e-docs > WebLogic Server > WebLogic Server Performance and Tuning > Tuning WebLogic Server |
WebLogic Server Performance and Tuning |
The following sections describe how to tune WebLogic Server to match your application needs.
Setting Performance-Related config.xml Parameters
The WebLogic Server configuration file (config.xml) contains a number of performance-related parameters that can be fine-tuned depending on your environment and applications. The config.xml file, located on the machine that hosts the Administration Server, provides persistent storage of Mbean attribute values. Every time you change an attribute using the system administration tools (using either the command-line interface or the Administration Console), its value is stored in the appropriate administration MBean and written to the config.xml file. Each WebLogic Server domain has its own config.xml file.
For more information about using WebLogic Server system administration tools, see the "System Administration Tools" section in the Administration Guide.
Table 3-1 lists the config.xml file parameters that affect server performance.
See Setting Thread Count. |
||
Using WebLogic Server Performance Packs
Benchmarks show major performance improvements when you use native performance packs on machines that host WebLogic Server instances. Performance packs use a platform-optimized, native socket multiplexor to improve server performance. However, if you must use the pure-Java socket reader implementation for host machines, you can still improve the performance of socket communication by configuring the proper number of socket reader threads for each server instance and client machine.
Which Platforms Have Performance Packs?
To see which supported platforms currently have performance packs available:
Make sure the NativeIOEnabled attribute of the Server element is defined in your config.xml file. The default config.xml file shipped with your distribution enables this attribute by default: NativeIOEnabled=true.
To use the Administration Console to make sure performance packs are enabled:
The value of the ThreadCount attribute of an ExecuteQueue element in the config.xml file equals the number of simultaneous operations that can be performed by applications that use the execute queue. As work enters an instance of WebLogic Server, it is placed in an execute queue. This work is then assigned to a thread that does the work on it. Threads consume resources, so handle this attribute with care—you can degrade performance by increasing the value unnecessarily.
By default, a new WebLogic Server instance is configured with a default execute queue (named "default") that contains 15 threads, which are used by all applications running on the server instance. A WebLogic Server instances also contains two built-in execute queues named __weblogic_admin_html_queue and __weblogic_admin_rmi_queue, but these queues are reserved for communicating with the Administration Console. If you configure no additional execute queues, all Web applications and RMI objects use the default queue.
Note: If native performance packs are not being used for your platform, you may need to tune the default number of execute queue threads and the percentage of threads that act as socket readers to achieve optimal performance. For more information, see Allocating Threads to Act as Socket Readers.
Should You Modify the Default Thread Count?
Adding more threads to the default execute queue does not necessarily imply that you can process more work. Even if you add more threads, you are still limited by the power of your processor. Because threads consume memory, you can degrade performance by increasing the value of the ThreadCount attribute unnecessarily. A high execute thread count causes more memory to be used and increases context switching, which can degrade performance.
The value of the ThreadCount attribute depends very much on the type of work your application does. For example, if your client application is thin and does a lot of its work through remote invocation, that client application will spend more time connected — and thus will require a higher thread count — than a client application that does a lot of client-side processing.
If you do not need to use more than 15 threads (the default) for your work, do not change the value of this attribute. As a general rule, if your application makes database calls that take a long time to return, you will need more execute threads than an application that makes calls that are short and turn over very rapidly. For the latter case, using a smaller number of execute threads could improve performance.
Default Thread Count Scenarios
To determine the ideal thread count for an execute queue, monitor the queue's throughput while all applications in the queue are operating at maximum load. Increase the number of threads in the queue and repeat the load test until you reach the optimal throughput for the queue. (At some point, increasing the number of threads will lead to enough context switching that the throughput for the queue begins to decrease.)
Note: The WebLogic Server Administration Console displays the cumulative throughput for all of a server's execute queues. To access this throughput value, follow steps 1-6 in Modifying the Thread Count in the Default Execute Queue.
Table 3-2 shows default scenarios for adjusting available threads in relation to the number of CPUs on the WebLogic Server system. These scenarios also assume that the WebLogic Server is running under maximum load, and that all thread requests are satisfied by using the default execute queue. If you configure additional execute queues and assign applications to specific queues, monitor results on a pool-by-pool basis.
