Re: Read/Write block sizes (Was: Caching by Postgres)

From: Michael Stone
Subject: Re: Read/Write block sizes (Was: Caching by Postgres)
Date: ,
Msg-id: 20050823231238.GL8667@mathom.us
(view: Whole thread, Raw)
In response to: Re: Read/Write block sizes (Was: Caching by Postgres)  (Jignesh Shah)
Responses: Re: Read/Write block sizes (Was: Caching by Postgres)  ("Jeffrey W. Baker")
List: pgsql-performance

Tree view

Re: Read/Write block sizes (Was: Caching by Postgres)  (Jignesh Shah, )
 Re: Read/Write block sizes  (Chris Browne, )
  Re: Read/Write block sizes  (Michael Stone, )
  Re: Read/Write block sizes  ("Jim C. Nasby", )
   Re: Read/Write block sizes  ("Jignesh K. Shah", )
    Re: Read/Write block sizes  (Bruce Momjian, )
  Re: Read/Write block sizes  (Steve Poe, )
   Re: Read/Write block sizes  (Josh Berkus, )
    Re: Read/Write block sizes  (Alan Stange, )
    Re: Read/Write block sizes  ("Jeffrey W. Baker", )
     Re: Read/Write block sizes  (Tom Lane, )
      Re: Read/Write block sizes  (Josh Berkus, )
     Re: Read/Write block sizes  (Guy Thornley, )
      Re: Read/Write block sizes  ("Jeffrey W. Baker", )
       Re: Read/Write block sizes  (Tom Lane, )
        Re: Read/Write block sizes  ("Jeffrey W. Baker", )
         Re: Read/Write block sizes  (Josh Berkus, )
          Re: Read/Write block sizes  (Ron, )
        Re: Read/Write block sizes  (PFC, )
 Re: Read/Write block sizes (Was: Caching by Postgres)  (Michael Stone, )
  Re: Read/Write block sizes (Was: Caching by Postgres)  ("Jeffrey W. Baker", )
 Re: Read/Write block sizes  (Chris Browne, )
  Re: Read/Write block sizes  ("Jim C. Nasby", )
 Re: Read/Write block sizes  (Chris Browne, )
  Re: Read/Write block sizes  (Ron, )

On Tue, Aug 23, 2005 at 05:29:01PM -0400, Jignesh Shah wrote:
>Actually some of that readaheads,etc  the OS does  already if it does
>some sort of throttling/clubbing of reads/writes.

Note that I specified the fully cached case--even with the workload in
RAM the system still has to process a heck of a lot of read calls.

>* Introduce a multiblock or extent tunable variable where you can
>define a multiple of 8K (or BlockSize tuneable) to read a bigger chunk
>and store it in the bufferpool.. (Maybe writes too) (Most devices now
>support upto 1MB chunks for reads and writes)

Yeah. The problem with relying on OS readahead is that the OS doesn't
know whether you're doing a sequential scan or an index scan; if you
have the OS agressively readahead you'll kill your seek performance.
OTOH, if you don't do readaheads you'll kill your sequential scan
performance. At the app level you know which makes sense for each
operation.

Mike Stone


pgsql-performance by date:

From: Tom Lane
Date:
Subject: Re: Read/Write block sizes
From: Guy Thornley
Date:
Subject: Re: Read/Write block sizes