Will higher shared_buffers improve tpcb-like benchmarks?

From: Saurabh Nanda <saurabhnanda(at)gmail(dot)com>
To: pgsql-performance(at)lists(dot)postgresql(dot)org
Subject: Will higher shared_buffers improve tpcb-like benchmarks?
Date: 2019-01-29 11:39:14
Message-ID: CAPz=2oEnzLofpgZnAwmwaZtW-OLn5+MFRux9uFnnWs2+YecZAA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Hi,

I'm going crazy trying to optimise my Postgres config for a production
setting [1] Once I realised random changes weren't getting my anywhere, I
finally purchased PostgreSQL 10 - Higher Performance [2] and understood the
impact of shared_buffers.

IIUC, shared_buffers won't have any significant impact in the following
scenario, right?

-- DB size = 30GB
-- shared_buffers = 2GB
-- workload = tpcb-like

This is because the tpcb-like workload selects & updates random rows from
the DB [3]. Therefore, with a 2GB shared buffer, there is only a 6-7%
chance (did I get my probability correct?) that the required data will be
in the shared_buffer. Did I understand this correctly?

If nothing else becomes the bottleneck (eg. periodically writing dirty
pages to disk), increasing the shared_buffers to 15GB+ should have a
significant impact, for this DB-size and workload, right? (The system has
64 GB RAM)

[1] Related thread at
https://www.postgresql.org/message-id/flat/CAPz%3D2oGdmvirLNX5kys%2BuiY7LKzCP4sTiXXob39qq6eDkEuk2Q%40mail.gmail.com
[2]
https://www.packtpub.com/big-data-and-business-intelligence/postgresql-10-high-performance
[3] https://www.postgresql.org/docs/11/pgbench.html#id-1.9.4.10.7.2

-- Saurabh.

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Joe Mirabal 2019-01-29 12:12:38 Re: Will higher shared_buffers improve tpcb-like benchmarks?
Previous Message Fabio Isabettini 2019-01-29 11:32:47 Re: dsa_allocate() faliure