Thread Count > number of CPUs (by a moderate number of threads) |
Practically ideal, with a moderate amount of context switching and a high CPU utilization rate. |
Tune the moderate number of threads and compare performance results. |
Thread Count > number of CPUs (by a large number of threads) |
Too much context switching, which can lead to significant performance degradation. Your performance may increase as you decrease the number of threads |
Reduce the number of threads so that it equals the number of CPUs, and then add only the number of "stuck" threads that you have determined. For example, if you have four processors, then four threads can be running concurrently with the number of stuck threads. So, you want the execute threads to be 4 + the number of stuck threads. To determine the amount of stuck threads, see Detecting "Stuck" Threads. Note: This recommendation is highly application-dependent. For instance, the length of time the application might block threads can invalidate the formula. |
Modifying the Thread Count in the Default Execute Queue
To modify the default execute queue thread count using the Administration Console:
Assigning Applications to Execute Queues
Although you can configure the default execute queue to supply the optimal number threads for all WebLogic Server applications, configuring multiple execute queues can provide additional control for key applications. By using multiple execute queues, you can guarantee that selected applications have access to a fixed number of execute threads, regardless of the load on WebLogic Server. See Using Execute Queues to Control Thread Usage for more information on assigning applications to configured execute queues.
Allocating Threads to Act as Socket Readers
For best socket performance, BEA recommends that you use the native socket reader implementation, rather than the pure-Java implementation, on machines that host WebLogic Server instances (see Using WebLogic Server Performance Packs). However, if you must use the pure-Java socket reader implementation for host machines, you can still improve the performance of socket communication by configuring the proper number of execute threads to act as socket reader threads for each server instance and client machine.
The ThreadPoolPercentSocketReaders attribute sets the maximum percentage of execute threads that are set to read messages from a socket. The optimal value for this attribute is application-specific. The default value is 33, and the valid range is 1-99.
Allocating execute threads to act as socket reader threads increases the speed and the ability of the server to accept client requests. It is essential to balance the number of execute threads that are devoted to reading messages from a socket and those threads that perform the actual execution of tasks in the server.
Set the Number of Socket Reader Threads on a WebLogic Server
To use the Administration Console to set the maximum percentage of execute threads that read messages from a socket:
Set the Number of Socket Reader Threads on Client Machines
On client machines, you can configure the number socket reader threads in the Java Virtual Machine (JVM) that runs the client. Specify the socket readers by defining the -Dweblogic.ThreadPoolSize=value and -Dweblogic.ThreadPoolPercentSocketReaders=value options in the java command line for the client.
Tuning Execute Queues for Overflow Conditions
You can configure a server to detect and optionally address potential overflow conditions in the default execute queue or any user-defined execute queue. WebLogic Server considers a queue to have a possible overflow condition when its current size reaches a user-defined percentage of its maximum size. When this threshold is reached, the server changes its health state to "warning" and can optionally allocate additional threads to perform the outstanding work in the queue, thereby reducing its size.
To automatically detect and address overflow conditions in a queue, you configure the following items:
To tune an execute queue using the WebLogic Server Administration Console:
WebLogic Server automatically detects when a thread in the default execute queue becomes "stuck." Because a stuck thread cannot complete its current work or accept new work, the server logs a message each time it diagnoses a stuck thread. If all threads in an execute queue become stuck, the server changes its health state to either "warning" or "critical" depending on the execute queue:
WebLogic Server diagnoses a thread as stuck if it is continually working (not idle) for a set period of time. You can tune a server's thread detection behavior by changing the length of time before a thread is diagnosed as stuck, and by changing the frequency with which the server checks for stuck threads.
Note: Although you can change the criteria WebLogic Server uses to determine whether a thread is stuck, you cannot change the default behavior of setting the "warning" and "critical" health states when all threads in the default execute queue become stuck. For more information, see Overview of WebLogic Logging Services.
To configure WebLogic Server thread detection behavior:
Tuning Connection Backlog Buffering
Use the AcceptBacklog attribute of the Server element in the config.xml file to set the number of connection requests the WebLogic Server instance will accept before refusing additional requests. The AcceptBacklog attribute specifies how many Transmission Control Protocol (TCP) connections can be buffered in a wait queue. This fixed-size queue is populated with requests for connections that the TCP stack has received, but the application has not accepted yet. The default value is 50 and the maximum value is operating system dependent.
To tune the Accept Backlog value from the Administration Console.
How JDBC Connection Pools Enhance Performance
Establishing a JDBC connection with a DBMS can be very slow. If your application requires database connections that are repeatedly opened and closed, this can become a significant performance issue. WebLogic connection pools offer an efficient solution to the problem.
