IBM Support

Collect troubleshooting data for performance problems in WebSphere Process Server (WPS)

Technote (troubleshooting)


You are having a performance problem in WebSphere Process Server. You would like to know what documentation you must collect (MustGather) so that the WebSphere Process Server Support team can diagnose your problem. If you gather this documentation before contacting support it will expedite the troubleshooting process, and save you time.

Resolving the problem

Collect the MustGather information for performance problems in WebSphere Process Server and WebSphere Enterprise Service Bus (WebSphere ESB). Before you submit a problem management report (PMR) to IBM, review the additional resources that are listed in this technote to understand performance problems and how they relate to WebSphere Process Server or WebSphere ESB. Also, consider applying interim fixes that address performance problems and improvements, or apply the latest cumulative fix pack for WebSphere Process Server and the latest cumulative fix pack for WebSphere ESB.

Resolution of performance problems with WebSphere Process Server or WebSphere ESB comes from two sources:

  • IBM Support - For problems with performance or performance degradation that are caused by a defect in the related IBM product.
  • IBM Services - For tuning and configuration problems, such as meeting a performance goal.

Determining if a problem is a code defect can be difficult because expectations of the product performance and the intended product performance can differ. Isolating the problem to a specific bottleneck or atomic function might be a requirement for the scenario to be considered a code defect. For example, a support request such as "an incoming sample message of 100 bytes takes 1 hour to parse before it is received by the first component" might be evaluated as a potential code defect. In comparison, a request such as "the server runs only 100 events per hour" might be a tuning problem.

Setting the trace string

Set the following trace string, then reproduce the problem. Also include trace strings for additional components that are involved in the performance problem. For instructions, review the main WebSphere Process Server troubleshooting document.

Note: Enabling the trace might slow down your system. This delay can lead to transaction timeouts and subsequent errors.

Required diagnostic information

Gather the following information and files:

General diagnostic information

  • A detailed problem description
  • Environment description (hardware and software in use)
  • Server logs: All files from the following directory:
  • Project interchange (PI) or enterprise archive (EAR) file from the relevant applications

Performance diagnostic information
  • Your analysis of the performance problem and an explanation of it if you use third-party software to collect performance metrics and data

Instructions for collecting the diagnostic information

Collect the following diagnostic information:

  • Your analysis of the performance problem, including the quantified (numerical) results from specific tests and the expected results.

  • Collect the general diagnostic information. For instructions, see the "Manually collect troubleshooting information" section in Collect troubleshooting data for WebSphere Process Server Version 6. You can collect the server log files and some of the environment details using ISA Lite.

What to do next

  1. Review the log files and traces at the time of the problem to try to determine the source of the problem.
  2. Use the IBM Support Assistant to search for known problems in the information center, forums, and technotes.
  3. If you cannot find related problems or cannot solve the problem, send the information you have collected to the IBM Support by following the instructions in Exchanging Information with IBM Technical Support.

For a listing of all technotes, downloads, and educational materials that are specific to WebSphere Process Server and WebSphere ESB, search the WebSphere Process Server and the WebSphere Enterprise Service Bus support pages.

Additional resources
This section contains links to articles that pertain to performance of WebSphere Process Server and WebSphere ESB. Note that assistance with these documents or tuning techniques is beyond the scope of IBM Support. Because WebSphere Process Server is based on several components and underlying products that can be tuned, information on these components and products is included.
  • Gathering performance metrics
To gather performance metrics from service component event points, use the Performance Monitoring Infrastructure (PMI), Tivoli Performance Viewer (TPV), or both. Using PMI data, you can identify and fix the performance bottlenecks in the application and server. PMI data can also be used to monitor the health of the application server.
  • WebSphere Process Server tuning papers and recommendations
WebSphere Business Integration V6.0.2 Performance Tuning – Redpaper

WebSphere Business Integration V6.0.1 Performance Tuning – IBM Redbooks publication
  • WebSphere Application Server tuning and recommendations

WebSphere Application Server performance overview site
  • DB2 database-specific tuning papers

Quick reference for DB2 tuning

Administration Guide: Performance (PDF FILE)

Design recommendations
In addition to the tuning steps in the corresponding run times and databases, you must also consider tuning steps during design time. The following general recommendations for the design of the application can significantly improve performance during execution:
  • Use synchronous bindings when possible.
  • Use microflows (short-running Business Process Execution Language (BPEL) processes) whenever possible.
  • Consider transaction boundaries in long-running BPEL processes with regard to transactional behavior of invoke activities. Use participates transaction qualifiers whenever possible.
  • Disable the Common Event Infrastructure (CEI), if possible.
  • Disable the "Enable persistence and queries of business-relevant data" feature, if possible.
  • Set the "suspend transaction" Service Component Architecture (SCA) qualifier to false, if possible.
  • Disable unnecessary traces (especially in a production environment).
  • Declare variables as read only, if applicable, in Java snippets. You can find an example in BPEL process variable changes in Java.

Even with runtime and design-time performance improvements, network performance might cause a bottleneck between the communications of services and components. Also, it can cause slow response times and timeouts, for example, in web service calls, database operations, and the communication between a deployment manager and its node agents in clustered and distributed environments. Consider evaluating network problems with the network administrator.

Related information

Apache JMeter
Rational Performance Tester

Cross reference information
Segment Product Component Platform Version Edition
Business Integration WebSphere Enterprise Service Bus Performance AIX, HP-UX, i5/OS, Linux, Linux iSeries, Linux pSeries, Linux xSeries, Linux zSeries, Solaris, Windows, z/OS 7.0, 6.2, 6.1.2, 6.1, 6.0.2, 6.0.1

Product Alias/Synonym


Document information

More support for: WebSphere Process Server

Software version: 6.1, 6.1.2, 6.2, 7.0

Operating system(s): AIX, HP-UX, Linux, Solaris, Windows, z/OS

Reference #: 1269374

Modified date: 21 October 2009