Re: HOT patch - version 15

From: Bruce Momjian <bruce(at)momjian(dot)us>
To: Simon Riggs <simon(at)2ndquadrant(dot)com>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Heikki Linnakangas <heikki(at)enterprisedb(dot)com>, Florian Pflug <fgp(dot)phlo(dot)org(at)gmail(dot)com>, Pavan Deolasee <pavan(dot)deolasee(at)gmail(dot)com>, Gregory Stark <stark(at)enterprisedb(dot)com>, PostgreSQL-patches <pgsql-patches(at)postgresql(dot)org>
Subject: Re: HOT patch - version 15
Date: 2007-09-09 11:55:06
Message-ID: 200709091155.l89Bt6224828@momjian.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-patches

Simon Riggs wrote:
> On Sat, 2007-09-08 at 23:13 -0400, Bruce Momjian wrote:
> > In summary, I feel we have the HOT mechanics down well, but the open
> > issue is _when_ to activate each operation.
>
> That has been the only open issue for months. The interplay of behaviour
> is where HOT succeeds and fails. Some behaviours sound like they will be
> good, but aren't.

Looking at the patch, it seems defragmentation is checked every time a
heap page is pinned in heapam.c by calling heap_page_prune_defrag(). If
page free space < 1.2 * average_tuple_size the page is defragmented. It
would seem defragmentation is checked as much as possible. Seems
pruning is also as aggressive as possible.

I think an open question is whether defragmentation should happen _only_
when we are trying to add something to the page and it doesn't fit.
Another open question is whether pruning should happen only when the
chain gets to a certain length.

The larger question is how are we going to get these answers?

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +

In response to

Responses

Browse pgsql-patches by date

  From Date Subject
Next Message Simon Riggs 2007-09-09 12:25:44 Re: HOT patch - version 15
Previous Message Andrew Dunstan 2007-09-09 11:46:19 Re: invalidly encoded strings