When WebLogic Server starts, connections from the connection pools are opened and are available to all clients. When a client closes a connection from a connection pool, the connection is returned to the pool and becomes available for other clients; the connection itself is not closed. There is little cost to opening and closing pool connections.
How many connections should you create in the pool? A connection pool can grow and shrink according to configured parameters, between a minimum and a maximum number of connections. The best performance occurs when the connection pool has as many connections as there are concurrent client sessions.
In addition to the following subsections, see "Performance Tuning Your JDBC Application" in Programming WebLogic JDBC.
Tuning JDBC Connection Pool Initial Capacity
The InitialCapacity attribute of the JDBCConnectionPool element enables you to set the number of physical database connections to create when configuring the pool. If the server cannot create this number of connections, the creation of this connection pool will fail.
During development, it is convenient to set the value of the InitialCapacity attribute to a low number. This helps the server start up faster.
In production systems, consider setting InitialCapacity equal to the MaxCapacity so that all database connections are acquired during server start-up. If InitialCapacity is less than MaxCapacity, the server needs to create additional database connections when its load is increased. When the server is under load, all resources should be working to complete requests as fast as possible, rather than creating new database connections.
Tuning JDBC Connection Pool Maximum Capacity
The MaxCapacity attribute of the JDBCConnectionPool element allows you to set the maximum number of physical database connections that a connection pool can contain. Different JDBC drivers and database servers might limit the number of possible physical connections.
In production, it is advisable that the number of connections in the pool equal the number of concurrent client sessions that require JDBC connections. The pool capacity is independent of the number of execute threads in the server. There may be many more ongoing user sessions than there are execute threads.
For each connection pool that you create on a WebLogic Server, you can specify a prepared statement cache size. When you set the prepared statement cache size, WebLogic Server stores each prepared statement used in applications and EJBs until it reaches the number of prepared statements that you specify. For example, if you set the prepared statement cache size to 10, WebLogic Server will store the first 10 prepared statements called by applications or EJBs.
Using the prepared statement cache can dramatically increase performance, but you must consider its limitations before you decide to use it. For more details, see Increasing Performance with the Prepared Statement Cache in the Administration Guide.
Setting Performance-Related weblogic-ejb-jar.xml Parameters
The weblogic-ejb-jar.xml deployment file contains the WebLogic Server-specific EJB DTD that defines the concurrency, caching, clustering, and behavior of EJBs. It also contains descriptors that map available WebLogic Server resources to EJBs. WebLogic Server resources include security role names and data sources such as JDBC pools, JMS connection factories, and other deployed EJBs.
For information on how to modify the weblogic-ejb-jar.xml deployment file, see "Specifying and Editing the EJB Deployment Descriptors" in Programming WebLogic Enterprise JavaBeans.
Table 3-3 lists the weblogic-ejb-jar.xml file parameters that affect performance.
The following sections describe these elements.
WebLogic Server maintains a free pool of EJBs for every stateless session bean class. The max-beans-in-free-pool element of the weblogic-ejb-jar.xml file defines the size of this pool. By default, max-beans-in-free-pool has no limit; the maximum number of beans in the free pool is limited only by the available memory.
Do not change the value of the max-beans-in-free-pool parameter unless you frequently create session beans, do a quick operation, and then throw them away. If you do this, enlarge your free pool by 25 to 50 percent and see if performance improves. If object creation represents a small fraction of your workload, increasing this parameter will not significantly improve performance. For applications where EJBs are database intensive, do not change the value of this parameter.
Caution: Tuning this parameter too high uses extra memory. Tuning it too low causes unnecessary object creation. If you are in doubt about changing this parameter, leave it unchanged.
In addition to the following subsections, see max-beans-in-free-pool in Programming WebLogic Enterprise JavaBeans
Allocating Pool Size for Session and Message Beans
When EJBs are created, the session bean instance is created and given an identity. When the client removes a bean, the bean instance is placed in the free pool. When you create a subsequent bean, you can avoid object allocation by reusing the previous instance that is in the free pool. The max-beans-in-free-pool element can improve performance if EJBs are frequently created and removed.
The EJB container creates new instances of message beans as needed for concurrent message processing. The max-beans-in-pool element puts an absolute limit on how many of these instances will be created. The container may override this setting according to the runtime resources that are available.
For the best performance for stateless session and message beans, use the default setting max-beans-in-free-pool element. The default allows you to run beans in parallel, using as many threads as possible. The only reason to change the setting is to limit the number of beans running in parallel.
