| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Gregory Stark <stark(at)enterprisedb(dot)com> |
| Cc: | "Robert Haas" <robertmhaas(at)gmail(dot)com>, "ITAGAKI Takahiro" <itagaki(dot)takahiro(at)oss(dot)ntt(dot)co(dot)jp>, Postgres <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: posix_fadvise v22 |
| Date: | 2009-01-12 13:02:24 |
| Message-ID: | 27519.1231765344@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Gregory Stark <stark(at)enterprisedb(dot)com> writes:
> Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
>> 2. I fixed it so that setting effective_io_concurrency to zero disables
>> prefetching altogether; there was no way to do that in the patch as
>> submitted.
> Hm. the original intent was that effective_io_concurrency 1 meant no
> prefetching because there was only one drive.
Well, "no prefetch" is an entirely different behavior from "prefetch one
block ahead". Given the way you've defined the GUC, a setting of one
has to mean the latter. My complaint was basically that with the patch
applied, the code was physically incapable of providing the former.
Which you'd surely want if only for testing/comparison purposes.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Simon Riggs | 2009-01-12 13:04:23 | Re: Hot standby, slot ids and stuff |
| Previous Message | Alvaro Herrera | 2009-01-12 13:01:52 | Re: autovacuum and reloptions |