Received: from netcom.com (zalman@netcom18.netcom.com [199.183.9.118]) by hub.org (8.9.3/8.9.3) with ESMTP id DAA94114 for ; Wed, 17 Nov 1999 03:47:03 -0500 (EST) (envelope-from zalman@netcom.com) Received: (from zalman@localhost) by netcom.com (8.9.3/8.9.3) id AAA25336; Wed, 17 Nov 1999 00:44:19 -0800 (PST) From: Zalman Stern Message-Id: <199911170844.AAA25336@netcom.com> Subject: Re: [SQL] zombie postmasters To: tgl@sss.pgh.pa.us (Tom Lane) Date: Wed, 17 Nov 1999 00:44:15 -0800 (PST) Cc: hamid@emarq.com, pgsql-sql@postgreSQL.org In-Reply-To: <24411.942815127@sss.pgh.pa.us> from "Tom Lane" at Nov 17, 99 00:05:27 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit > "Hamid Khoshnevis" writes: > > In a web environment, a user can repeatedly starts and abort large queries > > which create a large number of zombie postmasters. The first question is whether you mean real "zombies" (as indicated by ps) or a bunch more process than you expected ("process leak"). I've seen the latter a number of times and it generally indicates failure to properly close connections in your application. I find the same techniques used for debuging file descriptor leaks work for finding these problems. -Z-