[Zope] clarification neede: postgresDA, SQL method cache and connection pool

[Bak @ kedai]

hello again

after the bad experience yesterday, i want to scale my site so that i don't
face the same problems again.  ZEO would do that.  but not now.  not when i
don't know much about zope and zeo.

but before that comes, i'd like to store data in a RDB (postgres) and split the
frontend into two or more but accessing the same RDB.  that way, i can split
user access,  give faster response, get more users.what's the pros?  cons?


my Q:
can the gurus spare some time and clarify the database connection pool (?) per
zope thread(postgres specifically).  and how is cache handled?  i've seen
errors saying that the object is not in cache.  is this related to the
connection per thread?

--traceback----
 File /home/kdie/Zope/lib/python/Shared/DC/ZRDB/DA.py, line 389, in _cached_result    (Object: qry_msg_thread)
KeyError: ('select *\012from msg\012where \012par_id=188\012', 1000)
------------

is this a good way of approaching the problems?  or is there a better way?
thanks.


-- 
------------------------------------------------------
http://www.kedai.com.my/kk
Am I Evil?