Re: Performance Bottleneck

From: Dennis Bjorklund <db(at)zigo(dot)dhs(dot)org>
To: Martin Foster <martin(at)ethereal-realms(dot)org>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: Performance Bottleneck
Date: 2004-08-03 18:35:36
Message-ID: Pine.LNX.4.44.0408032029080.9559-100000@zigo.dhs.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Tue, 3 Aug 2004, Martin Foster wrote:

> to roughly 175 or more. Essentially, the machine seems to struggle
> to keep up with continual requests and slows down respectively as
> resources are tied down.

I suggest you try to find queries that are slow and check to see if the
plans are optimal for those queries.

There are some logging options for logging quries that run longer then a
user set limit. That can help finding the slow queries. Just doing some
logging for some typical page fetches often show things that can be done
better. For example, it's not uncommon to see the same information beeing
pulled several times by misstake.

Maybe you can also try something like connection pooling. I'm not sure how
much that can give, but for small queries the connection time is usually
the big part.

> Would disabling 'fsync' provide more performance if I choose that
> information may be lost in case of a crash?

I would not do that. In most cases the performance increase is modest and
the data corruption risk after a crash is much bigger so it's not worth
it.

If you have a lot of small inserts then it might be faster with this, but
if possible it's much better to try to do more work in a transaction then
before.

--
/Dennis Björklund

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Gaetano Mendola 2004-08-03 18:41:23 Re: Performance Bottleneck
Previous Message Scott Marlowe 2004-08-03 18:26:28 Re: NOT IN query takes forever