From:
Gregory Stark <stark(at)enterprisedb(dot)com>
To:
"Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc:
"Heikki Linnakangas" <heikki(at)enterprisedb(dot)com>, "Simon Riggs" <simon(at)2ndquadrant(dot)com>, "Alvaro Herrera" <alvherre(at)commandprompt(dot)com>, <pgsql-hackers(at)postgresql(dot)org>, <pgsql-patches(at)postgresql(dot)org>
Subject:
Re: [HACKERS] WAL logging freezing
Date:
2006-10-31 19:49:27
Message-ID:
877iygz2ns.fsf@enterprisedb.com (view raw or flat )
Thread:
2006-10-24 07:20:13 from Heikki Linnakangas <heikki(at)enterprisedb(dot)com>
2006-10-24 12:11:33 from Alvaro Herrera <alvherre(at)commandprompt(dot)com>
2006-10-24 14:11:43 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-27 15:44:52 from "Heikki Linnakangas" <heikki(at)enterprisedb(dot)com>
2006-10-27 16:01:12 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-27 21:19:54 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-10-27 21:23:38 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-27 22:03:17 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-10-30 13:42:27 from "Heikki Linnakangas" <heikki(at)enterprisedb(dot)com>
2006-10-30 14:08:37 from "Heikki Linnakangas" <heikki(at)enterprisedb(dot)com>
2006-10-30 16:20:56 from Alvaro Herrera <alvherre(at)commandprompt(dot)com>
2006-10-30 17:05:19 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-30 20:30:19 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-10-30 21:39:58 from Alvaro Herrera <alvherre(at)commandprompt(dot)com>
2006-10-30 22:10:08 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-30 21:58:14 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-31 00:01:00 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-10-31 00:18:50 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-31 00:58:40 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-10-31 01:40:28 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-31 14:24:12 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-10-31 10:41:10 from Heikki Linnakangas <heikki(at)enterprisedb(dot)com>
2006-10-31 15:06:40 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-31 16:04:55 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-31 19:49:27 from Gregory Stark <stark(at)enterprisedb(dot)com>
2006-10-31 19:56:59 from Alvaro Herrera <alvherre(at)commandprompt(dot)com>
2006-10-31 21:25:18 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-31 21:10:31 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-11-01 14:50:56 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-11-02 21:50:31 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-11-02 23:57:08 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-11-03 10:40:14 from "Zeugswetter Andreas ADI SD" <ZeugswetterA(at)spardat(dot)at>
2006-11-03 13:45:04 from Alvaro Herrera <alvherre(at)commandprompt(dot)com>
2006-11-04 17:35:16 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-11-05 09:21:05 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-11-05 16:26:20 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-11-05 17:01:54 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-11-05 18:24:55 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-11-05 18:28:34 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-11-05 19:47:28 from Gregory Stark <stark(at)enterprisedb(dot)com>
2006-11-05 21:39:25 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2006-11-05 21:44:43 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-11-05 09:52:36 from "Heikki Linnakangas" <heikki(at)enterprisedb(dot)com>
2006-11-05 16:29:06 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2006-10-31 16:23:40 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Lists:
pgsql-hackers pgsql-patches
"Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
> The added WAL volume should be pretty minimal, because only tuples that have
> gone untouched for a long time incur extra work.
That seems like a weak point in the logic. It seems like it would make VACUUM
which is already an i/o hog even more so. Perhaps something clever can be done
with vacuum_cost_delay and commit_siblings.
Something like inserting the delay between WAL logging and syncing the log and
writing to the heap. So if another transaction commits in the meantime we can
skip the extra fsync and continue.
--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com
In response to
Responses
pgsql-hackers by date
Next :From: Alvaro HerreraDate: 2006-10-31 19:56:59Subject : Re: [HACKERS] WAL logging freezingPrevious :From : Alvaro HerreraDate : 2006-10-31 19:29:00Subject : Re: [HACKERS] Index greater than 8k
pgsql-patches by date
Next :From: Alvaro HerreraDate: 2006-10-31 19:56:59Subject : Re: [HACKERS] WAL logging freezingPrevious :From : Jim C. NasbyDate : 2006-10-31 17:27:27Subject : Re: --single-transaction doc clarification
