From mboxrd@z Thu Jan 1 00:00:00 1970 From: NeilBrown Subject: Re: failed to write superblock errors, and more! Date: Tue, 15 Feb 2011 15:24:36 +1100 Message-ID: <20110215152436.1351c219@notabene.brown> References: Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-raid-owner@vger.kernel.org To: Jon Nelson Cc: LinuxRaid List-Id: linux-raid.ids On Mon, 14 Feb 2011 22:00:46 -0600 Jon Nelson wrote: > I'm wondering what's going on here: > > turnip:~ # mdadm /dev/md1 --create --level=1 --bitmap=internal > --raid-devices=2 /dev/sdd1 missing > mdadm: array /dev/md1 started. So far, so good. > turnip:~ # mdadm /dev/md2 --create --level=1 --bitmap=internal > --raid-devices=2 /dev/md1 missing > mdadm: failed to write superblock to /dev/md1 Probably your device is an exact multiple of 64K in size, and is of such a size that mdadm choose a bitmap size between 60K and 64K. This doesn't fit and initialising the bitmap fail. Fixed in newer releases. > mdadm: array /dev/md2 started. oops. mdadm ignores the error status after printing the 'failed to write superblock' message. Note added to my to-do list. > turnip:~ # mdadm /dev/md2 --stop Normal usage is: mdadm --stop /dev/md2 Putting them in the other order should either work or give an error message. It currently does neither. Note added to my to-do list. Remaining messages are because /dev/md2 is not stopped. NeilBrown > turnip:~ # mdadm --zero-superblock /dev/md1 > mdadm: Couldn't open /dev/md1 for write - not zeroing > turnip:~ # mdadm /dev/md1 --stop > turnip:~ # mdadm --zero-superblock /dev/sdd1 > mdadm: Couldn't open /dev/sdd1 for write - not zeroing > turnip:~ # mdadm /dev/md2 --stop > turnip:~ # mdadm /dev/md2 --stop --force > turnip:~ # ls /sys/block/md1/holders/ > md2 > turnip:~ # > > The kernel is 2.6.34.7-0.7-default (x86_64). openSUSE 11.3. > mdadm 3.0.3 >