Received: from localhost (maia-3.hub.org [200.46.204.184]) by postgresql.org (Postfix) with ESMTP id 404619FBEDD; Thu, 26 Apr 2007 10:29:05 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.184]) (amavisd-maia, port 10024) with ESMTP id 41879-07; Thu, 26 Apr 2007 10:28:54 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.4 X-Greylist: from auto-whitelisted by SQLgrey-1.7.4 Received: from mail.potentialtech.com (internet.potentialtech.com [66.167.251.6]) by postgresql.org (Postfix) with ESMTP id 0E0059FBE9B; Thu, 26 Apr 2007 10:29:00 -0300 (ADT) Received: from vanquish.pgh.priv.collaborativefusion.com (pr40.pitbpa0.pub.collaborativefusion.com [206.210.89.202]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by mail.potentialtech.com (Postfix) with ESMTP id 83FEAEBC80; Thu, 26 Apr 2007 09:28:58 -0400 (EDT) Date: Thu, 26 Apr 2007 09:28:57 -0400 From: Bill Moran To: "Sorin N. Ciolofan" Cc: , Subject: Re: [GENERAL] pg_buffercache view Message-Id: <20070426092857.82e2da22.wmoran@potentialtech.com> In-Reply-To: <20070426130419.9371A8E40FC@mailhost.ics.forth.gr> References: <20070426083211.06c1c990.wmoran@potentialtech.com> <20070426130419.9371A8E40FC@mailhost.ics.forth.gr> X-Mailer: Sylpheed 2.3.1 (GTK+ 2.10.11; i386-portbld-freebsd6.1) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200704/248 X-Sequence-Number: 25312 In response to "Sorin N. Ciolofan" : > > I don't know the algorithm on which Postgre uses the shared buffers but I'd > like to find the principles behind it. Let's assume the following scenario: > I've set shared_buffers=3000 > At the starting of Postgres there are 115 buffers used by database A > After the execution of some processing caused by a java methodA1() > invocation, 2850 buffers are used by A. > What happens next if these 2850 buffers remains used even if the methodA1() > finished its execution? They'll be reused the next time a query needs the same data, or discarded when the buffer space is needed for something else. > Suppose that now a methodA2() invocation occurs and this method works with > database A, too. Will be the 2850 buffers reused yes > or will postgre throw an > "out of shared memory" exception? no > What happens if a methodB() invocation occurs, assuming that this method > tries to work with database B? Buffers will be allocated/reallocate/reused as best fits the usage pattern of the server. > How Postgres decides the allocation of shared_buffers? They're "buffers". They follow the design of just about any other type of buffer: http://foldoc.org/index.cgi?query=buffer&action=Search "Buffers are used to decouple processes so that the reader and writer may operate at different speeds or on different sized blocks of data." > -----Original Message----- > From: pgsql-general-owner@postgresql.org > [mailto:pgsql-general-owner@postgresql.org] On Behalf Of Bill Moran > Sent: Thursday, April 26, 2007 3:32 PM > To: Sorin N. Ciolofan > Cc: pgsql-general@postgresql.org; pgsql-admin@postgresql.org > Subject: Re: [ADMIN] [GENERAL] pg_buffercache view > > In response to "Sorin N. Ciolofan" : > > > > > Hello! > > > > Do you know which could be the reasons that could conduce an application > to > > not release the shared buffers, even after the application was shut down? > > I noticed that only if a pg_ctl restart command is issued some of the > > buffers are set free. > > The reason would be "by design." > > If the server flushes its cache every time the application restarts, the > cache isn't going to be very effective. > > If PostgreSQL is using more shared buffers than you're comfortable with, > reduce the shared_buffers setting in the config. That will allow the OS > to decide how to use the memory instead. > > -- > Bill Moran > http://www.potentialtech.com > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Have you searched our list archives? > > http://archives.postgresql.org/ > > -- Bill Moran http://www.potentialtech.com