From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-qt1-f180.google.com ([209.85.160.180]:34941 "EHLO mail-qt1-f180.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726218AbeLKWzF (ORCPT ); Tue, 11 Dec 2018 17:55:05 -0500 Received: by mail-qt1-f180.google.com with SMTP id v11so18472102qtc.2 for ; Tue, 11 Dec 2018 14:55:04 -0800 (PST) Received: from lud1.home ([177.98.82.44]) by smtp.gmail.com with ESMTPSA id y14sm7590799qkb.2.2018.12.11.14.55.02 for (version=TLS1_2 cipher=ECDHE-RSA-CHACHA20-POLY1305 bits=256/256); Tue, 11 Dec 2018 14:55:03 -0800 (PST) Date: Tue, 11 Dec 2018 20:54:59 -0200 From: Luciano ES Subject: Re: I got file system corruption with XFS Message-ID: <20181211205459.106b35c9@lud1.home> In-Reply-To: <3daac366-6e0f-34c8-1a2f-e5d07ae4fee1@sandeen.net> References: <20181211183203.7fdbca0f@lud1.home> <3daac366-6e0f-34c8-1a2f-e5d07ae4fee1@sandeen.net> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: linux-xfs@vger.kernel.org On Tue, 11 Dec 2018 14:57:45 -0600, Eric Sandeen wrote: > On 12/11/18 2:32 PM, Luciano ES wrote: > > I needed to restore something from my backups, an external hard > > disk that is kept separately, always disconnected until I really > > need it. > > > > But the file system refused to be mounted: "structure needs > > cleaning," it said. > > > > I googled and didn't find much hope about it. I followed what > > little advice I found: I ran xfs_repair and it didn't work. So I > > ran it again with -L and it worked, but the software itself warned > > me that some files could not be recovered. I'll never know which > > ones. > > > > I always liked XFS and thought those dreaded days of file system > > corruption and lost files were far behind. So my only question is: > > Why does that happen? The disk is not even used 99% of the time. > > How does an XFS file system go belly up just like that? > > There is no way for us to know. You didn't provide nearly enough > information to even hazard a guess. > > But ok fine, I'll hazard a wild guess anyway: your external drive had > a corrupt log because the enclosure didn't honor a cache flush issued > by the filesystem after some previous mount. > > -Eric ************************** I understand you don't have much to work with, but I can't tell you more than I have. It happened several days ago and I didn't write anything down. As far as I can remember, there wasn't really much to note. "Structure needs cleaning" was pretty much all I was ever told. There must have been more when I ran xfs_repair, but it looked incomprehensible to me and I thought I shouldn't bother anyone else about it. Your theory about the enclosure sounds good. Do you think it is so flawed that issuing a 'sync' command manually before umounting wouldn't have made any difference? -- Luciano ES >>