Check out the "Inside the PostgreSQL Buffer Cache" link here:
Thanks to Greg Smith (active here).
On Fri, May 27, 2011 at 3:36 PM, preetika tyagi <preetikatyagi(at)gmail(dot)com>wrote:
> Hi All,
> I am little confused about the internal working of PostgreSQL. There is a
> parameter shared_buffer in postgres.conf and I am assuming that it is used
> for buffer management in PostgreSQL. If there is a need to bring in a new
> page in the buffer and size exceeds the shared_buffer limit, a victim dirty
> page will be written back to the disk.
> However, I have read on many links that PostgreSQL depends on the OS for
> caching. (http://www.varlena.com/GeneralBits/Tidbits/perf.html#shbuf)
> So my question is, the actual limit of the shared buffer will be defined by
> OS or the shared_buffer parameter in the postgres.conf to figure whether a
> victim dirty page needs to be selected for disk write or not?
In response to
pgsql-general by date
|Next:||From: Little, Douglas||Date: 2011-05-27 21:14:08|
|Subject: String library|
|Previous:||From: David Johnston||Date: 2011-05-27 20:48:42|
|Subject: Re: SELECT COUNT(*) execution time on large tables (v9.0.4-1)|