Re: Scaling SELECT:s with the number of disks on a stripe

From: Andrew - Supernews
Subject: Re: Scaling SELECT:s with the number of disks on a stripe
Date: ,
Msg-id: slrnf17508.2i67.andrew+nonews@atlantis.supernews.net
(view: Whole thread, Raw)
In response to: Scaling SELECT:s with the number of disks on a stripe  (Peter Schuller)
Responses: Re: Scaling SELECT:s with the number of disks on a stripe  (Peter Schuller)
List: pgsql-performance

Tree view

Scaling SELECT:s with the number of disks on a stripe  (Peter Schuller, )
 Re: Scaling SELECT:s with the number of disks on a stripe  ("Marc Mamin", )
  Re: Scaling SELECT:s with the number of disks on a stripe  (Peter Schuller, )
 Re: Scaling SELECT:s with the number of disks on a stripe  (Andrew - Supernews, )
  Re: Scaling SELECT:s with the number of disks on a stripe  (Peter Schuller, )
 Re: Scaling SELECT:s with the number of disks on a stripe  (Andrew - Supernews, )
  Re: Scaling SELECT:s with the number of disks on a stripe  (Peter Schuller, )
   Re: Scaling SELECT:s with the number of disks on a stripe  (Dave Cramer, )
 Re: Scaling SELECT:s with the number of disks on a stripe  (Andrew - Supernews, )
  Re: Scaling SELECT:s with the number of disks on a stripe  (Peter Schuller, )

On 2007-04-04, Peter Schuller <> wrote:
>> The next question then is whether anything in your postgres configuration
>> is preventing it getting useful performance from the OS. What settings
>> have you changed in postgresql.conf?
>
> The only options not commented out are the following (it's not even
> tweaked for buffer sizes and such, since in this case I am not
> interested in things like sort performance and cache locality other
> than as an afterthought):
>
> shared_buffers = 1000

I'd always do benchmarks with a realistic value of shared_buffers (i.e.
much higher than that).

Another thought that comes to mind is that the bitmap index scan does
depend on the size of work_mem.

Try increasing your shared_buffers to a reasonable working value (say
10%-15% of RAM - I was testing on a machine with 4GB of RAM, using a
shared_buffers setting of 50000), and increase work_mem to 16364, and
see if there are any noticable changes in behaviour.

--
Andrew, Supernews
http://www.supernews.com - individual and corporate NNTP services


pgsql-performance by date:

From: Peter Schuller
Date:
Subject: Re: Scaling SELECT:s with the number of disks on a stripe
From: Geoff Tolley
Date:
Subject: Re: SCSI vs SATA