Web module or application server stops processing requests

• If possible, isolate the problem by installing the web modules on different servers. If it runs on one but not the other, it is probably an environment problem.

• Check the product directory structure for a file with a name like javacore[number].txt. This file is a Java thread dump file that the JVM creates if an application server process spontaneously closes.

• Use the Tivoli performance viewer to determine if any of the application server resources, such as the Java heap, or database connections, have reached their maximum capacity. If there is a resource problem, review the application code for a possible cause:

  • If database connections are being assigned to a request but are not being released when the requests finish processing, ensure that the application code performs a close() on any opened Connection object within a finally{} block.

  • If there is a steady increase in servlet engine threads in use, review application synchronized code blocks for possible deadlock conditions.

  • If there is a steady increase in a JVM heap size, review application code for memory leak opportunities, such as static (class-level) collections, that can cause objects to never get garbage-collected.

• Enable verbose garbage collection on the application server to help determine if we have a memory leak problems. This feature adds detailed statements about the amount of available and in-use memory to the JVM error log file.

  1. In the administrative console, click...
     Servers > Server Types > Application servers > server > Server Infrastructure > Java and process management > Process definition > Java virtual machine > Verbose garbage collection

  2. Stop and restart the application server.

  3. Browse the log file for garbage collection statements. Look for statements beginning with "allocation failure". This string indicates that a need for memory allocation has triggered a JVM garbage collection, to release unused memory. Allocation failures are normal and do not necessarily indicate a problem. However, the statements that follow the allocation failure statement show how many bytes are needed and how many are allocated. If these bytes needed statements indicate that the JVM keeps allocating more memory for its own use, or that the JVM is unable to allocate as much memory as it needs, there might be a memory leak.

  We can also use the Performance and Diagnostic Advisor to detect memory leak problems.

• Determine if the application server is running out of memory.

  If we determine that the application server is running out of memory, one of the following situations might be occurring:

  • There is a memory leak in application code that we must address. To pinpoint the cause of a memory leak, enable the runHProf property on the Java Virtual Machine page of the administrative console. server is the name of the problem application server. Once we enable the runHProf property, we must:

    • Set the HProf Arguments field to a value similar to...
      depth=20,file=heapdmp.txt

      This value shows exception stacks to a maximum of 20 levels, and saves the heapdump output to...

      app_server_root/bin/heapdmp.txt

    • Save the settings.

    • Stop and restart the application server.

    • If possible, reenact the scenario or access the resource that caused the application server's process to spontaneously close, or its web modules to stop responding to new requests. Then stop the application server. If we cannot reenact the scenario or access the resource, wait until the problem reoccurs, and then stop the application server.

    • Examine the file into which the heap dump was saved...
      app_server_root/bin/heapdmp.txt file

      ...then...

      • Search for the string, "SITES BEGIN". This finds the location of a list of Java objects in memory, which shows the amount of memory allocated to the objects.

      • The list of Java objects occurs each time there was a memory allocation in the JVM. There is a record of what type of object the memory instantiated and an identifier of a trace stack, listed elsewhere in the dump, that shows the Java method that made the allocation.

      • The list of Java object is in descending order by number of bytes allocated. Depending on the nature of the leak, the problem class should show up near the beginning of the list, but this is not always the case. Look throughout the list for large amounts of memory or frequent instances of the same class being instantiated. In the latter case, use the ID in the trace stack column to identify allocations occurring repeatedly in the same class and method.

      • Examine the source code indicated in the related trace stacks for the possibility of memory leaks.

  • The JVM is using the maximum heap size that it is allowed to use. In this situation, we should increase the maximum heap size setting for application server if we have enough storage available to do so.

  • The server runtime is experiencing a problem. If we determine that there is a problem with the server runtime, apply all of the service updates for the product. If, after applying all of the service updates, the problem still exists, contact IBM Support.

Force a thread dump

The JVM creates a thread dump whenever an application server process spontaneously closes. We can also force an application to create a thread dump. A thread dump, or a javacore dump as it is also called, is one of the primary problem determination documents an application server creates. After a dump is created, we can check the dump for clues as to why new requests are not being processed.

1. Use the wsadmin command prompt, get a handle to the problem application server:
   wsadmin>set jvm [$AdminControl completeObjectName type=JVM,process=server,*]

   where server is the name of the server.

2. Generate the thread dump:
   wsadmin>$AdminControl invoke $jvm dumpThreads

   The dumpThreads command creates other types of dump files depending on the -Xdumps settings. Dump output varies depending on the platform and might include system core files, heap, and snap dumps.

3. Look for an output file, in the installation root directory for the product, with a name like...
   javacore.date.time.id.txt

After the application creates the dump, we can check for the following clues:

• "Error" or "exception information" strings at the beginning of the file. These strings indicate the thread that caused the application server process to spontaneously close. These strings are not be present if you forced the dump.

• Look at the snapshot of each thread in the process. The thread dump contains a snapshot of each thread in the process, starting in the section labeled "Full thread dump".
  • Look for threads with a description containing "state:R". Such threads are active and running when the dump is forced, or the process exited.

  • Look for multiple threads in the same Java application code source location. Multiple threads from the same location might indicate a deadlock condition (multiple threads waiting on a monitor) or an infinite loop, and help identify the application code with the problem.
  • Look for threads that are waiting on locks held by other threads.

  • Look for multiple threads in the same Java application code source location. Multiple threads from the same location might indicate a deadlock condition (multiple threads waiting on a monitor) or an infinite loop, and help identify the application code with the problem.

    It is normal for certain components in the product runtime to have certain types of threads in the same Java code source location. These components include the web container, the EJB container and the ORB thread pool.

IBM Support Assistant has analyzer and visualization tools that can save you time gathering information needed to resolve problems.

IBM Heap Analyzer

Monitor performance with Tivoli Performance Viewer

Start the wsadmin scripting client