From: Jim C. Nasby <decibel@decibel.org>
To: Dave Chapeskie <pgsql@ddm.wox.org>
To: pgsql-perform <pgsql-performance@postgresql.org>
Subject: Re: Index bloat problem?
Date: Fri, 22 Apr 2005 22:12:38 -0500
Message-ID: <20050423031238.GA58835@decibel.org> (raw)
In-Reply-To: <20050421193305.GA84813@ddm.wox.org>
References: <20050421173855.76286.qmail@web51410.mail.yahoo.com>
<200504211044.48794.josh@agliodbs.com>
<4267EFF9.2000908@rentec.com>
<200504211128.43937.josh@agliodbs.com>
<20050421193305.GA84813@ddm.wox.org>
You would be interested in
http://archives.postgresql.org/pgsql-hackers/2005-04/msg00565.php
On Thu, Apr 21, 2005 at 03:33:05PM -0400, Dave Chapeskie wrote:
> On Thu, Apr 21, 2005 at 11:28:43AM -0700, Josh Berkus wrote:
> > Michael,
> >
> > > ....Every five minutes, DBCC INDEXDEFRAG will report to the user an
> > > estimated percentage completed. DBCC INDEXDEFRAG can be terminated at
> > > any point in the process, and *any completed work is retained.*"
> >
> > Keen. Sounds like something for our TODO list.
> >
> > --
> > Josh Berkus
> > Aglio Database Solutions
> > San Francisco
>
> See http://archives.postgresql.org/pgsql-general/2005-03/msg01465.php
> for my thoughts on a non-blocking alternative to REINDEX. I got no
> replies to that message. :-(
>
>
> I've almost got a working solution integrated in the backend that does
> correct WAL logging and everything. (Writing the code to write and
> replay WAL logs for complicated operations can be very annoying!)
>
> For now I've gone with a syntax of:
>
> REINDEX INDEX btree_index_name INCREMENTAL;
>
> (For now it's not a proper index AM (accessor method), instead the
> generic index code knows this is only supported for btrees and directly
> calls the btree_compress function.)
>
> It's not actually a REINDEX per-se in that it doesn't rebuild the whole
> index. It holds brief exclusive locks on the index while it shuffles
> items around to pack the leaf pages fuller. There were issues with the
> code I attached to the above message that have been resolved with the
> new code. With respect to the numbers provided in that e-mail the new
> code also recycles more pages than before.
>
> Once I've finished it up I'll prepare and post a patch.
>
> --
> Dave Chapeskie
> OpenPGP Key ID: 0x3D2B6B34
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
>
--
Jim C. Nasby, Database Consultant decibel@decibel.org
Give your computer some brain candy! www.distributed.net Team #1828
Windows: "Where do you want to go today?"
Linux: "Where do you want to go tomorrow?"
FreeBSD: "Are you guys coming, or what?"
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-performance@postgresql.org
Cc: decibel@decibel.org, pgsql@ddm.wox.org
Subject: Re: Index bloat problem?
In-Reply-To: <20050423031238.GA58835@decibel.org>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox