From mboxrd@z Thu Jan 1 00:00:00 1970 From: Phil Turmel Subject: Re: problem with raid after reboot Date: Mon, 10 Dec 2012 10:04:53 -0500 Message-ID: <50C5FA15.9000600@turmel.org> References: <50C5EA89.40804@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: bobzer Cc: linux-raid List-Id: linux-raid.ids Hi Mathieu, [Top-posting repaired. Please don't.] [List address restored. Please use reply-to-all on kernel.org lists.] On 12/10/2012 09:14 AM, bobzer wrote: >> Hi Mathieu, [trim /] >> It looks like the superblocks didn't get updated together due to your >> improper reboot. (Event counts don't match.) >> >> I suggest you stop the array and reassemble with "--force", like so: >> >> mdadm --stop /dev/md0 >> mdadm --assemble --force /dev/md0 /dev/sd[bcd]1 >> >> The "--force" option will instruct mdadm to ignore the event count >> mismatch. >> >> You probably want to investigate why your array wasn't properly stopped >> during shutdown. > Hi Phil, > thanks to help me. > > what do you think of the array state ?it' s look strange no ? > > Device Role : Active device 0 > Array State : A.. ('A' == active, '.' == missing) > > Device Role : Active device 1 > Array State : AAA ('A' == active, '.' == missing) > > Device Role : Active device 2 > Array State : AAA ('A' == active, '.' == missing) > > do you think it's a problem ? It's a side-effect of the improper shutdown. Two devices' superblocks weren't saved properly, and still show everything OK. But their event counts are older than the one device that shows a problem. So mdadm sees that one device as most up-to-date, and cannot start the raid5 from just that one. Please try the commands I recommended. Phil