diff mbox series

[v2] common/xfs: wipe the XFS superblock of each AGs

Message ID 20190920062327.14747-1-zlang@redhat.com (mailing list archive)
State New, archived
Headers show
Series [v2] common/xfs: wipe the XFS superblock of each AGs | expand

Commit Message

Zorro Lang Sept. 20, 2019, 6:23 a.m. UTC
xfs/030 always fails after d0e484ac699f ("check: wipe scratch devices
between tests") get merged.

Due to xfs/030 does a sized(100m) mkfs. Before we merge above commit,
mkfs.xfs detects an old primary superblock, it will write zeroes to
all superblocks before formatting the new filesystem. But this won't
be done if we wipe the first superblock(by merging above commit).

That means if we make a (smaller) sized xfs after wipefs, those *old*
superblocks which created by last time mkfs.xfs will be left on disk.
Then when we do xfs_repair, if xfs_repair can't find the first SB, it
will go to find those *old* SB at first. When it finds them,
everyting goes wrong.

So I try to get XFS AG geometry(by default) and then try to erase all
superblocks. Thanks Darrick J. Wong helped to analyze this issue.

Reported-by: Yang Xu <xuyang2018.jy@cn.fujitsu.com>
Signed-off-by: Zorro Lang <zlang@redhat.com>
---

Hi,

V2 did below changes:
1) Use xfs_db to detect the real xfs geometry
2) Do a $FSTYP specified wipe before trying to wipefs all scratch devices

Thanks,
Zorro

 common/rc  |  8 ++++++++
 common/xfs | 20 ++++++++++++++++++++
 2 files changed, 28 insertions(+)

Comments

Darrick J. Wong Sept. 20, 2019, 3:34 p.m. UTC | #1
On Fri, Sep 20, 2019 at 02:23:27PM +0800, Zorro Lang wrote:
> xfs/030 always fails after d0e484ac699f ("check: wipe scratch devices
> between tests") get merged.
> 
> Due to xfs/030 does a sized(100m) mkfs. Before we merge above commit,
> mkfs.xfs detects an old primary superblock, it will write zeroes to
> all superblocks before formatting the new filesystem. But this won't
> be done if we wipe the first superblock(by merging above commit).
> 
> That means if we make a (smaller) sized xfs after wipefs, those *old*
> superblocks which created by last time mkfs.xfs will be left on disk.
> Then when we do xfs_repair, if xfs_repair can't find the first SB, it
> will go to find those *old* SB at first. When it finds them,
> everyting goes wrong.
> 
> So I try to get XFS AG geometry(by default) and then try to erase all
> superblocks. Thanks Darrick J. Wong helped to analyze this issue.
> 
> Reported-by: Yang Xu <xuyang2018.jy@cn.fujitsu.com>
> Signed-off-by: Zorro Lang <zlang@redhat.com>
> ---
> 
> Hi,
> 
> V2 did below changes:
> 1) Use xfs_db to detect the real xfs geometry
> 2) Do a $FSTYP specified wipe before trying to wipefs all scratch devices
> 
> Thanks,
> Zorro
> 
>  common/rc  |  8 ++++++++
>  common/xfs | 20 ++++++++++++++++++++
>  2 files changed, 28 insertions(+)
> 
> diff --git a/common/rc b/common/rc
> index 66c7fd4d..56329747 100644
> --- a/common/rc
> +++ b/common/rc
> @@ -4045,6 +4045,14 @@ _try_wipe_scratch_devs()
>  {
>  	test -x "$WIPEFS_PROG" || return 0
>  
> +	# Do specified filesystem wipe at first
> +	case "$FSTYP" in
> +	"xfs")
> +		_try_wipe_scratch_xfs
> +		;;
> +	esac
> +
> +	# Then do wipefs on all scratch devices
>  	for dev in $SCRATCH_DEV_POOL $SCRATCH_DEV $SCRATCH_LOGDEV $SCRATCH_RTDEV; do
>  		test -b $dev && $WIPEFS_PROG -a $dev
>  	done
> diff --git a/common/xfs b/common/xfs
> index 1bce3c18..082a1744 100644
> --- a/common/xfs
> +++ b/common/xfs
> @@ -884,3 +884,23 @@ _xfs_mount_agcount()
>  {
>  	$XFS_INFO_PROG "$1" | grep agcount= | sed -e 's/^.*agcount=\([0-9]*\),.*$/\1/g'
>  }
> +
> +# Wipe the superblock of each XFS AGs
> +_try_wipe_scratch_xfs()
> +{
> +	local num='^[0-9]+$'
> +	local agcount
> +	local agsize
> +	local dbsize
> +
> +	agcount=`_scratch_xfs_get_sb_field agcount 2>/dev/null`
> +	agsize=`_scratch_xfs_get_sb_field agblocks 2>/dev/null`
> +	dbsize=`_scratch_xfs_get_sb_field blocksize 2>/dev/null`
> +
> +	if [[ $agcount =~ $num && $agsize =~ $num && $dbsize =~ $num ]];then
> +		for ((i = 0; i < agcount; i++)); do
> +			$XFS_IO_PROG -c "pwrite $((i * dbsize * agsize)) $dbsize" \
> +				$SCRATCH_DEV >/dev/null;
> +		done
> +	fi

What happened to the loop that simulates a _scratch_mkfs_xfs run (to get
the AG geometry) and then zaps that too?  You need both zeroing loops to
make sure xfs/030 doesn't trip over old superblocks, right?

