From owner-freebsd-stable@FreeBSD.ORG Wed Nov 6 14:45:19 2013 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [8.8.178.115]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTP id 2ACCC416 for ; Wed, 6 Nov 2013 14:45:19 +0000 (UTC) (envelope-from jpc@porterclark.com) Received: from mail.cluebytwelve.org (clueby12.org [198.186.190.239]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.freebsd.org (Postfix) with ESMTPS id 02E542456 for ; Wed, 6 Nov 2013 14:45:17 +0000 (UTC) Received: from 24-183-226-240.dhcp.kgpt.tn.charter.com (24-183-226-240.dhcp.kgpt.tn.charter.com [24.183.226.240]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by mail.cluebytwelve.org (Postfix) with ESMTPSA id DF2551B5DA for ; Wed, 6 Nov 2013 08:35:51 -0600 (CST) Received: by 24-183-226-240.dhcp.kgpt.tn.charter.com (Postfix, from userid 501) id AC994D4A26E9; Wed, 6 Nov 2013 08:34:42 -0600 (CST) Date: Wed, 6 Nov 2013 08:34:42 -0600 From: "J. Porter Clark" To: freebsd-stable@freebsd.org Subject: Re: FreeBSD 9.2 UFS + GELI softdep_deallocate_dependencies: unrecovered I/O error Message-ID: <20131106143442.GA29775@MacPorter.local> Mail-Followup-To: freebsd-stable@freebsd.org References: <2AA765E7-1F17-4C6F-98BD-004AEFF88D32@lexasoft.ru> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <2AA765E7-1F17-4C6F-98BD-004AEFF88D32@lexasoft.ru> Organization: http://www.angelfire.com/ego/porterclark/ User-Agent: Mutt/1.5.21 (2010-09-15) X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 06 Nov 2013 14:45:19 -0000 On Sat, Oct 26, 2013 at 01:47:18PM +0400, Alexey Tarasov wrote: > > I've upgraded server to 9.2 and now it hangs every 2-3 hours > of intensive I/O to UFS SUJ + GELI disk. On 9.1 everything was > good for a half of a year. I'm glad this is not just my problem. I've been collecting data on it for a couple of months and still can't figure out how to fill out a bug report. Too much data, so little of which is potentially useful. My situation is: A Windows 7 PC running either Outlook 2010 or SCANPST.EXE therefrom, attempting to "repair" a .pst file that will exceed 2^31-1 bytes on completion. i386 9.2-STABLE server (currently r256846) Samba 3.6 UFS (either 1 or 2) GELI GPT partition No other combination can be made to produce this behavior. In particular, changing from UFS to ZFS (even on this 2 GB i386 system) fixes it. I cannot reproduce the problem by running a program on the server; apparently only smbd has the necessary mojo. Adding data authentication to GELI doesn't help. Tweaking block sizes in UFS or GELI doesn't help. Turning off soft updates doesn't help. Samba AIO is as off as I can get it. During the repair (writing) phase, the g_vfs_dones will hit, and the system will be useless until either they stop or the system panics--about equal likelihood. > g_vfs_done():da1.eli[WRITE(offset=614630752256, length=32768)]error = 11 Oh, man, have I got a fine collection of these! I'll try the tunable. Seems odd that a tunable would fix it, though. Yes, it has driven me completely crazy. -- J. Porter Clark