dieter@handshake.de wrote at Sat, 29 Mar 2003 01:04:17 +0100:

...

Yes, we do close the connection. This is the full fragment:

app = Zope.app() obj = app.restrictedTraverse(path) get_transaction().commit() app._p_jar.close() del app return obj

The server freezes inside the restrictedTraverse call (never returns).

I suspect there is some hidden and rarely exposed locking bug in Zope causing Zope to stop responding to requests due to a deadlock. I dearly would like to locate and fix it. But, I am unable to reproduce the situation reliably.

Maybe, your setup provides a way for reproduction.

The good (or bad) news is that the problem is gone. But I can tell you the following: - it happened inside restrictedTraverse. I replaced that call with a quick, dirty and stupid iterative function using split + getattr, and the code worked well. - it was random. Sometimes the same call that worked fine a few seconds ago would hang the Zope server then. - right now, I cannot reproduce the error, even using the original restrictedTraverse call. Anyway, we changed our design and now we are using a single instance of a Product located at site root, so there is no need to access Zope from 'outside' the ZODB. I cannot tell what has changed in order to make this code work now. The database (FileStorage) was packed (and then we got some "Bad reference" errors); there were some __setstate__ methods not being called properly; some classes were improperly declared as descendants of both SimpleItem and Persistent; the machine running Zope was being used for other purposes at the same time (like CD recording). Maybe the problem had to do with some of these odd things, maybe not, but I can't tell. Gabriel Genellina Softlab SRL ------------ ¡Navegá y ayudá a un chico! Por cada usuario que se conecte a Internet con Yahoo! Conexión en abril, Yahoo! Argentina donará un plato de comida para un niño, a través de la asociación "Por los chicos". http://ar.online.yahoo.com