If the server fails at a specific, reproducible point, take the following steps:
Check the server log for errors, to assess when the crash occurred, and to determine areas for further investigation.
Use the Runtime Monitoring feature in EAServer Manager to find out which components are active, the connections being used, and so on—see “Runtime monitoring tools”.
Using the insights from steps 1 and 2, turn on debug and trace flags for specific components, connections, and so on.
Explore what is happening in the application or component. Look for common application trouble spots—see “Generic issues”.
Check the component and server property settings.
Debug the component.
Check the server configuration, specifically:
Patch levels—see “EAServer log” for details.
Resource and memory usage.
DLL versions, library versions, and Java class loading—see “Verifying your configuration” for library and class information, and “ListDLLs”.
Invoke a service component just before the crash to write information to the EAServer log—see “ListDLLs”:
JVM properties
Free and total memory
EAServer monitoring data
EAServer memory dump
If a C++ component is involved, try running it within a dedicated external process, identified by the com.sybase.jaguar.component.cpp.process property.
There are restrictions on running externally, and not all C++ components are suitable candidates. See “Running C++ components externally” in Chapter 14, “Creating CORBA C++ Components,” of the EAServer Programmer’s Guide for details.
Run the debug server for more information in the dump file. Use the stack trace, dump file, and crash address for advanced troubleshooting—see “Stack traces, dump files, and core files”.
