Hi,
We are looking at using QP for new development and would like to hear
any thoughts any one has about our particular architecture requirements.

We have a number of sites that run a common code base and share the
same database, this is done in our current setup by setting Apache up
with proxy rewrite rules to pass through to a single backend python
medusa process that massages the request paths.

eg.
http://www.example.com -> apache_www.example.com -> localhost:8000/
www.example.com/index.html
http://another.example.com -> apache_another.example.com -> localhost:
8000/another.example.com/index.html

We do this since we currently make heavy use of SQLObject and running
it with multiple processes and getting reasonable performance wasn't
something we were able to satisfactorily achieve.

What I'd like to do is use QP and Durus, both very nice and Durus in
particular works much better with our data access patterns. My
question is really whether I should just run an external Durus
process to deal with sharing of the data across the different QP
sites or if there is another way to organise things that would work
better.

Any thoughts much appreciated.

Peter W