Allocating Pool Size for Entity Beans
There is a pool of anonymous entity beans (that is, beans without a primary key assigned to them) that is used to invoke finders and home methods, and to create entity beans. The max-beans-in-free-pool element also controls the size of this pool.
If you are running lots of finders or home methods or creating lots of beans, you may want to tune the max-beans-in-free-pool element so that there are enough beans available for use in the pool.
Tuning Initial Beans in Free Pool
Use the initial-beans-in-free-pool element of the weblogic-ejb-jar.xml file to specify the number of stateless session bean instances in the free pool at startup.
If you specify a value for initial-beans-in-free-pool, WebLogic Server populates the free pool with the specified number of bean instances at startup. Populating the free pool in this way improves initial response time for the EJB, because initial requests for the bean can be satisfied without generating a new instance.
initial-beans-in-free-pool defaults to 0 if the element is not defined.
The initial-beans-in-free-pool element is described in Programming WebLogic Enterprise JavaBeans.
WebLogic Server enables you to configure the number of active beans that are present in the EJB cache (the in-memory space where beans exist).
The max-beans-in-cache element of the weblogic-ejb-jar.xml file specifies the maximum number of objects of this class that are allowed in memory. When max-beans-in-cache is reached, WebLogic Server passivates some EJBs that have not been recently used by a client. The max-beans-in-cache element also affects when EJBs are removed from the WebLogic Server cache.
The value of this element sets the cache size for both stateful session and entity beans.
For more information, see "EJB Concurrency Strategy"in Programming WebLogic Enterprise JavaBeans
The max-beans-in-cache element is described in Programming WebLogic Enterprise JavaBeans.
Activation and Passivation of Stateful Session EJBs
Set the appropriate cache size with the max-beans-in-cache element to avoid excessive passivation and activation. Activation is the transfer of an EJB instance from secondary storage to memory. Passivation is the transfer of an EJB instance from memory to secondary storage. Tuning max-beans-in-cache too high consumes memory unnecessarily.
The EJB container performs passivation when it invokes the ejbPassivate() method. When the EJB session object is needed again, it is recalled with the ejbActivate() method. When the ejbPassivate() call is made, the EJB object is serialized using the Java serialization API or other similar methods and stored in secondary memory (disk). The ejbActivate() method causes the opposite.
The container automatically manages this working set of session objects in the EJB cache without the client's or server's direct intervention. Specific callback methods in each EJB describe how to passivate (store in cache) or activate (retrieve from cache) these objects. Excessive activation and passivation nullifies the performance benefits of caching the working set of session objects in the EJB cache—especially when the application has to handle a large number of session objects.
WebLogic Server supports database locking and exclusive locking mechanisms. The default and recommended locking mechanism for EJB 1.1 and EJB 2.0 is deferred database locking.
Database locking improves concurrent access to entity EJBs. The WebLogic Server container improves concurrent access by deferring locking services to the underlying database. Unlike exclusive locking, with deferred database locking, the underlying data store can provide finer granularity for locking EJB data, in most cases, and provide deadlock detection.
You specify the locking mechanism used for an EJB by setting the concurrency-strategy deployment parameter in the weblogic-ejb-jar.xml file. For details about database locking, see Database Concurrency Strategy in Programming WebLogic Enterprise JavaBeans.
Setting Transaction Isolation Level
Data accessibility is controlled through the transaction isolation level mechanism. Transaction isolation level determines the degree to which multiple interleaved transactions are prevented from interfering with each other in a multi-user database system. Transaction isolation is achieved through use of locking protocols that guide the reading and writing of transaction data. This transaction data is written to the disk in a process called "serialization." Lower isolation levels give you better database concurrency at the cost of less transaction isolation.
For more information, see the description of theisolation-level element of the weblogic-ejb-jar.xml file in Programming WebLogic Enterprise JavaBeans.
Refer to your database documentation for more information on the implications and support for different isolation levels.
Setting Java Parameters for Starting WebLogic Server
Java parameters must be specified whenever you start WebLogic Server. For simple invocations, this can be done from the command line with the weblogic.Server command. However, because the arguments needed to start a WebLogic Server from the command line can be lengthy and prone to error, we recommend that you incorporate the command into a script. To simply this process, you can modify the default values in the sample scripts that are provided with the WebLogic distribution to start WebLogic Server, as described in "Starting an Administration Server Using a Script".
The scripts for starting the Administration Server are called startWLS.sh (UNIX) and startWLS.cmd (Windows). These scripts are located in the WL_HOME\server\bin directory, where WL_HOME is the location in which you installed WebLogic Server.
