Re: : PG9.0 - Checkpoint tuning and pg_stat_bgwriter

From: Venkat Balaji <venkat(dot)balaji(at)verse(dot)in>
To: Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
Cc: PGSQL Performance <pgsql-performance(at)postgresql(dot)org>
Subject: Re: : PG9.0 - Checkpoint tuning and pg_stat_bgwriter
Date: 2011-10-04 11:23:10
Message-ID: CAFrxt0jx7ToE0=Rje9+bgm=2GQow__NGcnnm6q4vAVUXF5tT_Q@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Thanks Heikki !

Regards,
VB

On Tue, Oct 4, 2011 at 4:38 PM, Heikki Linnakangas <
heikki(dot)linnakangas(at)enterprisedb(dot)com> wrote:

> On 04.10.2011 13:50, Venkat Balaji wrote:
>
>> I have got a situation where in i see the production system is loaded with
>> the checkpoints and at-least 1000+ buffers are being written for every
>> checkpoint.
>>
>
> 1000 buffers isn't very much, that's only 8 MB, so that's not alarming
> itself.
>
>
> I am thinking of increasing the checkpoint_segments.
>>
>> Below are our current settings -
>>
>> checkpoint_segments = 8
>> checkpoint_timeout = 5 mins
>> checkpoint_completion_target = 0.5
>> bgwriter_delay = 200ms
>> bgwriter_lru_maxpages = 100
>> bgwriter_lru_multiplier = 2
>>
>> Looking forward for suggestions.
>>
>
> Yep, increase checkpoint_segments. And you probably want to raise
> checkpoint_timeout too.
>
> --
> Heikki Linnakangas
> EnterpriseDB http://www.enterprisedb.com
>

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message hubert depesz lubaczewski 2011-10-04 12:55:36 Re: pkey is not used on productive database
Previous Message Heikki Linnakangas 2011-10-04 11:08:57 Re: : PG9.0 - Checkpoint tuning and pg_stat_bgwriter