Re: High Disk write and space taken by PostgreSQL

From: Bruce Momjian <bruce(at)momjian(dot)us>
To: "anarazel(at)anarazel(dot)de" <andres(at)anarazel(dot)de>
Cc: Ondrej Ivanič <ondrej(dot)ivanic(at)gmail(dot)com>, J Ramesh Kumar <rameshj1977(at)gmail(dot)com>, pgsql-performance(at)postgresql(dot)org
Subject: Re: High Disk write and space taken by PostgreSQL
Date: 2012-08-16 16:23:03
Message-ID: 20120816162303.GC8359@momjian.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Thu, Aug 16, 2012 at 06:07:26PM +0200, anarazel(at)anarazel(dot)de wrote:
>
>
> Bruce Momjian <bruce(at)momjian(dot)us> schrieb:
>
> >On Thu, Aug 16, 2012 at 03:48:57PM +1000, Ondrej Ivanič wrote:
> >> Hi,
> >>
> >> On 16 August 2012 15:40, J Ramesh Kumar <rameshj1977(at)gmail(dot)com>
> >wrote:
> >> > As you said, MySQL with MyISAM is better choice for my app. Because
> >I don't
> >> > need transaction/backup. May be I'll try with InnoDB and find the
> >disk
> >> > write/space difference. Is there any similar methods available in
> >postgresql
> >> > like MyISAM engine ?
> >>
> >> You can try unlogged tables:
> >> http://www.postgresql.org/docs/9.1/static/sql-createtable.html
> >>
> >> If specified, the table is created as an unlogged table. Data written
> >> to unlogged tables is not written to the write-ahead log (see Chapter
> >> 29), which makes them considerably faster than ordinary tables.
> >> However, they are not crash-safe: an unlogged table is automatically
> >> truncated after a crash or unclean shutdown. The contents of an
> >> unlogged table are also not replicated to standby servers. Any
> >indexes
> >> created on an unlogged table are automatically unlogged as well;
> >> however, unlogged GiST indexes are currently not supported and cannot
> >> be created on an unlogged table.
> >
> >I would set full_page_writes = off too.
> Why? There shouldn't be any such writes on unlogged tables.

True. I was thinking more of the logged tables, and the system tables.

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

+ It's impossible for everything to be true. +

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message delongboy 2012-08-16 16:31:51 Re: Increasing WAL usage followed by sudden drop
Previous Message Jeff Janes 2012-08-16 16:16:45 Re: cluster on conditional index?