From:
Greg Stark <gsstark(at)mit(dot)edu>
To:
Greg Smith <greg(at)2ndquadrant(dot)com>
Cc:
Bruce Momjian <bruce(at)momjian(dot)us>, Dimitri Fontaine <dfontaine(at)hi-media(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Josh Berkus <josh(at)agliodbs(dot)com>, pgsql-hackers(at)postgresql(dot)org
Subject:
Re: Hot Standby query cancellation and Streaming Replication integration
Date:
2010-02-27 03:40:57
Message-ID:
407d949e1002261940n43c521d6i1bb3220997886e09@mail.gmail.com (view raw or flat )
Thread:
2010-02-26 08:33:35 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-26 12:10:07 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-26 18:53:16 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2010-02-26 19:01:30 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-26 19:06:51 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-02-26 19:16:40 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2010-02-26 20:02:08 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-02-26 20:22:56 from "Joshua D(dot) Drake" <jd(at)commandprompt(dot)com>
2010-02-26 20:04:23 from Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
2010-02-26 20:21:05 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2010-02-26 20:10:11 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-26 20:25:33 from Aidan Van Dyk <aidan(at)highrise(dot)ca>
2010-02-26 20:30:01 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2010-02-26 21:06:10 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-02-26 21:11:32 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-26 21:19:39 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2010-02-26 21:22:47 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-26 21:44:20 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2010-02-27 04:02:24 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-28 21:58:51 from Simon Riggs <simon(at)2ndQuadrant(dot)com>
2010-02-27 06:53:24 from Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
2010-02-26 21:39:12 from Dimitri Fontaine <dfontaine(at)hi-media(dot)com>
2010-02-26 21:48:38 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-26 22:11:28 from Dimitri Fontaine <dfontaine(at)hi-media(dot)com>
2010-02-26 22:20:37 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-27 09:07:32 from Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
2010-02-27 09:33:34 from Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
2010-02-26 23:56:55 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 00:36:57 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-27 02:43:58 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 03:40:57 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-27 04:38:38 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 04:53:48 from Aidan Van Dyk <aidan(at)highrise(dot)ca>
2010-02-27 05:08:35 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 20:31:06 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-02-27 22:09:59 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 22:26:22 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-02-28 02:14:28 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-28 05:28:17 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-28 06:01:49 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-28 06:18:00 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 00:43:48 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-27 01:53:13 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 02:30:56 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-27 03:11:49 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 22:28:49 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-02-27 23:49:09 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 03:59:41 from "Joshua D(dot) Drake" <jd(at)commandprompt(dot)com>
2010-02-27 04:31:12 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 14:34:08 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-26 23:33:57 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 00:21:11 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-02-27 06:59:10 from Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
2010-02-27 14:56:51 from Tatsuo Ishii <ishii(at)postgresql(dot)org>
2010-02-28 01:00:52 from Robert Haas <robertmhaas(at)gmail(dot)com>
2010-02-28 06:07:20 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-28 13:54:48 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-28 15:56:28 from Joachim Wieland <joe(at)mcknight(dot)de>
2010-02-28 19:47:38 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-02-28 20:48:51 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-28 22:38:28 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-03-01 03:00:09 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-03-01 17:50:52 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-03-01 18:26:09 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-03-01 18:46:29 from Greg Stark <gsstark(at)mit(dot)edu>
2010-03-01 19:21:48 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-03-01 19:32:24 from Greg Stark <gsstark(at)mit(dot)edu>
2010-03-01 19:35:37 from Stefan Kaltenbrunner <stefan(at)kaltenbrunner(dot)cc>
2010-03-01 19:43:13 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2010-03-01 20:04:41 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-03-02 01:34:53 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-03-02 20:57:39 from Simon Riggs <simon(at)2ndQuadrant(dot)com>
