From mboxrd@z Thu Jan 1 00:00:00 1970 From: Phil Turmel Subject: Re: reaid problem at reboot Date: Tue, 21 Jun 2011 23:53:37 -0400 Message-ID: <4E016741.8080606@turmel.org> References: ,<4E00FC8A.1050908@turmel.org> ,<4E0113FE.503@turmel.org> ,<4E0146B6.40101@turmel.org> ,<4E0153C1.1040708@turmel.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-raid-owner@vger.kernel.org To: william L'Heureux Cc: linux-raid@vger.kernel.org List-Id: linux-raid.ids On 06/21/2011 11:34 PM, william L'Heureux wrote: > > /dev/sdj1 cant be first, its /dev/sda1. 'j' had the lvm metadata in the right spot. It was first. If it's not now, then your partial rebuild scrambled it. I don't think I have any more expertise to offer at this point. > My friend made a python script. take a look. > > cat rebuild.py [...] Nice to automate the search, especially with many combinations. > i can see the data with the combinatin below, but there is a lot of errors, sounds like a rebuild was started(beghining of the disk) then was canceled That hurts. > we made a snapshot, run a e2fsck on it and im still(not now, gotta sleep) add 1 missing or 2. after the e2fsck i mount the snap, look at the data, check corruption, rince and repeat with another combination of missing and not. i rate the ones with less corruption > > the only e2fsck working is with --->>> a,j,c,h,i,d Good luck. Phil