| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | shridhar_daithankar(at)persistent(dot)co(dot)in |
| Cc: | pgsql-general(at)postgresql(dot)org |
| Subject: | Re: Urgent need of (paid) PostgreSQL support in New |
| Date: | 2002-12-12 05:28:40 |
| Message-ID: | 3361.1039670920@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
"Shridhar Daithankar" <shridhar_daithankar(at)persistent(dot)co(dot)in> writes:
> On 11 Dec 2002 at 12:52, Nicolai Tufar wrote:
>> OS caching is generally considered a waste of resource in databases.
>> Try to allocate as much as possible to shared buffers and set OS caching
>> to minimum.
> That is an exactly opposite of the truth. Leave as much for OS cache and do
> minimum use of shared buffers.
There are varying opinions about that. Some say "push PG's
shared-buffers setting as high as you can make it". Some say "use a
conservatively small shared-buffers setting and expect the kernel to use
the rest of physical memory as kernel disk buffers". But both camps
agree that a shared-buffers setting near 50% of physical RAM will suck:
then any given page of database is likely to get cached *both* in PG's
buffers and in kernel buffers. That behavior can't win by any measure.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2002-12-12 05:41:33 | Re: Docs: GIST |
| Previous Message | Tom Lane | 2002-12-12 05:13:28 | Re: infinity as a date |