Re: Pgbench performance tuning?

From: Andres Freund <andres(at)anarazel(dot)de>
To: Greg Stark <stark(at)mit(dot)edu>
Cc: PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Pgbench performance tuning?
Date: 2016-08-04 18:20:37
Message-ID: 20160804182037.nqsqix4yr7abdjop@alap3.anarazel.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 2016-08-04 19:15:43 +0100, Greg Stark wrote:
> On Thu, Aug 4, 2016 at 6:53 PM, Andres Freund <andres(at)anarazel(dot)de> wrote:
> > What's the config? Version? What activity does pidstat -d -l indicate?
> > How much WAL was generated?
>
> I know the specifics matter but I was also trying to avoid dumping too
> much into the email.
>
> The shared buffers is set to 16384 (128MB). Otherwise it's a default
> config (For 9.4 and before I set checkpoint_segments to 32 as well).

Well, with 128MB I don't find that a very surprising result. You're
going to push data out to disk constantly. Given the averaged random
access pattern of pgbench that's not really something that interesting.

> Never seen pidstat before but pidstat -d looks like it prints very
> similar output to the iostat output I was gathering already. There's
> nothing else running on the machine.

The question is which backends are doing the IO.

Greetings,

Andres Freund

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2016-08-04 18:21:01 Re: improved DefElem list processing
Previous Message Simon Riggs 2016-08-04 18:19:25 Re: Lossy Index Tuple Enhancement (LITE)