Re: optimal memory

From: "A(dot) Kretschmer" <andreas(dot)kretschmer(at)schollglas(dot)com>
To: pgsql-general(at)postgresql(dot)org
Subject: Re: optimal memory
Date: 2010-08-03 09:00:04
Message-ID: 20100803090004.GH15342@a-kretschmer.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

In response to Sim Zacks :
>
>
> On 03-Aug-2010 11:18 AM, A. Kretschmer wrote:
> > In response to Sim Zacks :
> >
> >> Is there a way to tell what the optimal memory is for a specific
> >> postgresql instance?
> >>
> >> I am configuring Xen virtual machines and I don't want to give it more
> >> then it needs.
> >>
> >> Would looking at the swap be an indication? As soon as it starts to use
> >> swap, that means I need more, but until that point, I have enough?
> >>
> > You can't have enough ;-)
> >
> > Fits your DB into the RAM?
> >
> > If you don't have enough, for instance, work_mem, sort-operations
> > performed on disk and not in the ram. That's much slower. So, as i said,
> > you can't have enough ram ;-)
> >
> In theory that's a great answer.

;-)

> If my database is 400MB (du on the base directory) and there are 10
> active users running functions and queries, that for the most part take
> less then 1 sec each.
> I would assume that 10GB of RAM is overkill.

Maybe.

> Is 2 GB RAM also overkill? Is there a way to know when you have reached
> the overkill level?

I think, you should try it. Set your virtual machine to 2 GByte, set
shared_buffers to 512 MByte, effective_cache_size to 1.5 gbyte and
work_mem to, for instance, 20 mbyte. Monitor the machine, watch the
logfile (set log_min_duration_statement properly).

Reduce all parameters to 50% and compare the results.

Andreas
--
Andreas Kretschmer
Kontakt: Heynitz: 035242/47150, D1: 0160/7141639 (mehr: -> Header)
GnuPG: 0x31720C99, 1006 CCB4 A326 1D42 6431 2EB0 389D 1DC2 3172 0C99

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Scott Marlowe 2010-08-03 09:35:42 Re: optimal memory
Previous Message A. Kretschmer 2010-08-03 08:18:17 Re: optimal memory