Client exit handling. Is there an example?

I am working on a project to build a remote proxy database access server.
[The goal is to permit a django client running on Linux to manipulate a
Microsoft SQL Server database.] Pyro4 is used to carry queries from the
Linux client to a Windows host. It is working well enough that I am
starting to get ready to put it into limited production. There are a
hundred little things (*) that it does not do perfectly yet, but all of the
basics are working.

One major problem remaining is that the server is too reliable. If a client
program goes away unexpectedly (this is Nigeria, the Internet, and the
electricity, are not always reliable) the proxy still keeps the database
connection open, hanging there in mid transaction forever.  My only way of
clearing this at present is to hit <control-Break> on the proxy server.
This may be fine for testing, but cannot be the answer in full production.
I need some method or methods to trigger cleanup when a client ceases to
communicate.  A Python exit trigger on the client would help, but
ultimately something like a heartbeat timer may be the only answer.

Does anyone have such a partner-failure detector that you would be willing
to share?
--
Vernon Cole
-----------------------------------------
(*) re: little things. The results from the django test suite are...
Ran 4957 tests in 28832.856s

FAILED (failures=77, errors=86, skipped=209, expected failures=5)

------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev

_______________________________________________
Pyro-core mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/pyro-core