From mboxrd@z Thu Jan 1 00:00:00 1970 From: Phil Turmel Subject: Re: Offline array, events count mismatch Date: Sun, 8 Nov 2015 22:35:13 -0500 Message-ID: <56401471.2040605@turmel.org> References: <1903758.3ih4aa4jR2@carbon.lepommierdujardin.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1903758.3ih4aa4jR2@carbon.lepommierdujardin.fr> Sender: linux-raid-owner@vger.kernel.org To: Guillaume Paumier , linux-raid@vger.kernel.org List-Id: linux-raid.ids Hi Guillaume, On 11/08/2015 09:49 PM, Guillaume Paumier wrote: [trim /] > Looking at those event counts, my understanding is this: > * Two of the disks (sde, sdf) were dropped from the array for some reason. > * I didn't notice this immediately (an issue I'm addressing separately). > * A third disk (sdj) encountered a small issue today. > * The array went offline because it didn't have enough disks to function > cleanly any more. > > If I understand the documentation [1] correctly, since the event count for sdj > is very close to the event count of sd[b,c,d,g,h,i], I should be able to re- > assemble the array with these 7 disks using --force, leaving sde and sdf > aside. Once the array is assembled, I should be able to re-add sde and sdf, > and they will be re-sync'd. Yes, that is the correct response. Your situation is common. Please see the thread this weekend started by Franscisco Parada. https://marc.info/?t=144691643300001&r=1&w=2&n=12 You should provide "smartctl -i -A -l scterc /dev/sdX" reports for your drives. If you can find an old syslog for when your two worst drives fell out, it might help. Phil