You need to modify some default Java values in these scripts to fit your environment and applications. The important performance tuning parameters in these files are the JAVA_HOME parameter and the Java heap size parameters:
"%JAVA_HOME%\bin\java" -hotspot -Xms512m -Xmx512m -classpath %CLASSPATH% -
See Specifying Heap Size Values for details about setting heap size options.
The standard Java compiler for compiling JSP servlets is javac. You can improve performance significantly by setting your server's java compiler to sj or jikes instead of javac. The following sections discuss this procedure and other compiler considerations.
Changing Compilers in the Administration Console
To change your compiler in the Administration Console:
Setting Your Compiler in weblogic.xml
In the weblogic.xml file, the jsp-descriptor element defines parameter names and values for servlet JSPs.
For more information about setting your server's java compiler in the weblogic.xml file, see the jsp-descriptor element.
Compiling EJB Container Classes
Use the weblogic.ejbc utility to compile EJB 2.0 and 1.1 container classes. If you compile .jar files for deployment into the EJB container, you must use weblogic.ejbc to generate the container classes. By default, ejbc uses the javac compiler. For faster performance, specify a different compiler (such as Symantec sj) using the -compiler flag.
For more information, see "WebLogic Server EJB Utilities".
If you receive the following error message received when compiling JSP files on a UNIX machine:
failed: java.io.IOException: Not enough space
Try any or all of the following solutions:
set rlim_fd_max = 4096
set rlim_fd_cur = 1024
Using WebLogic Server Clusters
A WebLogic Server cluster is a group of WebLogic Servers instances that together provide fail-over and replicated services to support scalable high-availability operations for clients. A cluster appears to its clients as a single server but is in fact a group of servers acting as one.
Scalability and High Availability
Scalability is the ability of a system to grow in one or more dimensions as more resources are added to the system. Typically, these dimensions include (among other things), the number of concurrent users that can be supported and the number of transactions that can be processed in a given unit of time.
Given a well-designed application, it is entirely possible to increase performance by simply adding more resources. To increase the load handling capabilities of WebLogic Server, add another WebLogic Server instance to your cluster—without changing your application. Clusters provide two key benefits that are not provided by a single server: scalability and availability.
WebLogic Server clusters bring scalability and high-availability to J2EE applications in a way that is transparent to application developers. Scalability expands the capacity of the middle tier beyond that of a single WebLogic Server or a single computer. The only limitation on cluster membership is that all WebLogic Servers must be able to communicate by IP multicast. New WebLogic Servers can be added to a cluster dynamically to increase capacity.
A WebLogic Server cluster guarantees high-availability by using the redundancy of multiple servers to insulate clients from failures. The same service can be provided on multiple servers in a cluster. If one server fails, another can take over. The ability to have a functioning server take over from a failed server increases the availability of the application to clients.
For complete information about clusters, see "Using WebLogic Server Clusters".
Caution: Provided that you have resolved all application and environment bottleneck issues, adding additional servers to a cluster should provide linear scalability. When doing benchmark or initial configuration test runs, isolate issues in a single server environment before moving to a clustered environment.
Performance Considerations for Multi-CPU Machines
With multi-processor machines, additional consideration must be given to the ratio of the number of available CPUs to clustered WebLogic Server instances. Because WebLogic Server has no built-in limit to the number of server instances that reside in a cluster, large, multi-processor servers, such as Sun Microsystems' Sun Enterprise 10000, can potentially host very large clusters or multiple clusters.
In order to determine the optimal ratio of CPUs to WebLogic server instances, you must first ensure that an application is truly CPU-bound, rather than network or disk I/O-bound. Use the following steps to determine the optional ratio of CPUs to server instances:
If you discover that an application is primarily network I/O-bound, consider measures to increase network throughput before increasing the number of available CPUs. For truly network I/O-bound applications, installing a faster network interface card (NIC) may increase performance more than additional CPUs, because most CPUs would remain idle while waiting to read available sockets.
If you discover that an application is primarily disk I/O-bound, consider upgrading the number of disk spindles or individual disks and controllers before allocating additional CPUs.
Monitoring a WebLogic Server Domain
The tool for monitoring the health and performance of your WebLogic Server domain is the Administration Console. Using the Administration Console, you can view status and statistics for WebLogic Server resources such as servers, HTTP, the JTA subsystem, JNDI, security, CORBA connection pools, EJB, JDBC, and JMS.
For details, see "Monitoring a WebLogic Server Domain" .