MustGather: Crash on Linux produces no core or truncated core

Technote (FAQ)
Problem
MustGather for when a WebSphere® Application Server crash on Linux® produces no core or truncated core. Gathering this information before calling IBM support helps familiarize you with the troubleshooting process and saves you time.
Solution
If you already contacted support, continue to the component-specific MustGather information. Otherwise, click: MustGather: Read first for all WebSphere Application Server products.

Crash specific MustGather information:

The following lists what needs to be done to ensure that a core file is produced on Linux when the Java™ virtual machine (JVM™) crashes:

  1. Add the following export statement in the file, startupServer.sh (for releases of V3.5 and V4.0) or startServer.sh (for releases V5.0, V5.1 and V6.0):

    export JAVA_DUMP_OPTS="ONERROR(JAVADUMP,SYSDUMP), ONEXCEPTION(JAVADUMP,SYSDUMP),ONDUMP(JAVADUMP)"

    Note: The preceding 2 lines are broken due to length. The 2 lines should be joined in the script.
  2. Ensure that the following ulimit settings are enabled:

    ulimit -c unlimited
    ulimit -n unlimited
    ulimit -u unlimited


    It is recommended to set the values to unlimited and ensure that you have enough free file system space in install_root/bin (at least 750 MB to 2 GB).
  3. Ensure that the WebSphere Application Server process can write to the install_root/bin directory.
  4. Issue the kill -11 command (against PID of the Java process) to test getting a core file.
  5. If you still do not get a core file after setting these parameters correctly, do the following:
    1. Add the following export statement to the beginning of the startupServer.sh (for releases of V3.5 and V4.0) or startServer.sh (for releases of V5.0, V5.1 and V6.0):

      export DISABLE_JAVADUMP=true
    2. Restart the administrative server and all Java processes.
    3. Issue the kill -11 command (against PID of the Java process) to test getting a core file.
    4. If you cannot get a core file from the test, determine if you can obtain a core file from any process on the machine by issueing a kill -11 against the PID of another running process such as vi.
    5. If you can get a core file from the test, wait for a crash to occur during an error condition.

    Note:
    You can also search for core files in the following directories:
    • install_root/bin
    • Configured Application Server working directory
    • /tmp

    If you cannot find a core file in any of these directories, search your entire machine for core* files.

What to do if the crashing JVM still does not produce a core after a crash:
  1. Collect the following diagnostic information:
    • All logs from the WebSphere Application Server that is crashing.
    • /usr/var/messages
    • Description of the application, including use of back-end systsms and native code.
  2. Enable the Verbose Java Native Interface (JNI) trace.
  3. Restart the Application Server.
  4. After the next crash of the Application Server, collect the following diagnostic information:
    • All logs from the WebSphere Application Server that is crashing.
    • /usr/var/messages
  5. Follow instructions tosend diagnostic information to IBM support.

For a listing of all technotes, downloads, and educational materials specific to the Crash component, search the WebSphere Application Server support site.
Related information
Submitting information to IBM Support

Steps to getting support

MustGather: Readme first

Troubleshooting guide












Document Information

Product categories: Software, Application Servers, Distributed Application & Web Servers, WebSphere Application Server, Crash
Operating system(s): Linux
Software version: 3.5, 4.0, 5.0, 5.1, 6.0
Software edition: Advanced, Base, Express, Network Deployment, Single Server
Reference #: 1115658
IBM Group: Software Group
Modified date: 2004-06-07