From:
Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To:
Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
Cc:
Jan Wieck <JanWieck(at)Yahoo(dot)com>, pgsql-hackers(at)postgreSQL(dot)org
Subject:
Re: Stats collector performance improvement
Date:
2006-01-02 20:20:24
Message-ID:
11924.1136233224@sss.pgh.pa.us (view raw or flat )
Thread:
2005-12-12 18:33:27 from "Merlin Moncure" <merlin(dot)moncure(at)rcsonline(dot)com>
2005-12-12 18:50:16 from Michael Fuhr <mike(at)fuhr(dot)org>
2005-12-12 23:01:01 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2005-12-13 01:07:51 from Michael Fuhr <mike(at)fuhr(dot)org>
2005-12-13 03:20:45 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2005-12-15 23:44:48 from Michael Fuhr <mike(at)fuhr(dot)org>
2005-12-16 00:06:39 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2005-12-16 05:44:58 from Kevin Brown <kevin(at)sysexperts(dot)com>
2005-12-16 13:17:25 from Simon Riggs <simon(at)2ndquadrant(dot)com>
2005-12-17 02:44:19 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-01-02 18:40:40 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-01-02 18:45:21 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-01-02 19:13:47 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-01-02 20:20:24 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-01-02 21:03:20 from "Qingqing Zhou" <zhouqq(at)cs(dot)toronto(dot)edu>
2006-01-02 21:48:45 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-01-03 09:40:53 from Simon Riggs <simon(at)2ndquadrant(dot)com>
2006-01-03 16:35:56 from "Jim C(dot) Nasby" <jnasby(at)pervasive(dot)com>
2006-01-03 23:28:34 from Greg Stark <gsstark(at)mit(dot)edu>
2006-01-03 21:42:53 from Hannu Krosing <hannu(at)skype(dot)net>
2006-01-08 16:49:12 from Greg Stark <gsstark(at)mit(dot)edu>
2006-01-09 15:48:21 from Hannu Krosing <hannu(at)skype(dot)net>
2006-01-02 21:48:15 from Hannu Krosing <hannu(at)skype(dot)net>
2006-01-03 04:06:57 from Jan Wieck <JanWieck(at)Yahoo(dot)com>
2006-01-03 09:54:57 from Simon Riggs <simon(at)2ndquadrant(dot)com>
2006-01-03 16:43:23 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-01-05 00:39:42 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-06-15 04:05:38 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-06-15 06:09:43 from "Qingqing Zhou" <zhouqq(at)cs(dot)toronto(dot)edu>
2006-06-15 14:27:33 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-06-15 15:57:36 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-06-15 21:38:01 from Josh Berkus <josh(at)agliodbs(dot)com>
2006-06-15 21:42:16 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-06-15 21:46:29 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-06-16 01:34:12 from "Qingqing Zhou" <zhouqq(at)cs(dot)toronto(dot)edu>
2006-06-16 01:56:34 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-06-16 03:14:06 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-06-16 03:27:50 from "Qingqing Zhou" <zhouqq(at)cs(dot)toronto(dot)edu>
2006-06-16 03:57:22 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-06-16 04:48:27 from "Qingqing Zhou" <zhouqq(at)cs(dot)toronto(dot)edu>
2006-06-16 16:03:22 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-06-16 16:12:50 from Stefan Kaltenbrunner <stefan(at)kaltenbrunner(dot)cc>
2006-06-16 16:24:03 from Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
2006-06-16 18:14:47 from Stefan Kaltenbrunner <stefan(at)kaltenbrunner(dot)cc>
2006-06-17 17:43:06 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-06-15 08:00:27 from "Larry Rosenman" <ler(at)lerctr(dot)org>
2006-06-15 19:58:27 from Stefan Kaltenbrunner <stefan(at)kaltenbrunner(dot)cc>
2006-06-15 20:29:36 from Stefan Kaltenbrunner <stefan(at)kaltenbrunner(dot)cc>
Lists:
pgsql-hackers pgsql-patches pgsql-performance
[ moving to -hackers ]
Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> I did some research on this because the numbers Tom quotes indicate there
> is something wrong in the way we process stats_command_string
> statistics.
> [ ... proposed patch that seems pretty klugy to me ... ]
I wonder whether we shouldn't consider something more drastic, like
getting rid of the intermediate stats buffer process entirely.
The original design for the stats communication code was based on the
premise that it's better to drop data than to make backends wait on
the stats collector. However, as things have turned out I think this
notion is a flop: the people who are using stats at all want the stats
to be reliable. We've certainly seen plenty of gripes from people who
are unhappy that backend-exit messages got dropped, and anyone who's
using autovacuum would really like the tuple update counts to be pretty
solid too.
If we abandoned the unreliable-communication approach, could we build
something with less overhead?
regards, tom lane
In response to
Responses
pgsql-performance by date
Next :From: Qingqing ZhouDate: 2006-01-02 21:03:20Subject : Re: Stats collector performance improvementPrevious :From : Bruce MomjianDate : 2006-01-02 19:13:47Subject : Re: Stats collector performance improvement
pgsql-hackers by date
Next :From: Qingqing ZhouDate: 2006-01-02 21:03:20Subject : Re: Stats collector performance improvementPrevious :From : Tom LaneDate : 2006-01-02 20:00:04Subject : Re: What bison versions are installed on buildfarm machines?
pgsql-patches by date
Next :From: Qingqing ZhouDate: 2006-01-02 21:03:20Subject : Re: Stats collector performance improvementPrevious :From : Bruce MomjianDate : 2006-01-02 19:13:47Subject : Re: Stats collector performance improvement