--D

> +}
> -- 
> 2.20.1
>
Zorro Lang Sept. 22, 2019, 11:25 a.m. UTC | #2
On Fri, Sep 20, 2019 at 08:34:43AM -0700, Darrick J. Wong wrote:
> On Fri, Sep 20, 2019 at 02:23:27PM +0800, Zorro Lang wrote:
> > xfs/030 always fails after d0e484ac699f ("check: wipe scratch devices
> > between tests") get merged.
> > 
> > Due to xfs/030 does a sized(100m) mkfs. Before we merge above commit,
> > mkfs.xfs detects an old primary superblock, it will write zeroes to
> > all superblocks before formatting the new filesystem. But this won't
> > be done if we wipe the first superblock(by merging above commit).
> > 
> > That means if we make a (smaller) sized xfs after wipefs, those *old*
> > superblocks which created by last time mkfs.xfs will be left on disk.
> > Then when we do xfs_repair, if xfs_repair can't find the first SB, it
> > will go to find those *old* SB at first. When it finds them,
> > everyting goes wrong.
> > 
> > So I try to get XFS AG geometry(by default) and then try to erase all
> > superblocks. Thanks Darrick J. Wong helped to analyze this issue.
> > 
> > Reported-by: Yang Xu <xuyang2018.jy@cn.fujitsu.com>
> > Signed-off-by: Zorro Lang <zlang@redhat.com>
> > ---
> > 
> > Hi,
> > 
> > V2 did below changes:
> > 1) Use xfs_db to detect the real xfs geometry
> > 2) Do a $FSTYP specified wipe before trying to wipefs all scratch devices
> > 
> > Thanks,
> > Zorro
> > 
> >  common/rc  |  8 ++++++++
> >  common/xfs | 20 ++++++++++++++++++++
> >  2 files changed, 28 insertions(+)
> > 
> > diff --git a/common/rc b/common/rc
> > index 66c7fd4d..56329747 100644
> > --- a/common/rc
> > +++ b/common/rc
> > @@ -4045,6 +4045,14 @@ _try_wipe_scratch_devs()
> >  {
> >  	test -x "$WIPEFS_PROG" || return 0
> >  
> > +	# Do specified filesystem wipe at first
> > +	case "$FSTYP" in
> > +	"xfs")
> > +		_try_wipe_scratch_xfs
> > +		;;
> > +	esac
> > +
> > +	# Then do wipefs on all scratch devices
> >  	for dev in $SCRATCH_DEV_POOL $SCRATCH_DEV $SCRATCH_LOGDEV $SCRATCH_RTDEV; do
> >  		test -b $dev && $WIPEFS_PROG -a $dev
> >  	done
> > diff --git a/common/xfs b/common/xfs
> > index 1bce3c18..082a1744 100644
> > --- a/common/xfs
> > +++ b/common/xfs
> > @@ -884,3 +884,23 @@ _xfs_mount_agcount()
> >  {
> >  	$XFS_INFO_PROG "$1" | grep agcount= | sed -e 's/^.*agcount=\([0-9]*\),.*$/\1/g'
> >  }
> > +
> > +# Wipe the superblock of each XFS AGs
> > +_try_wipe_scratch_xfs()
> > +{
> > +	local num='^[0-9]+$'
> > +	local agcount
> > +	local agsize
> > +	local dbsize
> > +
> > +	agcount=`_scratch_xfs_get_sb_field agcount 2>/dev/null`
> > +	agsize=`_scratch_xfs_get_sb_field agblocks 2>/dev/null`
> > +	dbsize=`_scratch_xfs_get_sb_field blocksize 2>/dev/null`
> > +
> > +	if [[ $agcount =~ $num && $agsize =~ $num && $dbsize =~ $num ]];then
> > +		for ((i = 0; i < agcount; i++)); do
> > +			$XFS_IO_PROG -c "pwrite $((i * dbsize * agsize)) $dbsize" \
> > +				$SCRATCH_DEV >/dev/null;
> > +		done
> > +	fi
> 
> What happened to the loop that simulates a _scratch_mkfs_xfs run (to get
> the AG geometry) and then zaps that too?  You need both zeroing loops to
> make sure xfs/030 doesn't trip over old superblocks, right?

