From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S265188AbTLZObN (ORCPT ); Fri, 26 Dec 2003 09:31:13 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S265191AbTLZObN (ORCPT ); Fri, 26 Dec 2003 09:31:13 -0500 Received: from sina187-143.sina.com.cn ([202.106.187.143]:39953 "HELO sina.com") by vger.kernel.org with SMTP id S265188AbTLZObL (ORCPT ); Fri, 26 Dec 2003 09:31:11 -0500 Date: Fri, 26 Dec 2003 22:30:21 +0800 From: dlion X-Mailer: The Bat! (v2.00) Reply-To: dlion2004@sina.com.cn X-Priority: 3 (Normal) Message-ID: <17140565218.20031226223021@sina.com.cn> To: lkml Subject: Re: filesystem bug? In-Reply-To: <8BD3ED34-37A6-11D8-A9B5-00039341E01A@ybb.ne.jp> References: <8BD3ED34-37A6-11D8-A9B5-00039341E01A@ybb.ne.jp> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hello , Friday, December 26, 2003, 9:22:25 PM, you wrote: Bxynj> Hi, >>I got other errors on ext3 filesystem include: >>1. missing file >>2. corrupted file >>but when I used fsck.ext3 to check the ramdisk, the result was clean. Bxynj> Dlion, how did the corrupted file look like? Bxynj> (its file size, number of blocks etc.) 1. some corrupted files is truncated to 0 bytes. Blockcount is 0. 2. some corrupted files is truncated . the result is a shorter file. the new size is multiple of block size. 3. maybe all corrupted files' mtime is exactly the same wrong value. Should be around 2003.12.26 21:30:00, but is 2002.05.12 12:00:48(hex value is 0x3cdde8f0) . ctime and atime is correct. The system's clock time is unchanged. 4. it seems that the corrupted files tends to exist in the same directory. Use a 128000k bytes ramdisk you can get these results in less than 30 minutes. BTW, your test script is very good. Thank you.