Re: Sustained inserts per sec ... ?

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Mike Nolan <nolan(at)gw(dot)tssi(dot)com>
Cc: scrappy(at)postgresql(dot)org (Marc G(dot) Fournier), pgsql-performance(at)postgresql(dot)org
Subject: Re: Sustained inserts per sec ... ?
Date: 2005-04-01 21:03:37
Message-ID: 28575.1112389417@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Mike Nolan <nolan(at)gw(dot)tssi(dot)com> writes:
> Well, I just did an insert of 27,500 records with 9 fields, averaging
> around 118 bytes per record, each insert statement coming from a separate
> SQL statement fed to psql, and it took a bit over 4 minutes, or about
> 106 inserts per second.

Is that with a separate transaction for each insert command? I can get
significantly higher rates on my devel machine if the inserts are
bundled into transactions of reasonable length.

With fsync on, you can't expect to get more than about one commit per
disk rotation (with a single inserting process), so with say a 7200RPM
drive (120 revs/sec) the above is a pretty good fraction of the
theoretical limit.

regards, tom lane

In response to

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Alex Turner 2005-04-01 21:17:19 Re: Sustained inserts per sec ... ?
Previous Message Christopher Petrilli 2005-04-01 20:59:52 Re: Sustained inserts per sec ... ?