization and DeserializationUse StringBuffer to Concatenate Strings Declare Constants as static final String str = testing str = str + abcAssign null to Variables That Are No Longer Needed Declare Methods as final Only If Necessary Use DataHandlers for Soap Attachments Java Server Page and Servlet TuningSynchronize Only When Necessary Declare Method Arguments final Http Session Handling Suggested Coding PracticesGeneral Guidelines Avoid Shared Modified Class Variables Configuration and Deployment Tips Disable Security ManagerOptimize SSL EJB Performance Tuning GoalsMonitoring EJB Components Following is a sample of the monitoring output Monitoring Individual EJB Components Asadmin get -m monitorableObject General Guidelines Use High Performance BeansUse Caching Remove Unneeded Stateful Session Beans Using Local and Remote InterfacesCache and Pool Tuning Tips Use the Appropriate Stubs Prefer Local Interfaces Using Pass-By-Reference Semantics Improving Performance of EJB Transactions Version Consistency Using Special TechniquesConfigure Jdbc Resources as One-Phase Commit Resources Use the Least Expensive Transaction Attribute Entity-mapping cmp-field-mapping Cmp-field-mapping consistency To enable request partitioningWhen new.VCVERSIONNUMBER = old.VCVERSIONNUMBER Declare Request Partitioning Tuning Tips for Specific Types of EJB Components Entity BeansStateful Session Beans Checkpoint only when needed Stateless Session BeansRead-Only Entity Beans Refresh period Pre-fetching Container Managed Relationship CMR Beans OrderTable.OrderID = OrderLineTable.OrderLineOrderID Close Connections Jdbc and Database AccessUse Jdbc Directly Encapsulate Business Logic in Entity EJB Components Tune the Message-Driven Bean’s Pool Size Tuning Message-Driven BeansAsadmin set server.mdb-container.max-pool-size = value Use getConnection Limit Use of JMS Connections Cache Bean-Specific Resources Deployment Settings Tuning the Enterprise Server Use Pre-compiled JavaServer Pages Logger SettingsDisable Auto-deployment Disable Dynamic Application Reloading Log Levels Web Container SettingsGeneral Settings Session Properties Session Timeout Manager Properties Reap Interval Disable Dynamic JSP Reloading Overview of EJB Pooling and Caching EJB Container SettingsTuning the EJB Container Monitoring the EJB Container EJB Pool Settings Tuning the EJB Pool EJB Cache Settings Tuning the EJB Cache Pool and Cache Settings for Individual EJB Components Determining the best commit option Commit Option Viewing Monitoring Information Java Message Service SettingsTransaction Service Settings Monitoring the Transaction Service Tuning the Transaction Service Disable Distributed Transaction LoggingAsadmin get -m serverInstance.transaction-service Keypoint Interval Http Service SettingsMonitoring the Http Service Recover On Restart Automatic Recovery DNS Cache Information dns EnabledList --user admin --port Property name=statsProfilingEnabled value=false NameLookups HitRatioCaching DNS Entries Limit DNS Lookups to Asynchronous Keep Alive keep-alive File Cache Information file-cache Request Processing Tuning the Http ServiceAccess Log Connection Queue Request Timeout Thread CountInitial Thread Count Keep Alive Buffer LengthMax Connections Time Out DNS Lookup EnabledHttp Protocol Http File Cache Small/Medium File Size and File Size Limit Max Files CountHash Init Size Max Age Tuning Http Listener Settings Network AddressAcceptor Threads Monitoring the ORB ORB SettingsOverview How a Client Connects to the ORB Thread Pools Tuning the ORB3Tunable ORB Settings Connection Statistics ORB Thread Pool Parameters Tunable ORB Parameters Using multiple connections Client ORB PropertiesControlling connections between client and server ORB Thread Pool Sizing Examining Iiop MessagesLoad Balancing Improving ORB Performance with Java Serialization Using JSG for Application Clients Thread Pool SettingsTuning Thread Pools Unix /Linux only To Enable Java Serialization Monitoring Jdbc Connection Pools Jdbc Connection Pool SettingsTuning Jdbc Connection Pools Resources Pool Size Settings Timeout Settings4Connection Pool Sizing Connection Validation Settings Isolation Level Settings Transaction Support Connector Connection Pool Settings Testpool Page Java Virtual Machine Settings Tuning the Java Runtime System Managing Memory and Garbage Collection Tuning the Garbage Collector Choosing the Garbage Collection Algorithm To use the CMS collectorAdditional Information Tracing Garbage Collection Other Garbage Collector Settings Tuning the Java Heap Guidelines for Java Heap Sizing1Maximum Address Space Per Process Xmsvalue Xmxvalue Heap Tuning Parameters Example 4-1Heap Configuration on Solaris Rebasing DLLs on WindowsSurvivor Ratio Sizing To rebase the Application Server’s DLLs Enter this command rebase -b 0x6000000 *.dll For more information on tuning the JVM, see Page Processors Server Scaling Memory Disk SpaceNetworking Tuning Parameters Sizing the Connection Hash Table Ndd -get /dev/tcp tcpconnhashNetstat -nP tcpwc -l File Descriptor Setting Linux Configuration File Descriptors on IP Stack Settings on Tuning for Solaris on File Descriptors IP Stack Settings Tuning for Linux platforms Virtual Memory Soft Nofile 65535 HardSession required /lib/security/pamlimits.so Network Interface Disk I/O SettingsTCP/IP Settings To tune the TCP/IP settings Tuning Operating System and TCP Settings Tuning UltraSPARC T1-Based Systems 2Tuning 64-bit Systems for Performance Benchmarking Disk Configuration Network ConfigurationStart Options 106 Tuning Hadb Disk UseCalculating Hadb Data Device Size Hadbm set TotalDatadeviceSizePerNode Tuning Data Device Size Vmstat -S Memory AllocationPlacing Hadb files on Physical Disks Beware last flush/fputs took too long Performance DataBufferPoolSizeNodeNo TotalSize FreeSize Usage 512 504 Tuning DataBufferPoolSize LogBufferSizeHadbm set DataBufferPoolSize Node No Avail Free Size Tuning LogBufferSizeInternalLogbufferSize Hadbm resourceinfo --logbuf Hadbm set InternalLogbufferSize Tuning InternalLogbufferSizeNumberOfLocks Hadbm resourceinfo --nilogbuf Tuning NumberOfLocks Calculating the number of locksSummary Node No Avail Free Waits 50000 TimeoutsJdbc connection pool timeouts Load Balancer timeouts No space left on device Tuning the Enterprise Server for High-AvailabilityOperating System Configuration Hadb timeouts Tuning Session Persistence Frequency Web-methodTime-based Modified-session Session Persistence ScopeSession Size Session Checkpointing Stateful Session Beans Configuring the Jdbc Connection Pool Enabling the Health Checker Configuring the Load Balancer Tuning for High-Availability 121 122 Index Index 125 Monitoring transaction service Session timeout 128