From owner-freebsd-current Mon Jul 6 08:33:30 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA25691 for freebsd-current-outgoing; Mon, 6 Jul 1998 08:33:30 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from dot.crosswinds.net (dot.crosswinds.net [209.47.139.147]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA25670; Mon, 6 Jul 1998 08:33:20 -0700 (PDT) (envelope-from tony@crosswinds.net) Received: (from tony@localhost) by dot.crosswinds.net (8.8.7/8.8.8) id LAA27822; Mon, 6 Jul 1998 11:33:26 -0400 (EDT) (envelope-from tony) From: Tony Holmes Message-Id: <199807061533.LAA27822@dot.crosswinds.net> Subject: Load related wd problem To: freebsd-stable@FreeBSD.ORG, freebsd-current@FreeBSD.ORG Date: Mon, 6 Jul 1998 11:33:25 -0400 (EDT) X-Mailer: ELM [version 2.4ME+ PL31H (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Greetings, I posted to -stable and -current since these seem to be the most appropriate places for this question - please correct me if this is wrong. I run a Network Service Provider and utilize FreeBSD exclusively on all my systems - I have come to trust and rely on FreeBSD's performance and stability. My systems range from 2.2.1 to 2.2.6 (with some stable patches thrown in) and provide a myriad of services, but I have only recently been able to verify the existence of a problem with heavy access to IDE disks. When the IDE disks are under heavy load, they will take down the system with no warning. This occurs across all versions - it find it easy to duplicate the problem - on a web server, while the http server is up, I will do some custom processing on the log files (cat, grepping, etc.) which hits the IDE disks hard. After a time (usually after 10-15 minutes) the system hangs and can only be rescued with a hard reset. I've seen the same effect on a system set up as a Squid proxy with 2 ide drives - loggings to one disk and caching on the other, where my peak wd interrupt rates reach 150-200/sec - this machine will hang after anywhere between 1.5 hours and 2 days of uptime. If it makes any difference, I use Seagate IDE disks on all my systems. Unfortunately, due to the nature of the crash and my limited resources, I cannot isolate the problem further, but am very curious as to whether or not anyone else has seen similar problems? Tony Holmes Senior Sysadmin Crosswinds Internet Communications Inc. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message