Collector tool output

Use the collector tool to gather and analyze output from WebSphere Application Server.

The first step in using the collector tool on our WAS product is to run the tool to produce a JAR file as output. The second step in using the collector tool is to analyze its output. The preferred method of performing this analysis is to send the JAR file to IBM Support for analysis. However, we can use this topic to understand the content of the JAR file if we perform our own analysis.

We can view the files contained in the JAR file without extracting the files from the JAR file. However, it is easier to extract all files and view the contents of each file individually. To extract the files, use one of the following commands:

• jar -xvf WASenv.jar
• unzip WASenv.jar

WASenv.jar stands for the name of the JAR file that the collector tool creates.

The JAR file contains:

• A collector tool log file, collector.log
• Copies of stored WAS files and their full paths located under directory root in the JAR file
• Operating system information in a directory named OS
• Java information in a directory named Java
• WAS information in a directory named WAS
• Collector shell script (or batch file) execution information in a directory named debug
• MQ information in a directory named MQ, if we installed IBM MQ or the embedded messaging feature
• A JAR file manifest

Tips and suggestions

• Unzip the JAR file to an empty directory.

• Check the collector.log file for errors:
  • Some errors might be normal or expected. For example, when the collector attempts to gather files or directories that do not exist for our specific installation, it logs an error about the missing files.

  • A non-zero return code means that a command that the collector tool attempted to run does not exist. This might be expected in some cases. If this type of error occurs repeatedly, there might actually be a problem.

• (Linux) (UNIX) The OS/commands file has the location of all commands used. If we are missing command output, check this file to see if the command was found.

• (Linux) (UNIX) The collector runs some shell scripts. The shell script output is saved in files in the OS directory, while the corresponding debug information is saved in the debug directory. If the output of a shell script is missing, check the corresponding file in the debug directory.

• When we issue the collector command when there are multiple installation instances, the tool that runs depends on what is in the PATH statement. For example, if we install both the base WAS and the Deployment Manager product on the same machine, the bin directory that first appears in the PATH variable is the one that furnishes the collector tool. To work around this problem, use a fully qualified file path when calling the collector tool as shown in this example for a Windows platform:
  c:\WebSphere\AppServer\bin\collector.bat

• (Windows) The OS directory contains a file named installed.out. This file contains a list of programs found in the Add/Remove Programs list. This same information is contained in the file Desktop\My Computer\Control Panel\Add/Remove Programs\Install/Uninstall.

Gathering information with the collector tool

collector command - summary option