X-Original-To: Pgsql-performance@postgresql.org Received: from torque.intervideoinc.com (mail.intervideo.com [206.112.112.151]) by postgresql.org (Postfix) with ESMTP id 724A7475E91 for ; Mon, 17 Mar 2003 15:14:56 -0500 (EST) Received: from ronpc [63.68.5.2] by torque.intervideoinc.com (SMTPD32-5.05) id A16640B0082; Mon, 17 Mar 2003 12:34:46 -0800 From: "Ron Mayer" To: "Josh Berkus" , Cc: Subject: Re: Performance on large data transformations Date: Mon, 17 Mar 2003 12:06:39 -0800 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 In-Reply-To: <200303170938.38365.josh@agliodbs.com> Importance: Normal X-Archive-Number: 200303/124 X-Sequence-Number: 1422 Josh Berkus wrote: > >There is, for that matter, no kernel swap activity, but I'm not >sure how to measure Postgres temp file activity. Of course you could: mv /wherever/data/base/16556/pgsql_tmp /some_other_disk/ ln -s /some_other_disk/pgsql_tmp /wherever/data/base/16556 and use "iostat" from the "systat" package to watch how much you're using the disk the temp directory's on. In fact, for OLAP stuff I've had this help performance because quite a few data warehousing operations look like: First read from main database, do a big hash-or-sort, (which gets written to pgsql_tmp), then read from this temporary table and write result to main database PS: Yes, I know this doesn't help the FSM stuff you asked about. Ron