From: Theodore Tso Subject: Re: More ext4 acl/xattr corruption - 4th occurence now Date: Tue, 19 May 2009 07:36:44 -0400 Message-ID: <20090519113644.GA7927@mit.edu> References: <20090513062634.GE4972@kulgan> <20090514044011.GC11352@mit.edu> <20090514110659.GA5146@kulgan> <20090514132506.GD5146@kulgan> <20090514140732.GI11352@mit.edu> <20090514143014.GH5146@kulgan> <20090514161254.GJ11352@mit.edu> <20090514210244.GL5146@kulgan> <20090514212325.GG21316@mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Kevin Shanahan , Andreas Dilger , Alex Tomas , linux-ext4@vger.kernel.org To: Thierry Vignaud Return-path: Received: from THUNK.ORG ([69.25.196.29]:38910 "EHLO thunker.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750971AbZESLg7 (ORCPT ); Tue, 19 May 2009 07:36:59 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: On Tue, May 19, 2009 at 12:00:56PM +0200, Thierry Vignaud wrote: > I wonder if that would explain the corruption I reported a couple weeks > ago. > > Now I remember I wrongly got 2 parallel cp from the same source > directory to the same target directory. > > Could this be the cause? Yes, it's possible. There could be a problem either if the two cp's tried targetting the same file at the same time, or if the directory was getting expanded at the same time by the two different processes. Since we don't cache logical->physical mapping for directories (since because of the journalling requirements directories are stored in the buffer cache, not the page cache), rather more likely to run into problems with directories; more so since directoris will tend to be fragmented. On the other hand, most of the time writes into the directory will tend to be into pre-existing free space; but if you had two parallel cp's copying a large number of files into the same directory, that could certainly happen. - Ted