Tips and Tricks

Digg This!

This article will provide some useful tips for debugging your BEA WebLogic Server applications when a system core file occurs. It describes debugging tips, problem troubleshooting, and tools available to assist you in this process.

A system core file is usually indicative of an error in some native code. This could be from the application code of a user (if you are using native code [JNI] in your application), an error in the Java Virtual Machine version you are using, or in BEA WebLogic Server itself. There are a couple of places where native code could have caused the system core file to happen on your operating system. The following ideas and suggestions will help narrow the problem and stabilize the application until the exact cause of the system core can be determined.

The Java Virtual Machine
The first place to look is at the JVM itself. The JVM is a native program and can cause such errors. When in doubt, you can try another certified JVM and/or a later release to determine if a JVM bug is at fault. Many JVM bugs involve the use of the JIT compiler and disabling this feature will often resolve this type of problem. This can be done by supplying the "-Djava.compiler=none" option to the Java command line. Most of the time the JVM will produce a small log file that may contain useful information as to which library the system core may have come from; however, this is not true all of the time. The file is produced in the directory where BEA WebLogic Server was started and is of the form "hs_err_pid<PID #>.log", where "<PID #>" is the process ID of the BEA WebLogic Server process. On AIX this file would be "javacore<PID>.<ID Number>.txt", where "<PID #>" is the process ID of the BEA WebLogic Server process and "<ID Number>" is a number generated by the operating system. You can go to http://java.sun.com and search the Bug Database with the HotSpot "Error ID" number or the method that was happening when the error was reported.

Native Code
The next place to look is at any native libraries that are accessed with JNI (native code) calls in your application, if applicable. This can also cause a system core file to be produced. If your application uses such libraries, they should be carefully examined. It may be difficult to rule out these libraries, as their functionality may not be easily removed from the application. Extensive logging may be needed to determine if a pattern of use can be correlated with the system core dump/Dr. Watson error.

Another place is any Type 2 JDBC driver that makes use of native DBMS libraries, which could also produce this type of error. An option is to switch to a pure Java (Type 4) JDBC driver in order to determine if that is the cause. You can also check with the Type 2 JDBC driver provider to see if there are any known issues or if there is an updated version of the native libraries available.

A final place that could have caused this could be in the BEA WebLogic Server performance pack. This is also native code and when enabled could potentially produce such an error. One option is to disable this feature to determine if that is the cause. This can be done by adding "-Dweblogic.NativeIOEnabled=false" to the command line to start your server. If you are on an older service-pack version you can also check the list of fixed change requests for your particular BEA WebLogic Server version at http://edocs.bea.com.

System Core File
If after looking at the above items you cannot determine the cause of the error, then we can examine the system core file that was produced. If a system core file is produced it will be in the directory where BEA WebLogic Server was started. You will need to run an operating system-specific debugger on the resulting system core file to get a stacktrace that may help in pointing to the offending code.

If you are in development and the system core consistently happens, you can set the following flags to allow a thread dump to be taken of the server right before a core happens. This will get the state of the threads at that moment and may help narrow the problem to application code or point to a bug. The option is "-XX:+ShowMessageBoxOnError" on the Sun JVM (which is not officially documented on the Sun Web site). When the JVM crashes, the program will prompt: "Do you want to debug the problem?" You can then take a thread dump of the JVM. This option will be available on the 8.1 SP2 version of the BEA JRockit JVM when it becomes available. However, in that version the corresponding option will be "-Djrockit.waitonerror".

