Jacob Kaplan-Moss wrote: > > On Apr 5, 2006, at 2:09 PM, Ivan Sagalaev wrote: >> How can this work? Django app connects to Postgres through psycopg, >> how >> Apache can "hold" the connections that it doesn't know anything about? >> Or does it? > > It's not "holding" connections, per se... let me see if I can explain > it more clearly: > > So the "problem" with Python (when it comes to web apps) is that > starting the interpreter is very expensive; on my main web server > ``time python -c 'import django.models'`` takes 1.5 seconds! (This > is why anyone trying to do Django as a CGI is crazy). > > To make this a non-issue, any sane environment shares one or more > Python interpreters across multiple requests (mod_python has one > interpreter per Apache child). This means that any global variables > -- including the all-important ``sys.modules`` -- stay resident in > memory between requests. Since Django's database connection is an > attribute in ``django.db``, it also sticks around between requests, > hence Postgres reporting idle connections. > > Does that make sense? > > I've consulted a database guru (Frank Wiles, http://revsys.com/) > about these "idle in transaction" connections a few times, and he > reassures me that they are less of a problem than recreating a > database connection upon each request. His suggestion -- and mine -- > is that if it's causing issues you should install a connection pool > (pg_pool is the best one) between Django and your database. > > Again, I really want to back this change out. My off-the-cuff > measurements (not very good, probably) show about a 5% performance > hit when closing the connection each time, and although that's not > much, that's 5% less hardware I have to buy. > > I'm sorry I didn't speak up *before* this change got checked in; I > missed the thread until I saw the changeset. Are there any serious > objections to backing this out?
hi Jacob, many thanks for the explanation. (and sorry to jump into the discussion).. the original poster (luke plant) said that: " It appears that connections are not being closed at all. With every hit I'm getting another entry in the process table " so if he's only hit by the reuse-open-connection effect, it should be pretty easy to reproduce: setup apache with maxclients = 3, and do more than 3 requests. if the number of open psql connections stays 3, everything is ok. if it increases, there's a problem. perhaps this would help to clarify the situation. gabor --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Django developers" group. To post to this group, send email to django-developers@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers -~----------~----~----~----~------~----~------~--~---
