Received: from u1.abs.net (root@u1.abs.net [207.114.0.131]) by hub.org (8.9.2/8.9.1) with ESMTP id TAA65638 for ; Thu, 15 Apr 1999 19:37:17 -0400 (EDT) (envelope-from byronn@insightdist.com) Received: from insightdist.com (nobody@localhost) by u1.abs.net (8.9.0/8.9.0) with UUCP id TAA24570 for pgsql-interfaces@postgreSQL.org; Thu, 15 Apr 1999 19:22:23 -0400 (EDT) X-Authentication-Warning: u1.abs.net: nobody set sender to insightdist.com!byronn using -f Received: from ceodev by insightdist.com (AIX 3.2/UCB 5.64/4.03) id AA17467; Thu, 15 Apr 1999 18:46:15 -0400 Received: from ppp0_ceodev by ceodev (AIX 4.1/UCB 5.64/4.03) id AA31040; Thu, 15 Apr 1999 18:46:13 -0400 Message-Id: <37166E09.E02332D8@insightdist.com> Date: Thu, 15 Apr 1999 18:54:01 -0400 From: Byron Nikolaidis X-Mailer: Mozilla 4.05 [en] (Win95; I) Mime-Version: 1.0 To: Denis Sbragion Cc: pgsql-interfaces@postgreSQL.org Subject: Re: [INTERFACES] ODBC Memory leak References: <3.0.6.32.19990413105717.008787e0@MBox.InfoTecna.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Denis Sbragion wrote: > Hello, > > It looks like the postgres ODBC (6.40.0005) leaks a lot of memory. Under a > heavy usage test the memory occupation increase up to 2 Mb every 30 > seconds, until everything crashes. To be sure I did exactly the same test > using the MS SQL Server ODBC driver. This driver showed no memory leakage > at all. Anybody can confirm this? > Whoa! Well, nothing in the core driver has been changed in the version 0005 release, certainly nothing related to memory usage/allocation. In fact, nothing has been changed along these lines since like version 6.30.0250 or something. Something to keep in mind though, if the driver is *NOT* using the "Use Declare/Fetch option", then the driver will attempt to store the entire query results in its memory. Is it possible your usage test would have done a huge query or something like that? The Use Declare/Fetch option only allocates a cache of 100 rows (by default) and never goes over that so huge queries wouldn't bother it. If there is a memory problem like this in the driver, I would like to know more about your test so I can track it down. If so, you might try changing some options (such as the Declare/Fetch) and see if it changes anything. Byron