From: Eric Sandeen Subject: Re: ext4: (2.6.34-rc4): This should not happen!! Data will be lost Date: Fri, 16 Apr 2010 10:02:14 -0500 Message-ID: <4BC87BF6.2040104@redhat.com> References: <20100416123526.GW21495@skl-net.de> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: linux-ext4@vger.kernel.org, linux-driver@qlogic.com, Thomas Helle To: Andre Noll Return-path: Received: from mx1.redhat.com ([209.132.183.28]:40717 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753316Ab0DPPCg (ORCPT ); Fri, 16 Apr 2010 11:02:36 -0400 In-Reply-To: <20100416123526.GW21495@skl-net.de> Sender: linux-ext4-owner@vger.kernel.org List-ID: On 04/16/2010 07:35 AM, Andre Noll wrote: > Hi > > I'm seeing the following errors when writing to an ext4 file system: > > qla2xxx 0000:06:09.0: scsi(0:0:0): Abort command issued -- 1 fa6a73 2002. > end_request: I/O error, dev sda, sector 7812771880 Sounds like a hardware failure > Aborting journal on device sda-8. > EXT4-fs (sda): delayed block allocation failed for inode 752 at logical offset 1982464 with max blocks 17588 with error -30 > > This should not happen!! Data will be lost ext4 is just the messenger here, I think. -Eric > EXT4-fs error (device sda) in ext4_da_writepages: Journal has aborted > EXT4-fs error (device sda): ext4_journal_start_sb: Detected aborted journal > EXT4-fs (sda): Remounting filesystem read-only > EXT4-fs (sda): delayed block allocation failed for inode 756 at logical offset 177656 with max blocks 1205 with error -30 > EXT4-fs error (device sda) in ext4_da_write_end: IO failure > > This should not happen!! Data will be lost > EXT4-fs (sda): ext4_da_writepages: jbd2_start: 1205 pages, ino 756; err -30 > > EXT4-fs (sda): ext4_da_writepages: jbd2_start: 17588 pages, ino 752; err -30 > > EXT4-fs (sda): delayed block allocation failed for inode 755 at logical offset 669696 with max blocks 3268 with error -30 > > This should not happen!! Data will be lost > EXT4-fs (sda): ext4_da_writepages: jbd2_start: 3268 pages, ino 755; err -30 > > This happened while stress-testing the machine [1] with > > stress -d 5 --hdd-bytes 10G --hdd-noclean > > This command is supposed to fill the disk and then exit with "no space > left". Instead, the above errors occured when the fs was 100% full. > There's nothing in the logs of the Raid system. > > Is this a known issue? Who's to blame, qla2xxx or ext4? > > Thanks > Andre > > [1] Vanilla 2.6.34-rc4, a 7.3T ext4 fs on a FC Infortrend hardware > raid, connected to a qlogic hba. Let me know if you need more info.