RE: Problems with BTRFS formatted disk

["David C. Partridge" <david.partridge@perdrix.co.uk>]

Here's what the 2022.06.01 version of Archlinux had to say in the log when I issued:

mount -t btrfs -o rescue=all /dev/sdc1 /mnt

Jun 19 12:43:01 archiso kernel: BTRFS info (device sdc1): flagging fs with big metadata feature
Jun 19 12:43:01 archiso kernel: BTRFS info (device sdc1): enabling all of the rescue options
Jun 19 12:43:01 archiso kernel: BTRFS info (device sdc1): ignoring data csums
Jun 19 12:43:01 archiso kernel: BTRFS info (device sdc1): ignoring bad roots
Jun 19 12:43:01 archiso kernel: BTRFS info (device sdc1): disabling log replay at mount time
Jun 19 12:43:01 archiso kernel: BTRFS error (device sdc1): nologreplay must be used with ro mount option
Jun 19 12:43:01 archiso kernel: BTRFS error (device sdc1): open_ctree failed

Did I need to say:

mount -t btrfs -o ro,rescue=all /dev/sdc1 /mnt

D.

-----Original Message-----
From: Qu Wenruo <quwenruo.btrfs@gmx.com> 
Sent: 19 June 2022 12:51
To: David C. Partridge <david.partridge@perdrix.co.uk>; linux-btrfs@vger.kernel.org
Subject: Re: Problems with BTRFS formatted disk



On 2022/6/19 19:14, David C. Partridge wrote:
> LUbuntu 22.04 was definitely 5.15 kernel, what alternative distro do you propose I use?

I have no idea why 22.04 doesn't work here.

The upstream commit is 2b29726c473b ("btrfs: rescue: allow ibadroots to
skip bad extent tree when reading block group items"), which is already
in v5.15 kernels.

I double checked the current code base, as long as it's error reading
the block group items and rescue=all (implies ibadroots), it should go
fill_dummy_bgs().

For the alternative distros, OpenSUSE tumbleweed, Archlinux, etc. As
they are definitely upstream and v5.15+.

For example, Archlinux 2022.06.01, it goes with 5.18 kernel:

$ file arch/boot/x86_64/vmlinuz-linux
arch/boot/x86_64/vmlinuz-linux: Linux kernel x86 boot executable
bzImage, version 5.18.1-arch1-1 (linux@archlinux) #1 SMP PREEMPT_DYNAMIC
Mon, 30 May 2022 17:53:11 +0000, RO-rootFS, swap_dev 0XA, Normal VGA

If that still doesn't work, let me creating a similar fs with some block
groups items corrupted to see why it doesn't work.

Thanks,
Qu
>
> -----Original Message-----
> From: Qu Wenruo <quwenruo.btrfs@gmx.com>
> Sent: 19 June 2022 11:41
> To: David C. Partridge <david.partridge@perdrix.co.uk>; linux-btrfs@vger.kernel.org
> Subject: Re: Problems with BTRFS formatted disk
>
>
>
> On 2022/6/19 18:29, David C. Partridge wrote:
>> Booted from live USB 22.04 LUbuntu.
>
> Ubuntu kernel version doesn't seem to be that consistent even for its
> LTS releases:
>
> https://ubuntu.com/about/release-cycle#ubuntu-kernel-release-cycle
>
> Please use something rolling released distro/branch instead.
>
> Thanks,
> Qu
>>
>> root@lubuntu:/home/lubuntu# mount -t btrfs -o rescue=all /dev/sdc1 /mnt
>> mount: /mnt: wrong fs type, bad option, bad superblock on /dev/sdc1, missing codepage or helper program, or other error.
>> root@lubuntu:/home/lubuntu#
>>
>> Content of system journal
>>
>> Jun 19 10:08:03 lubuntu kernel: BTRFS info (device sdc1): flagging fs with big metadata feature
>> Jun 19 10:08:03 lubuntu kernel: BTRFS info (device sdc1): disk space caching is enabled
>> Jun 19 10:08:03 lubuntu kernel: BTRFS info (device sdc1): has skinny extents
>> Jun 19 10:08:03 lubuntu kernel: BTRFS error (device sdc1): parent transid verify failed on 12554992156672 wanted 130582 found 127355
>> Jun 19 10:08:03 lubuntu kernel: BTRFS error (device sdc1): parent transid verify failed on 12554992156672 wanted 130582 found 127355
>> Jun 19 10:08:03 lubuntu kernel: BTRFS error (device sdc1): failed to read block groups: -5
>> Jun 19 10:08:03 lubuntu kernel: BTRFS error (device sdc1): open_ctree failed
>>
>> David
>>
>> -----Original Message-----
>> From: Qu Wenruo <quwenruo.btrfs@gmx.com>
>> Sent: 19 June 2022 03:02
>> To: David C. Partridge <david.partridge@perdrix.co.uk>; linux-btrfs@vger.kernel.org
>> Subject: Re: Problems with BTRFS formatted disk
>>
>>>> You can try rescue=all mount option, which has the extra handling on
>>>> corrupted extent tree.
>>>
>>>> Although you have to use kernels newer than v5.15 (including v5.15) to
>>>> benefit from the change.
>>>
>>> Unfortunately:
>>> amonra@charon:~$ uname -a
>>> Linux charon 5.4.0-113-generic #127-Ubuntu SMP Wed May 18 14:30:56 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux
>>
>> Any special reason that you can not even use a liveUSB to boot a newer
>> kernel to do the salvage?
>>
>>
>> Thanks,
>> Qu
>>
>