From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alan Chandler Subject: Raid1 degraded mode Date: Thu, 06 Jan 2011 16:41:53 +0000 Message-ID: <4D25F0D1.1030009@hartley-consultants.com> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="------------020300090202060003070208" To: linux-btrfs@vger.kernel.org Return-path: List-ID: This is a multi-part message in MIME format. --------------020300090202060003070208 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit I am trying to understand how btrfs works with Raid1. Is it possible to create the filesystem with -m raid1 -d raid1 in which there is only one device available when the filesystem is created. Is it possible to refer to a second device as "missing" The use case I am thinking of is converting an existing raid1 setup from mdm + lvm + ext4 to btrfs with raid1 and subvolumes. I would like to fail half of my existing mdm devices and re-initialise them as btrfs. make subvolumes for each of my lvm logical volumes, copy the data over, then remove mdm and lvm entirely and add back in the devices that were supporting them as additional btrfs devices. The problem I see is that, according to the wiki, specifying -m raid1 and -d raid1 with a single device causes two copies to be made on the same device, whereas I really want it to start off degraded with only one copy until I add in the second device. Trying to cram two copies on the one device will make it run out of space. --------------020300090202060003070208 Content-Type: text/x-vcard; charset=utf-8; name="alan_chandler.vcf" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="alan_chandler.vcf" begin:vcard fn:Alan Chandler n:Chandler;Alan email;internet:alan.chandler@hartley-consultants.com tel;cell:+44 7540 419 854 version:2.1 end:vcard --------------020300090202060003070208--