From mboxrd@z Thu Jan 1 00:00:00 1970 From: Michael Tokarev Subject: Re: Assembly problems with kernel 2.6.30? Date: Fri, 12 Jun 2009 12:00:03 +0400 Message-ID: <4A320B03.5080009@msgid.tls.msk.ru> References: <20090611072538.GA7841@psychosis.jim.sh> <18993.45142.80174.321434@notabene.brown> <20090612023344.GA24806@psychosis.jim.sh> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20090612023344.GA24806@psychosis.jim.sh> Sender: linux-raid-owner@vger.kernel.org To: Jim Paris Cc: Neil Brown , linux-raid@vger.kernel.org List-Id: linux-raid.ids Jim Paris wrote: [] > Weird: /dev/sdc2 has both a 0.90 and 1.0 superblock. I guess that's > the problem. How would that have happened? mdadm --create over an > existing device without a --zero-superblock inbetween, I suppose? > > I wonder why it was never a problem before. For the root, only > /dev/sda2 and /dev/sdb2 are necessary. Maybe 2.6.30 changed the disk > order, or the ordering of /proc/partitions, such that mdadm now > examines this bad disk before the array is fully assembled, whereas it > never needed to scan it before? 2.6.30 turned on "fastboot" option by default -- it's when all disks/controllers gets scanned in parallel. /mjt