Re: [PATCH 2/5] populate: wipe external xfs log devices when restoring a cached image

From: Zorro Lang <zlang@redhat.com>
To: "Darrick J. Wong" <djwong@kernel.org>
Cc: linux-xfs@vger.kernel.org, fstests@vger.kernel.org
Subject: Re: [PATCH 2/5] populate: wipe external xfs log devices when restoring a cached image
Date: Thu, 13 Oct 2022 23:10:02 +0800	[thread overview]
Message-ID: <20221013151002.ctyhb4wg6ckeokeq@zlang-mailbox> (raw)
In-Reply-To: <166553913349.422450.12686256615707425089.stgit@magnolia>

On Tue, Oct 11, 2022 at 06:45:33PM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
> 
> The fs population code has the ability to save cached metadumps of
> filesystems to save time when running fstests.  The cached images should
> be unmounted cleanly, so we never save the contents of external log
> devices.
> 
> Unfortunately, the cache restore code fails to wipe the external log
> when restoring a clean image, so we end up with strange test failures
> because the log doesn't match the filesystem:
> 
> * ERROR: mismatched uuid in log
> *            SB : 5ffec625-d3bb-4f4e-a181-1f9efe543d9c
> *            log: 607bd75a-a63d-400c-8779-2139f0a3d384
> 
> Worse yet, xfs_repair will overwrite a filesystem's uuid with the log
> uuid, which leads to corruption messages later on:
> 
> Metadata corruption detected at 0x561f69a9a2a8, xfs_agf block 0x8/0x1000
> xfs_db: cannot init perag data (117). Continuing anyway.
> 
> Solve this by wiping the log device when restoring.
> 
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> ---

Make sense,

Reviewed-by: Zorro Lang <zlang@redhat.com>

>  common/populate |   21 +++++++++++++++++++--
>  1 file changed, 19 insertions(+), 2 deletions(-)
> 
> 
> diff --git a/common/populate b/common/populate
> index b501c2fe45..0bd78e0a0a 100644
> --- a/common/populate
> +++ b/common/populate
> @@ -11,7 +11,12 @@ _require_populate_commands() {
>  	_require_xfs_io_command "falloc"
>  	_require_xfs_io_command "fpunch"
>  	_require_test_program "punch-alternating"
> -	_require_command "$XFS_DB_PROG" "xfs_db"
> +	case "${FSTYP}" in
> +	"xfs")
> +		_require_command "$XFS_DB_PROG" "xfs_db"
> +		_require_command "$WIPEFS_PROG" "wipefs"
> +		;;
> +	esac
>  }
>  
>  _require_xfs_db_blocktrash_z_command() {
> @@ -851,7 +856,19 @@ _scratch_populate_restore_cached() {
>  
>  	case "${FSTYP}" in
>  	"xfs")
> -		xfs_mdrestore "${metadump}" "${SCRATCH_DEV}" && return 0
> +		xfs_mdrestore "${metadump}" "${SCRATCH_DEV}"
> +		res=$?
> +		test $res -ne 0 && return $res
> +
> +		# Cached images should have been unmounted cleanly, so if
> +		# there's an external log we need to wipe it and run repair to
> +		# format it to match this filesystem.
> +		if [ -n "${SCRATCH_LOGDEV}" ]; then
> +			$WIPEFS_PROG -a "${SCRATCH_LOGDEV}"
> +			_scratch_xfs_repair
> +			res=$?
> +		fi
> +		return $res
>  		;;
>  	"ext2"|"ext3"|"ext4")
>  		# ext4 cannot e2image external logs, so we cannot restore
> 