Re: in-transaction insert performance in 7.5devel - Mailing list pgsql-performance

From Merlin Moncure
Subject Re: in-transaction insert performance in 7.5devel
Date
Msg-id 6EE64EF3AB31D5448D0007DD34EEB34101AE5A@Herge.rcsinc.local
Whole thread Raw
In response to in-transaction insert performance in 7.5devel  ("Merlin Moncure" <merlin.moncure@rcsonline.com>)
Responses Re: in-transaction insert performance in 7.5devel
List pgsql-performance
Rod Taylor wrote:
> On Fri, 2004-06-11 at 14:40, Merlin Moncure wrote:
> > I am batch inserting insert statements into a database with fsync =
on.
> > My single disk system is on a 10k drive...even though I am inside a
> > transaction there is at least 1 file sync per row insert.
>
> Which filesystem?
>
> PostgreSQL isn't issuing the sync except at commit of a transaction,
but
> some filesystems do wacky things if you ask them too.

Um, NTFS. :)  I'm playing with the new 'syncer' to get a feel for the
new performance considerations.

As I understand it, sync() is never called anymore.  mdsync() hits the
all the files 1 by 1 with an fsync.  My understanding of the commit
process is that 30 tps is quite reasonable for my hardware.

What surprised was that I am getting at least 1 seek/insert inside a
transaction.

Merlin

pgsql-performance by date:

Previous
From: Rod Taylor
Date:
Subject: Re: in-transaction insert performance in 7.5devel
Next
From: Rod Taylor
Date:
Subject: Re: in-transaction insert performance in 7.5devel