Re: COMMIT NOWAIT Performance Option

From: Robert Treat <xzilla(at)users(dot)sourceforge(dot)net>
To: pgsql-hackers(at)postgresql(dot)org
Cc: "A(dot)M(dot)" <agentm(at)themactionfaction(dot)com>
Subject: Re: COMMIT NOWAIT Performance Option
Date: 2007-03-04 04:19:00
Message-ID: 200703032319.00744.xzilla@users.sourceforge.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Monday 26 February 2007 19:27, A.M. wrote:
> On Feb 26, 2007, at 18:58 , Simon Riggs wrote:
> > On Mon, 2007-02-26 at 23:25 +0000, Richard Huxton wrote:
> >> Simon Riggs wrote:
> >>> Proposal: Implement a new option for COMMIT, for enhancing
> >>> performance,
> >>> providing a MySQL-like trade-off between performance and
> >>> robustness for
> >>> *only* those that want it.
> >>>
> >>> COMMIT NOWAIT
>
> Isn't the domain of transient data relegated to certain tables
> instead of specific transactions? Wouldn't it be easier to create un-
> wal-logged global temp tables?
>

A similar idea we've been kicking around would be having a set storage
parameter = nologging option for alter table which would, as it's name
implies, cause the system to ignore writing wal logs for the table, much like
it does for temp tables now. One cavaet would be you would probably need to
forbid such a table from being the parent side of a FK relationship, but
otherwise this should be fairly safe even for replay since alter table needs
an exclusive lock, so you have finite points where data for the table would
be written or ignored.

--
Robert Treat
Build A Brighter LAMP :: Linux Apache {middleware} PostgreSQL

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Joshua D. Drake 2007-03-04 04:49:24 Re: Patch license update to developer's FAQ
Previous Message Bruce Momjian 2007-03-04 03:57:02 Re: IDENTITY/GENERATED v36 Re: Final version of IDENTITY/GENERATED patch