| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Hannu Krosing <hannu(at)tm(dot)ee> |
| Cc: | jim(at)nasby(dot)net, pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: More tablescanning fun |
| Date: | 2003-04-25 20:10:47 |
| Message-ID: | 29560.1051301447@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
Hannu Krosing <hannu(at)tm(dot)ee> writes:
> I have been pondering if keeping pages half-empty (or even 70% empty)
> could solve both clustering problems and longish updates for much data.
You could achieve that pretty easily if you simply don't ever VACUUM
FULL ;-)
UPDATE has always (AFAIR) attempted to place the new version on the same
page as the old, moving it elsewhere only if it doesn't fit. So that
part of the logic is already there.
> So "VACUUM FULL 65% EMPTY;" could make sense ?
Not so much that, as a parameter to CLUSTER telling it to fill pages
only x% full.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Cecilia Alvarez | 2003-04-25 22:33:13 | Indexes with different datatypes |
| Previous Message | Jan Wieck | 2003-04-25 19:52:16 | Re: [PERFORM] Foreign key performance |