2010-03-02 20:58:03 from Simon Riggs <simon(at)2ndQuadrant(dot)com>
2010-03-02 21:36:52 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-03-02 23:39:10 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-03-01 03:12:22 from Robert Haas <robertmhaas(at)gmail(dot)com>
2010-03-01 22:32:01 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-03-01 22:52:50 from "Kevin Grittner" <Kevin(dot)Grittner(at)wicourts(dot)gov>
2010-03-02 00:46:22 from Greg Stark <stark(at)mit(dot)edu>
2010-03-02 01:04:11 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2010-03-02 02:08:27 from Aidan Van Dyk <aidan(at)highrise(dot)ca>
2010-03-02 01:50:28 from Robert Haas <robertmhaas(at)gmail(dot)com>
2010-03-02 06:21:10 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-03-02 04:56:45 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-28 20:54:52 from Dimitri Fontaine <dfontaine(at)hi-media(dot)com>
2010-02-28 21:42:00 from Joachim Wieland <joe(at)mcknight(dot)de>
2010-02-28 20:21:23 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-03-02 04:54:04 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-03-02 05:50:08 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-03-02 18:30:15 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-03-02 18:34:33 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-03-02 20:11:00 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-03-02 23:44:16 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-03-10 06:29:19 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-03-10 11:38:12 from Greg Stark <gsstark(at)mit(dot)edu>
2010-03-10 18:59:50 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-03-11 00:51:29 from Fujii Masao <masao(dot)fujii(at)gmail(dot)com>
2010-03-11 01:39:37 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-03-11 05:58:35 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-03-02 19:03:14 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-03-02 21:00:24 from Simon Riggs <simon(at)2ndQuadrant(dot)com>
2010-02-26 14:10:26 from Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
2010-02-26 14:24:30 from Richard Huxton <dev(at)archonet(dot)com>
2010-02-26 18:47:09 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-27 00:11:36 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 00:54:57 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-27 02:52:12 from Greg Smith <greg(at)2ndquadrant(dot)com>
2010-02-27 02:56:01 from Bruce Momjian <bruce(at)momjian(dot)us>
2010-02-26 14:17:23 from Richard Huxton <dev(at)archonet(dot)com>
2010-02-26 14:45:07 from Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
2010-02-26 14:57:16 from Richard Huxton <dev(at)archonet(dot)com>
2010-02-26 15:21:03 from Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
2010-02-26 17:45:33 from Robert Haas <robertmhaas(at)gmail(dot)com>
2010-02-26 19:02:12 from Josh Berkus <josh(at)agliodbs(dot)com>
2010-02-26 16:43:24 from Richard Huxton <dev(at)archonet(dot)com>
2010-02-26 17:46:15 from Greg Stark <gsstark(at)mit(dot)edu>
2010-02-28 22:15:18 from Simon Riggs <simon(at)2ndQuadrant(dot)com>
Lists:
pgsql-hackers
On Sat, Feb 27, 2010 at 2:43 AM, Greg Smith <greg(at)2ndquadrant(dot)com> wrote:
>
> But if you're running the 8 hour report on the master right now, aren't you
> already exposed to a similar pile of bloat issues while it's going? If I
> have the choice between "sometimes queries will get canceled" vs. "sometimes
> the master will experience the same long-running transaction bloat issues as
> in earlier versions even if the query runs on the standby", I feel like
> leaning toward the latter at least leads to a problem people are used to.
If they move from running these batch queries on the master to running
them on the slave then sure, the situation will be no worse than
before.
But if they move from having a plain old PITR warm standby to having
one they can run queries on they might well assume that the big
advantage of having the standby to play with is precisely that they
can do things there that they have never been able to do on the master
previously without causing damage.
I agree that having queries randomly and unpredictably canceled is
pretty awful. My argument was that max_standby_delay should default to
infinity on the basis that any other value has to be picked based on
actual workloads and SLAs.
My feeling is that there are probably only two types of configurations
that make sense, a HA replica with a low max_standby_delay or a
reporting replica with a high max_standby_delay. Any attempt to
combine the two on the same system will only work if you know your
application well and can make compromises with both.
I would also like to be able to handle load balancing read-only
queries but that will be really tricky. You want up-to-date data and
you want to be able to run moderately complex queries. That kind of
workload may well require synchronous replication to really work
properly.
--
greg
In response to
Responses
pgsql-hackers by date
Next :From: Дмитрий ФефеловDate: 2010-02-27 03:59:01Subject : Re: Alpha4 Available NowPrevious :From : Bruce MomjianDate : 2010-02-27 03:38:49Subject : Re: ALTER ROLE/DATABASE RESET ALL versus security
