Start with the client and work your way backward to find the problem.
When tracing exceptions back to the original problem, be aware that the RMI/IIOP
protocol requires that some exceptions undergo conversion from one type to
another as the exception passes through the runtime. Usually this transformation
is between CORBA::SystemExceptions and RMI RemoteExceptions. Pay special attention
to the CORBA::SystemException minor codes which indicate that a type transformation
has occurred.
Caused
by: |
System
exception (thrown by runtime) |
User
exception (thrown by application code) |
Look
for: |
- CEEDUMPs in controller (region) or servant (region). These dumps indicate
that the runtime had an error
- JRAS error log entries, which can narrow the error down the exception
to a specific function within the runtime
|
- CEEDUMPs
- JRAS error log entries and traces
|
Actions: |
- Look at the minor code that is listed.
- Look for fixes that address similar symptoms or minor codes.
- System exceptions usually represent the detection of an unexpected error,
and therefore (unless directed by the documentation of the minor code) will
often require IBM assistance to identify the problem.
|
- Look at your application for any sign of error.
- Look for system failures, such as a system exception in the controller
(region). If you find a system exception, follow the steps to the left for
diagnosing a system exception.
|