| From: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
|---|---|
| To: | Satoshi Nagayasu <nagayasus(at)nttdata(dot)co(dot)jp> |
| Cc: | josh(at)agliodbs(dot)com, pgsql-hackers(at)postgresql(dot)org, "Jim C(dot) Nasby" <jnasby(at)pervasive(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Subject: | Re: Pre-allocated free space for row updating |
| Date: | 2005-08-31 06:41:13 |
| Message-ID: | 1125470473.4010.447.camel@localhost.localdomain |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, 2005-08-31 at 08:32 +0900, Satoshi Nagayasu wrote:
> Simon Riggs wrote:
> > The summary was:
> >
> > 1. Have a PCTFREE column added on a table by table basis
>
> I think a good place to keep PCTFREE value is a new column
> in the pg_class, and ALTER TABLE should be able to change this value.
Agreed
> > 2. Apply PCTFREE for Inserts only
> > 3. Allow Updates to use the full space in the block.
>
> 4. Allow to repair fragmentation in each page.
>
> Because updates cause fragmentation in the page.
>
> So we need to keep large continuous free space in each page,
> if we want to get more effective on PCTFREE feature.
...doesn't VACUUM already do that?
Anyway, if the setting is for each specific table then the performance
benefit is very clear.
Best Regards, Simon Riggs
| From | Date | Subject | |
|---|---|---|---|
| Next Message | rm_pg | 2005-08-31 07:30:55 | ERROR: bogus varno: 1 with 8.1beta and postgis 1.0.3 |
| Previous Message | Oliver Jowett | 2005-08-31 02:34:49 | statement logging / extended query protocol issues |