Re: Parallel Vacuum

From: Michael Stone
Subject: Re: Parallel Vacuum
Date: ,
Msg-id: 20070322184613.GJ11402@mathom.us
(view: Whole thread, Raw)
In response to: Re: Parallel Vacuum  (Dimitri)
Responses: Re: Parallel Vacuum  (Dimitri)
List: pgsql-performance

Tree view

Parallel Vacuum  (Dimitri, )
 Re: Parallel Vacuum  (Alvaro Herrera, )
  Re: Parallel Vacuum  (Dimitri, )
   Re: Parallel Vacuum  (Alvaro Herrera, )
    Re: Parallel Vacuum  (Dimitri, )
     Re: Parallel Vacuum  (Alvaro Herrera, )
     Re: Parallel Vacuum  (Michael Stone, )
      Re: Parallel Vacuum  (Dimitri, )
       Re: Parallel Vacuum  (Michael Stone, )
        Re: Parallel Vacuum  (Dimitri, )
         Re: Parallel Vacuum  (Michael Stone, )

On Thu, Mar 22, 2007 at 07:24:38PM +0100, Dimitri wrote:
>you're right until you're using a single disk :)
>Now, imagine you have more disks

I do have more disks. I maximize the I/O performance by dedicating
different sets of disks to different tables. YMMV. I do suggest watching
your I/O rates and wallclock time if you try this to see if your
aggregate is actually substantially faster than the single case. (I
assume that you haven't yet gotten far enough to actually do performance
testing.) You may also want to look into tuning your sequential I/O
performance.

Mike Stone


pgsql-performance by date:

From: Tino Wildenhain
Date:
Subject: Re: Performance of count(*)
From: Michael Stone
Date:
Subject: Re: Performance of count(*)