Re: performance for high-volume log insertion - Mailing list pgsql-performance

From Glenn Maynard
Subject Re: performance for high-volume log insertion
Date
Msg-id bd36f99e0905012113g1204133t5137b58c4d502284@mail.gmail.com
Whole thread Raw
In response to Re: performance for high-volume log insertion  (PFC <lists@peufeu.com>)
List pgsql-performance
On Fri, May 1, 2009 at 8:29 PM, PFC <lists@peufeu.com> wrote:
>        Roundtrips can be quite fast but they have a hidden problem, which is
> that everything gets serialized.

The client and server will serialize, but what usually matters most is
avoiding serializing against disk I/O--and that's why write-back
caching exists.  There's still a benefit to pipelining (not everything
the db might need to read to complete the write will always be in
cache), but if everything was being serialized it'd be an order of
magnitude worse.  That's why running each insert in a separate
transaction is so much slower; in that case, it *will* serialize
against the disk (by default).

--
Glenn Maynard

pgsql-performance by date:

Previous
From: david@lang.hm
Date:
Subject: Re: performance for high-volume log insertion
Next
From: Laurent Laborde
Date:
Subject: high shared buffer and swap