The best option is to run a debugger on the resulting system core file to get a stacktrace of the native calls. The following information was also given in my last article (WLDJ, Vol. 3, issue 2) but it is being repeated here for consistency. This information may help point out the offending code to you; or, if you are unsure, then contact BEA Customer Support with this information so they can investigate the stacktrace more thoroughly. If you are on a Windows platform, then a "Dr. Watson" file may be produced. Please send this file to BEA Customer Support when opening a case. Otherwise, check the following "Unix" operating system values to make sure that they have already been properly set in order to generate a core file:

1. Check the "ulimit -c" (configured size of the core file) at a system and user level to make sure that it is set and that the value is not set too low to produce a meaningful core file.
2. Check the available disk space for the user (for example, is there a disk quota?).
3. Also check the following parameter, which on Solaris is usually in /etc/system file and can be used to disable core files: set sys:coredumpsize=0.
4. On Linux, the core dump is turned off by default on all systems. In RedHat Advanced Server 2.1 it should be under "/etc/security". There should be a self-explanatory file called limits.conf; within that file look for the word "core". If it is set to "0", then coredump files are disabled.
5. On HPUX check the setting called "kernel parm maxdsiz" (max_per_proc_data_size, which increases the User Process Data Segment Size) from the old value of, say, 64M to something higher like 134M.

Please get a stacktrace (or backtrace) from your debugger. Here are the commands needed when using "dbx" or "gdb", which will work on most platforms:

a. dbx:

• $ java -version: Need to use right version of JDK
• $ ls /opt/bin/dbx: Need to know dbx location, or "which dbx"
• $ export: DEBUG_PROG=/opt/bin/dbx (or wherever "dbx" is located)
• $ <path to java command>/java corefile

Now you will be in the debugger. Execute the following commands:

• (dbx) where: Shows a summary of the stack
• (dbx) threads: Shows the state of the existing threads
• (dbx) quit

b. gdb

• java -version: Need to use right version of jdk
• ls /usr/local/bin/gdb: Need to know gdb location or "which gdb"
• export DEBUG_PROG=/usr/local/bin/gdb: Or wherever "gdb" is located)
• <path to java command>/java corefile

Now you will be in the debugger. Execute the following commands:

• (gdb) where: Shows a summary of the stack
• (gdb) thr: Switch among threads or show the current thread
• (gdb) info thr: Inquire about existing threads
• (gdb) thread apply 1 bt: Apply a command to a list of threads, specifically the backtrace to thread #1)
• (gdb) quit

If you don't have access to a debugger you can check to see if you have the "pstack" and "pmap" utilities on your operating system. If you do have those utilities (on some operating systems you have to download these utilities separately); you can run them on the system core file to gather information for support. The syntax of the command would be something like this:
$ /usr/proc/bin/pstack core
$ /usr/proc/bin/pmap core

Further Information
You can also go to http://support.bea.com and find some published solutions on core files. In the "Question" field type in "S-16147" to display some additional information about using pstack and pmap on Solaris.

Enabling/Disabling Dr. Watson
The drwtsn32.log files are similar to system core files on Unix. On Windows 2000, these files are found in the following directory: C:\Documents and Settings\All Users\Documents\DrWatson. After entering "drwtsn32 ?" in a DOS command window and hitting enter, the "Dr. Watson for Windows 2000" box appears. The DrWatson log file overview option will display a screen that explains the format of the drwtsn32.log files.

By default, Dr. Watson will be enabled when Windows NT is installed. To disable Dr. Watson, the following Registry value must be changed from 1 to 0 (zero): "\HKEY_LOCAL_MACHINE\SOFTWARE \Microsoft\Windows NT\CurrentVersion\AeDebug". An entry called "Auto" corresponds to how Dr. Watson will start up. To enable Dr. Watson, change the "Auto" value from 0 (zero) to 1. This will then launch whatever debugger, or application, is under the Debugger registry value. For Dr. Watson, the Debugger value should contain:

drwtsn32 -p %ld -e %ld -g

Links to On-Line Debugger Manuals

1. gdb: www.gnu.org/manual/gdb-5.1.1/gdb.html
2. dbx:
   a. Sun: http://docs.sun.com/db/doc/805-4948?q=DBX
   b. IBM: http://publib16.boulder.ibm.com/pseries/en_US/cmds/aixcmds2/dbx.htm
3. adb: HP: http://docs.hp.com/hpux/onlinedocs/B2355-90680/00/00/8-con.html

If none of these hints help direct you towards a solution or an identifier in your application, then you should contact BEA Customer Support for further diagnosis. You can open a case with a valid support contract by logging in at http://support.bea.com/login.jsp.