From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from c.mail.sonic.net ([64.142.111.80]:56861 "EHLO c.mail.sonic.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751478AbcAWXhh (ORCPT ); Sat, 23 Jan 2016 18:37:37 -0500 Subject: Re: "Some devices missing" only while not mounted To: Chris Murphy References: <56A13176.2050502@gmail.com> <56A15B6F.9070300@gmail.com> <56A19785.6090702@gmail.com> Cc: Btrfs BTRFS From: Konstantin Svist Message-ID: <56A40EBD.90900@gmail.com> Date: Sat, 23 Jan 2016 15:37:33 -0800 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8 Sender: linux-btrfs-owner@vger.kernel.org List-ID: On 01/21/2016 07:08 PM, Chris Murphy wrote: > On Thu, Jan 21, 2016 at 7:44 PM, Konstantin Svist wrote: > >> Wish the FS could be "edited" to remove the phantom device to make it work.. > Well that's why I'm saying it's a bug. If there's a missing device, > 'btrfs dev del missing' should work. But there's no device missing for > 'dev del' it's only missing for 'dev ready' which makes no sense. > > I suggest filing a bug, bugzilla.kernel.org. Something like 'bogus > 'some devices missing' prevents boot. And then give examples in the > summary: > > - mounts OK without degraded > - btrfs check does not report missing devices > - btrfs fi show does not report missing devices > - btrfs fi show -d does report missing devices > - btrfs dev del missing reports there are no missing devices > - btrfs dev ready exit value 1, so boot always fails > > I don't know what debug info is useful for this case either. Maybe > btrfs-image, btrfs-debug-tree output to a file maybe the chunk tree > has a clue about this phantom missing device. > > And then once you file the bug, and post the URL here, go on the IRC > channel tomorrow (business hours U.S. time) and reference it and ask > what other information devs would want before you blow away the volume > and start over. Because presumably you need to get back to work, and > to do that at this point, baring a dev having an idea how to fix it, > you'll need to blow it away. > > Make a backup while you can in the meantime. > I've made a dd copy of the problematic array, changed the UUID in the copy and whittled it down to 0 files (removed all subvolumes) Made an image of the 0-file array by "btrfs-image -c9", now posted to the bug (https://bugzilla.kernel.org/show_bug.cgi?id=111161) Need to know if that's enough to go on or more info is needed.. Also would like to know if it's worth waiting a day or two for someone to give me a way to fix the original "broken" array -- or better to just copy all the files and start over, as mentioned above?