From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <3F2A4A76.9070003@conterra.de> From: Dieter Stueken MIME-Version: 1.0 Subject: Re: [linux-lvm] pvmove and badsectors Content-Transfer-Encoding: 8bit Sender: linux-lvm-admin@sistina.com Errors-To: linux-lvm-admin@sistina.com Reply-To: linux-lvm@sistina.com List-Help: List-Post: List-Subscribe: , List-Unsubscribe: , List-Archive: Date: Fri Aug 1 06:11:01 2003 List-Id: Content-Type: text/plain; charset="utf-8"; format="flowed" To: linux-lvm@sistina.com Heinz J . Mauelshagen wrote: > pvmove moves data chunk by chunk (64kB). > > If such chunk read fails, it retries rereading the failed chunk sector > by sector in order to retrieve as much data as possible. I'm about to try it. But afterwards I like to analyze which files are affected using "debugfs". As far as I understand the code, any unreadable sector is zeroed. The printf reports about the unreadable sector, but only prints the # of the sector from the input device. Instead I'm interested on where the zeroed sector ends up on the destionation device, to locate it later on. Questions: the printed sector number counts 512 byte sectors? If I run pvmove -v, it reports the PE# and LE# of the PE currently copying. I hope I can calculate this way: Get sector# modulo PE size to get the relative position of the broken sector within the PE, and using the destionation LE# + the offset to get the sector # on the LV after the move. Right? Suggestion: write such informations to syslog to be sure this is saved somewhere. If this is printed on screen olny, I may loose this information easily and won't have any change to locate the broken sector later on. Dieter. -- Dieter St�ken, con terra GmbH, M�nster stueken@conterra.de http://www.conterra.de/ (0)251-7474-501