| From: | Bruce Momjian <bruce(at)momjian(dot)us> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | "Jim C(dot) Nasby" <jim(at)nasby(dot)net>, Ron Mayer <rm_pg(at)cheapcomplexdevices(dot)com>, Gregory Stark <stark(at)enterprisedb(dot)com>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: [GENERAL] Autovacuum Improvements |
| Date: | 2007-01-26 04:39:00 |
| Message-ID: | 200701260439.l0Q4d0P04691@momjian.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general pgsql-hackers |
Tom Lane wrote:
> Bruce Momjian <bruce(at)momjian(dot)us> writes:
> > Jim C. Nasby wrote:
> >> If we extended relations by more than one page at a time we'd probably
> >> have a better shot at the blocks on disk being contiguous and all read
> >> at the same time by the OS.
>
> > Actually, there is evidence that adding only a single page to the end
> > causes a lot of contention for that last page, and that adding a few
> > might be better.
>
> Evidence where? The code is designed so that the last page *isn't*
> shared --- go read the comments in hio.c sometime.
I was talking about the last page of a table, where INSERTs all cluster
on that last page and cause lots of page locking. hio.c does look like
it avoids that problem.
--
Bruce Momjian bruce(at)momjian(dot)us
EnterpriseDB http://www.enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2007-01-26 04:41:17 | Re: bytea performance issue |
| Previous Message | Jim Nasby | 2007-01-26 04:32:25 | Re: Converting 7.x to 8.x |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Bruce Momjian | 2007-01-26 04:41:30 | Re: Proposal: Commit timestamp |
| Previous Message | Tom Lane | 2007-01-26 04:13:20 | Re: [GENERAL] Autovacuum Improvements |