From:
"Zeugswetter Andreas ADI SD" <ZeugswetterA(at)spardat(dot)at>
To:
"Galy Lee" <lee(dot)galy(at)oss(dot)ntt(dot)co(dot)jp>,"Simon Riggs" <simon(at)2ndquadrant(dot)com>
Cc:
"Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us>,"Jim C(dot) Nasby" <jim(at)nasby(dot)net>,<pgsql-hackers(at)postgresql(dot)org>
Subject:
Re: Resumable vacuum proposal and design overview
Date:
2007-03-01 13:28:17
Message-ID:
E1539E0ED7043848906A8FF995BDA57901CAFFC5@m0143.s-mxs.net (view raw or flat )
Thread:
2007-02-26 09:21:50 from Galy Lee <lee(dot)galy(at)oss(dot)ntt(dot)co(dot)jp>
2007-02-26 10:20:09 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2007-02-27 02:44:28 from Galy Lee <lee(dot)galy(at)oss(dot)ntt(dot)co(dot)jp>
2007-02-27 03:14:47 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2007-02-27 16:37:45 from "Jim C(dot) Nasby" <jim(at)nasby(dot)net>
2007-02-27 16:49:44 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2007-02-27 17:11:04 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2007-02-27 17:15:52 from "Matthew T(dot) O'Connor" <matthew(at)tocr(dot)com>
2007-02-27 17:23:30 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2007-02-27 18:15:50 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2007-02-28 01:32:20 from Galy Lee <lee(dot)galy(at)oss(dot)ntt(dot)co(dot)jp>
2007-02-28 03:48:25 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2007-02-28 04:53:01 from Galy Lee <lee(dot)galy(at)oss(dot)ntt(dot)co(dot)jp>
2007-02-28 05:10:57 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2007-02-28 09:38:57 from Heikki Linnakangas <heikki(at)enterprisedb(dot)com>
2007-02-28 11:34:27 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2007-02-28 13:19:50 from Gregory Stark <stark(at)enterprisedb(dot)com>
2007-02-28 22:14:24 from Heikki Linnakangas <heikki(at)enterprisedb(dot)com>
2007-03-01 15:10:32 from "Jim C(dot) Nasby" <jim(at)nasby(dot)net>
2007-03-01 15:11:11 from Heikki Linnakangas <heikki(at)enterprisedb(dot)com>
2007-02-28 08:36:31 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2007-02-28 08:49:55 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2007-02-28 10:02:45 from Galy Lee <lee(dot)galy(at)oss(dot)ntt(dot)co(dot)jp>
2007-02-28 10:34:14 from "Zeugswetter Andreas ADI SD" <ZeugswetterA(at)spardat(dot)at>
2007-02-28 10:19:44 from "Zeugswetter Andreas ADI SD" <ZeugswetterA(at)spardat(dot)at>
2007-02-28 11:26:29 from "Simon Riggs" <simon(at)2ndquadrant(dot)com>
2007-03-01 04:48:50 from Galy Lee <lee(dot)galy(at)oss(dot)ntt(dot)co(dot)jp>
2007-03-01 06:17:03 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2007-03-01 13:28:17 from "Zeugswetter Andreas ADI SD" <ZeugswetterA(at)spardat(dot)at>
2007-03-01 13:52:16 from "Zeugswetter Andreas ADI SD" <ZeugswetterA(at)spardat(dot)at>
2007-02-26 10:45:56 from tomas(at)tuxteam(dot)de
2007-02-26 18:39:40 from Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
2007-02-28 08:30:10 from tomas(at)tuxteam(dot)de
Lists:
pgsql-hackers
> I admit that the implementation is much complex, but I can
> not see any big problems to save the dead tuples out and read
> it in again(like two phase commit does). Why do we need to
> hold the lock and transaction? We can open the lock and
> abandon the transaction ID, vacuum can take the lock and get
> a new ID when restarting. Why do we need to worry about if
> the dead tuple is still alive, only vacuum will sweep them,
> HOT can not touch the tuple until we have finished sweeping.
One imho important (not necessarily mandatory) aspect of HOT is, that it
does parts of what vacuum would usually do.
Thus:
1. resume, load ctid list
2. continue filling ctid list
3. remove index tuples for these ctids (* problem *)
You have just removed index entries for possibly now live tuples that
have been reused by HOT.
Unless ...
Andreas
In response to
Responses
pgsql-hackers by date
Next :From: Merlin MoncureDate: 2007-03-01 13:32:57Subject : Re: [HACKERS] urgent: upgraded to 8.2, getting kernel panicsPrevious :From : Zeugswetter Andreas ADI SDDate : 2007-03-01 13:24:47Subject : Re: COMMIT NOWAIT Performance Option
