From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pf0-f170.google.com ([209.85.192.170]:44195 "EHLO mail-pf0-f170.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751202AbeEBTnM (ORCPT ); Wed, 2 May 2018 15:43:12 -0400 Received: by mail-pf0-f170.google.com with SMTP id q22so12650427pff.11 for ; Wed, 02 May 2018 12:43:12 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <20edd001-152b-edbb-548b-dcb2647d828a@gmx.com> References:

<20edd001-152b-edbb-548b-dcb2647d828a@gmx.com> From: Paul Richards Date: Wed, 2 May 2018 20:43:11 +0100 Message-ID: Subject: Re: fs/btrfs/ctree.h:1559 warnings To: Qu Wenruo Cc: Patrik Lundquist , linux-btrfs@vger.kernel.org Content-Type: text/plain; charset="UTF-8" Sender: linux-btrfs-owner@vger.kernel.org List-ID: On 29 April 2018 at 02:50, Qu Wenruo wrote: > > > On 2018年04月29日 04:16, Paul Richards wrote: >> On 28 April 2018 at 20:39, Patrik Lundquist wrote: >>> On 28 April 2018 at 20:54, Paul Richards wrote: >>>> >>>> Hi, >>>> I recently upgraded from Linux 4.4.0 to 4.13.0 (Ubuntu 16.04 stock to >>>> hwe kernel). >>>> >>>> Since then, I've noticed lots of btrfs warnings in dmesg (example at >>>> the end). I believe these warnings to be benign, and they relate to >>>> my partition not being a multiple of 4KiB in size (I confirmed that >>>> alignment is okay in this instance). >>> >>> >>> Run btrfs rescue fix-device-size . >>> >> >> >> Excellent, that looks like exactly what I need. It's a shame my >> google-fu didn't uncover it before I posted. >> >> However, the "fix-device-size" sub-command is not available in >> btrfs-tools (v4.4) from Ubuntu 16.04. I guess my original question >> still stands. Is this warning safe for me to ignore? > > Safe to ignore. > > And there is another way to solve it. > > Shrink your fs by 4K, and newer kernel will do the round down for you > and result new device size to be aligned. > > Thanks, > Qu > Thanks - this fixed the issue for me (modulo my problem below..). I thought I should follow up with exactly what I did for anyone else looking to fix this. I used this command to identify which devices were not exact multiples of 4KiB in size: % btrfs fi show --raw /storage Label: none uuid: c4b33374-9006-47d7-b6f9-2136dc988f9a Total devices 3 FS bytes used 7027891666944 devid 1 size 8001560055808 used 4691211583488 path /dev/mapper/storage1 devid 3 size 8001560059392 used 4691211583488 path /dev/mapper/storage3 devid 4 size 8001560059392 used 4690104287232 path /dev/mapper/storage4 I used a calculator to determine that devids 3 and 4 were the offenders (they don't divide by 4006 nicely). I then ran: % btrfs fi resize 3:-1K /storage % btrfs fi resize 4:-1K /storage Which has resulted in this: % btrfs fi show --raw /storage Label: none uuid: c4b33374-9006-47d7-b6f9-2136dc988f9a Total devices 3 FS bytes used 7027891666944 devid 1 size 8001560055808 used 4691211583488 path /dev/mapper/storage1 devid 3 size 8001560055808 used 4691211583488 path /dev/mapper/storage3 devid 4 size 8001560055808 used 4690104287232 path /dev/mapper/storage4 PERFECT! :) The issue I have now is that the filesystem cannot be unmounted. "umount" reports "target is busy", but I cannot find anything relevant with "lsof" or "fuser" (this is a very quiet home NAS). Is this something related to the resize? For what it's worth, here are the sizes of the underlying devices (output somewhat redacted): % lsblk -b NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT storage3 253:2 0 8001560059392 0 crypt storage4 253:1 0 8001560059392 0 crypt storage1 253:0 0 8001560059392 0 crypt /storage/.snapshots It looks like all the devices are an odd size. One of these disks was added after the others, presumably with a newer kernel that rounded the btrfs dev size downwards.