| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> |
| Cc: | "P(dot)J(dot) \"Josh\" Rovero" <rovero(at)sonalysts(dot)com>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: TOAST performance (was Re: [GENERAL] Delete Performance) |
| Date: | 2001-11-22 01:40:46 |
| Message-ID: | 27152.1006393246@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general pgsql-hackers |
Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> I was thinking of just throwing the inactive page number into WAL. Much
> smaller than the entire page image. You don't touch the page. Does
> that help?
I don't think so. Somehow you have to tell the other backends that that
page is dead; merely recording it in WAL doesn't do that.
More to the point, you can't recycle (overwrite) that page until you've
checkpointed or WAL-logged the replacement page; so you still end up
with disk I/O for the replacement.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Hiroshi Inoue | 2001-11-22 02:00:09 | Re: Determining which index to create |
| Previous Message | Dado Feigenblatt | 2001-11-22 01:32:32 | backup: restoring, problems with permissions |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2001-11-22 01:46:34 | Re: Mention NLS option in INSTALL |
| Previous Message | Bruce Momjian | 2001-11-22 01:34:11 | Re: Quick question |