From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andreas Klauer Subject: Re: RAID5 grow interrupted. Date: Thu, 22 Sep 2016 19:19:15 +0200 Message-ID: <20160922171915.GA5392@metamorpher.de> References: <36947ef9-64e2-f084-4949-476107e2771e@spallek.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <36947ef9-64e2-f084-4949-476107e2771e@spallek.org> Sender: linux-raid-owner@vger.kernel.org To: Axel Spallek Cc: linux-raid@vger.kernel.org List-Id: linux-raid.ids On Thu, Sep 22, 2016 at 04:11:28PM +0200, Axel Spallek wrote: > I tried to get a disk out of a RAID5 with 8 hdds (4TB) to get a Hotspare > to change to RAID6 afterwards. I'm a bit confused here. You started out with 8 disk RAID-5? And final result is to be what? RAID-6 with 8 disks? > Therefore I issued the following commands: > > mdadm --grow -n7 /dev/md1 <-- just to get the size for the next command. > > mdadm --grow /dev/md1 --array-size 23441292288 > > mdadm --grow -n7 /dev/md1 --backup-file /var/backups/mdadm.backup I'm not sure, but you might not have needed any of these intermediate steps. But now you started it you might have to finish them. That said there is a reshape revert (which is not yet in the manpage?) which might help if the regular reshape just won't resume. > root@s10:~# mdadm -A --scan --verbose > mdadm: looking for devices for further assembly > mdadm: /dev/sdg1 is busy - skipping It says busy because already assembled (or in use by something else). So you have to stop first before assembeling again. I didn't see anything out of place in your --examine. It seems to be in sync, event count identical, .... Good luck Andreas Klauer