Re: Best practices for data buffer cache setting/tuning (v15)

From: Bruce Momjian <bruce(at)momjian(dot)us>
To: Christophe Pettus <xof(at)thebuild(dot)com>
Cc: David Gauthier <dfgpostgres(at)gmail(dot)com>, pgsql-general(at)lists(dot)postgresql(dot)org
Subject: Re: Best practices for data buffer cache setting/tuning (v15)
Date: 2024-01-30 03:56:56
Message-ID: ZbhziEoDaYwcVdEn@momjian.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Mon, Jan 29, 2024 at 01:46:30PM -0800, Christophe Pettus wrote:
>
>
> > On Jan 29, 2024, at 11:39, David Gauthier <dfgpostgres(at)gmail(dot)com>
> > wrote:
> >
> > Is there a document which makes recommendations on sizing data
> > buffer cache, tuning options which evict old/unused data in mem, and
> > cache fragmentation avoidance for a v15.3 DB ?
>
> On any modern system, set shared_buffers to 25% of instance RAM, and
> get on with life.
>
> I realize that seems glib, but small variations in shared_buffers make
> almost no significant difference in performance. There are unusual
> workloads (like very high insert rate with few queries) that require
> different settings, but for most typical ELTP or data warehouse
> workloads, that is sufficient.

This blog entry might explain why Postgres requires less tuning of
shared buffers than other database systems:

https://momjian.us/main/blogs/pgblog/2018.html#December_7_2018

--
Bruce Momjian <bruce(at)momjian(dot)us> https://momjian.us
EDB https://enterprisedb.com

Only you can decide what is important to you.

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Guyren Howe 2024-01-30 05:19:01 Good overview of programming IN Postgres?
Previous Message Klaus Darilion 2024-01-29 23:15:06 Re: Monitoring logical replication