[omniORB] unexpected behavior - CORBA automatic retries?

[Adrian Rosoga]

Hi.

I got quite an unexpected behavior I would like to
understand.

The configuration is quite simple - client A invoking a
remote method f() on other server component, B. The
processing of f() in B takes a long time - let's say 20
seconds.

The unexpected behavior is that after 13 seconds, the same
method f() is reinvoked on B. This is quite surprising, as
I would probably expect a CORBA timeout. Why the retry (or
retries)?

The above configuration and the time values may look not
natural, but basically this is what happens in some failure
conditions. We want to deal gracefully in these situations
- expecting a single method invocation with a CORBA timeout
exception. The subsequent method invocations are quite a
problem. 

Details about the configuration: omniORB v3.0.4 on Solaris.

Please let me know if extra information is required for
answering the question. I would really appreciate some
help.

Happy day and happy CORBA.

Regards,
Adrian

__________________________________________________
Do You Yahoo!?
Yahoo! - Official partner of 2002 FIFA World Cup
http://fifaworldcup.yahoo.com