Logging A Thread Dump

If Confluence stops responding, or is performing poorly, you should create a thread dump to help Atlassian determine the cause of the problem.

This will show the state of each thread in the JVM, including a stack trace and information about what locks that thread is holding and waiting for.

To take a thread dump from Windows:

Visit http://www.adaptj.com/root/main/download and click Launch
Click Run for any security warnings
Select Process -> Thread Dump
Under Process Id, select the '...' button.
From the drop-down list, select the Confluence process. Users running Confluence Standalone, select the 'Java (Tomcat) ...' option. Users running Confluence WAR should select their application server process.
Click OK to capture the thread dump.
Save the output to a file, eg 'threaddump.log'
If you were asked by Atlassian technical support to create the thread dump, attach the logfile to the support ticket.

Find the process id of the JVM and issue the command:

Use the ps command to get list of all processes.

kill -3 <pid>

The thread dump will be printed to the java standard output. This means that for Confluence this should be captured in the log files.

A free handy thread dump analyzer TDA 1.0 Final can be obtained from the java.net

HAT lets look at the HeapDump

Deciphering javacore thread dumps:

Labels