[omniORB] Timeouts and re-sending

[Matthew Berry]

Hi,

I have a CORBA server that is involved in a "transaction" in which it calls
two other back-end servers. Once the first back-end server is called, the
second one is called.

It seems that if my second server takes a long time (e.g. when I am
debugging it), a timeout in the client occurs and the request is resent to
the front end server. This is calls the first server again, and I am getting
unique constraint problems on the database associated with the first server.

Is there a simple way around this, or do I have to keep track of what steps
I have processed in the front-end server.

Thanks
Matthew Berry