From owner-freebsd-fs@FreeBSD.ORG Tue May 4 16:16:41 2010 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 34BAF106566C for ; Tue, 4 May 2010 16:16:41 +0000 (UTC) (envelope-from Lassi.Tuura@cern.ch) Received: from CERNMX30.cern.ch (cernmx30.cern.ch [137.138.144.177]) by mx1.freebsd.org (Postfix) with ESMTP id BDA608FC20 for ; Tue, 4 May 2010 16:16:40 +0000 (UTC) Received: from cernfe12.cern.ch (137.138.140.37) by cernmxgwlb2.cern.ch (137.138.144.177) with Microsoft SMTP Server (TLS) id 14.0.639.21; Tue, 4 May 2010 18:16:36 +0200 Received: from cmsmac01.cern.ch (137.138.52.75) by smtp.cern.ch (137.138.140.59) with Microsoft SMTP Server (TLS) id 8.2.176.0; Tue, 4 May 2010 18:16:35 +0200 MIME-Version: 1.0 (Apple Message framework v1078) Content-Type: text/plain; charset="us-ascii" From: Lassi Tuura In-Reply-To: <20100504004905.GA12233@icarus.home.lan> Date: Tue, 4 May 2010 18:16:35 +0200 Content-Transfer-Encoding: 7bit Message-ID: <43B635E5-A901-4EA4-AE03-8A61E8F3F200@cern.ch> References: <4529AF96-4BFF-4424-B77F-FE5BC8AE43D3@cern.ch> <20100504004905.GA12233@icarus.home.lan> To: Jeremy Chadwick X-Mailer: Apple Mail (2.1078) Keywords: CERN SpamKiller Note: -50 Cc: freebsd-fs@freebsd.org Subject: Re: Instant crash with ZFS + iozone? X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 04 May 2010 16:16:41 -0000 Hi, > This sounds like a hardware problem to me -- particularly excessive draw > on the PSU, voltage issues, or other whatnots. Especially if the screen > goes black (no cursor, monitor not in power save mode) and requires a > hard reset. Possibly the problem is only witnessed under extreme disk > I/O combined with high CPU usage + memory I/O. Thanks, that does look like plausible explanation here. I'll see what I can do. As you suggested, I made some tests with various number of disks and read/write patterns; ZFS is totally peripheral to this, I can get dd to induce similar failures. Two drives is ok, three only if there's not a lot of seeking, four only if two are reading and two are writing sequentially (no seeking at all). Selectively taking drives out and/or rearranging drives doesn't change this - it's not an issue with any one drive that I can tell - so will focus on PSU next. (Re Josh' reply I was in fact running iozone as root. Will keep in mind.) Regards, Lassi