Обсуждение: pgsql: TODO item not needed anymore now that the buffer cache is

Log Message:
-----------
TODO item not needed anymore now that the buffer cache is
scan-resistant:

<
< * Allow free-behind capability for large sequential scans, perhaps using
<   posix_fadvise()
<
<   Posix_fadvise() can control both sequential/random file caching and
<   free-behind behavior, but it is unclear how the setting affects other
<   backends that also have the file open, and the feature is not supported
<   on all operating systems.

Modified Files:
--------------
    pgsql/doc:
        TODO (r1.2197 -> r1.2198)
        (http://developer.postgresql.org/cvsweb.cgi/pgsql/doc/TODO.diff?r1=1.2197&r2=1.2198)
    pgsql/doc/src/FAQ:
        TODO.html (r1.698 -> r1.699)
        (http://developer.postgresql.org/cvsweb.cgi/pgsql/doc/src/FAQ/TODO.html.diff?r1=1.698&r2=1.699)

On Fri, 1 Jun 2007, Bruce Momjian wrote:

> Log Message:
> -----------
> TODO item not needed anymore now that the buffer cache is
> scan-resistant:
>
> <
> < * Allow free-behind capability for large sequential scans, perhaps using
> <   posix_fadvise()
> <
> <   Posix_fadvise() can control both sequential/random file caching and
> <   free-behind behavior, but it is unclear how the setting affects other
> <   backends that also have the file open, and the feature is not supported
> <   on all operating systems.
>

This todo item is about telling the OS cache that we don't want these
buffers kept around, not about pg's own buffer cache.  So I think it's
still valid.

Kris Jurka

Kris Jurka wrote:
>
>
> On Fri, 1 Jun 2007, Bruce Momjian wrote:
>
> > Log Message:
> > -----------
> > TODO item not needed anymore now that the buffer cache is
> > scan-resistant:
> >
> > <
> > < * Allow free-behind capability for large sequential scans, perhaps using
> > <   posix_fadvise()
> > <
> > <   Posix_fadvise() can control both sequential/random file caching and
> > <   free-behind behavior, but it is unclear how the setting affects other
> > <   backends that also have the file open, and the feature is not supported
> > <   on all operating systems.
> >
>
> This todo item is about telling the OS cache that we don't want these
> buffers kept around, not about pg's own buffer cache.  So I think it's
> still valid.

Agreed, re-added, and clarified it is for the kernel cache:

* Allow free-behind capability for large sequential scans to avoid
  kernel cache spoiling

  Posix_fadvise() can control both sequential/random file caching and
  free-behind behavior, but it is unclear how the setting affects other
  backends that also have the file open, and the feature is not supported
  on all operating systems.

--
  Bruce Momjian  <bruce@momjian.us>          http://momjian.us
  EnterpriseDB                               http://www.enterprisedb.com

  + If your life is a hard drive, Christ can be your backup. +