[development] drupal on postgresql benchmark

Bill Moran wmoran at potentialtech.com
Mon Nov 26 19:40:00 UTC 2007

In response to Michael Favia <michael at favias.org>:

> Bill Moran wrote:
> > In response to Gerhard Killesreiter <gerhard at killesreiter.de>:
> > My point is:
> > a) Let me send you a tweaked postgresql.conf file and well find
> >    out whether the performance difference is simply a matter of
> >    untuned configuration and ...
> > b) Let me have the query log after the test and I'll see if I
> >    can identify any queries that could be improved, then someone
> >    else can test to see if the improved queries are acceptable
> >    from the MySQL standpoint.
> >   
> I think this misses the point of the experiment which was to see which 
> ran faster out of the box without optimization, not which can be made to 
> run faster (which is another useful experiment but much more complicated 
> loads, read/write/update ratio, transactions, etc).

And my point is that "out of the box" performance is an oxymoron
with PostgreSQL and isn't even worth testing.  It's admittedly
bad ... this is a known issue with PostgreSQL.  If you shoot the
horse again, all you do is waste bullets, it's already dead.

> On the other hand i completely agree with your points above and realize 
> that those individuals most concerned with performance should have or 
> desire the knowledge to properly configure the database engine/server. 
> Unfortunately this isn't a black box many people like to touch readily.

But there's more to it than that ... I want to _help_ improve this.
I don't have a set up to test myself yet (hopefully after the
first of the year, but I won't know until then) ... However, if
someone is willing to feed me log output, I can help by
identifying what is slow and methods that might speed it up.
Frankly, I suspect this will help Drupal performance _overall_
on all database platforms.

> I'm all for making things go faster i just think this test wasn't meant 
> to prove database superiority or anything like that and that is what 
> this seems to have morphed into.

That's the way most people read it and, quite frankly, I don't

I'm interested in improving the performance of Drupal, the
performance of PostgreSQL, and the performance of Drupal running
on PostgreSQL.  I can help with that, but trying to tune Drupal
queries when PostgreSQL isn't tuned for the hardware it's
running on is a recipe for wasted time.

I'm not interested in proving that PostgreSQL is better than MySQL,
I'm interested in _making_ PostgreSQL better than MySQL.  To do
that, I need data.  Unfortunately, "PostgreSQL's default config
performs poorly" isn't helpful data ;)

Bill Moran

More information about the development mailing list