Hi,

Sorry I'm not sure what you mean. Do you mean I'd better to do two zeroing
loops, one is as above, the other is for default mkfs.xfs geometry as
below[1]?

Thanks,
Zorro

[1]
  local tmp=`mktemp -u`
  _scratch_mkfs_xfs -N 2>/dev/null | perl -ne '
    if (/^meta-data=.*\s+agcount=(\d+), agsize=(\d+) blks/) {
      print STDOUT "agcount=$1\nagsize=$2\n";
    }
    if (/^data\s+=\s+bsize=(\d+)\s/) {
      print STDOUT "dbsize=$1\n";
    }' > $tmp.mkfs

  . $tmp.mkfs
  if [ -n "$agcount" -a -n "$agsize" -a -n "$dbsize" ];then
    for ((i = 0; i < agcount; i++)); do
      $XFS_IO_PROG -c "pwrite $((i * dbsize * agsize)) $dbsize" \
         $SCRATCH_DEV >/dev/null;
    done
  fi
  rm -f $tmp.mkfs



> 
> --D
> 
> > +}
> > -- 
> > 2.20.1
> >
Darrick J. Wong Sept. 23, 2019, 3:12 p.m. UTC | #3
On Sun, Sep 22, 2019 at 07:25:12PM +0800, Zorro Lang wrote:
> On Fri, Sep 20, 2019 at 08:34:43AM -0700, Darrick J. Wong wrote:
> > On Fri, Sep 20, 2019 at 02:23:27PM +0800, Zorro Lang wrote:
> > > xfs/030 always fails after d0e484ac699f ("check: wipe scratch devices
> > > between tests") get merged.
> > > 
> > > Due to xfs/030 does a sized(100m) mkfs. Before we merge above commit,
> > > mkfs.xfs detects an old primary superblock, it will write zeroes to
> > > all superblocks before formatting the new filesystem. But this won't
> > > be done if we wipe the first superblock(by merging above commit).
> > > 
> > > That means if we make a (smaller) sized xfs after wipefs, those *old*
> > > superblocks which created by last time mkfs.xfs will be left on disk.
> > > Then when we do xfs_repair, if xfs_repair can't find the first SB, it
> > > will go to find those *old* SB at first. When it finds them,
> > > everyting goes wrong.
> > > 
> > > So I try to get XFS AG geometry(by default) and then try to erase all
> > > superblocks. Thanks Darrick J. Wong helped to analyze this issue.
> > > 
> > > Reported-by: Yang Xu <xuyang2018.jy@cn.fujitsu.com>
> > > Signed-off-by: Zorro Lang <zlang@redhat.com>
> > > ---
> > > 
> > > Hi,
> > > 
> > > V2 did below changes:
> > > 1) Use xfs_db to detect the real xfs geometry
> > > 2) Do a $FSTYP specified wipe before trying to wipefs all scratch devices
> > > 
> > > Thanks,
> > > Zorro
> > > 
> > >  common/rc  |  8 ++++++++
> > >  common/xfs | 20 ++++++++++++++++++++
> > >  2 files changed, 28 insertions(+)
> > > 
> > > diff --git a/common/rc b/common/rc
> > > index 66c7fd4d..56329747 100644
> > > --- a/common/rc
> > > +++ b/common/rc
> > > @@ -4045,6 +4045,14 @@ _try_wipe_scratch_devs()
> > >  {
> > >  	test -x "$WIPEFS_PROG" || return 0
> > >  
> > > +	# Do specified filesystem wipe at first
> > > +	case "$FSTYP" in
> > > +	"xfs")
> > > +		_try_wipe_scratch_xfs
> > > +		;;
> > > +	esac
> > > +
> > > +	# Then do wipefs on all scratch devices
> > >  	for dev in $SCRATCH_DEV_POOL $SCRATCH_DEV $SCRATCH_LOGDEV $SCRATCH_RTDEV; do
> > >  		test -b $dev && $WIPEFS_PROG -a $dev
> > >  	done
> > > diff --git a/common/xfs b/common/xfs
> > > index 1bce3c18..082a1744 100644
> > > --- a/common/xfs
> > > +++ b/common/xfs
> > > @@ -884,3 +884,23 @@ _xfs_mount_agcount()
> > >  {
> > >  	$XFS_INFO_PROG "$1" | grep agcount= | sed -e 's/^.*agcount=\([0-9]*\),.*$/\1/g'
> > >  }
> > > +
> > > +# Wipe the superblock of each XFS AGs
> > > +_try_wipe_scratch_xfs()
> > > +{
> > > +	local num='^[0-9]+$'
> > > +	local agcount
> > > +	local agsize
> > > +	local dbsize
> > > +
> > > +	agcount=`_scratch_xfs_get_sb_field agcount 2>/dev/null`
> > > +	agsize=`_scratch_xfs_get_sb_field agblocks 2>/dev/null`
> > > +	dbsize=`_scratch_xfs_get_sb_field blocksize 2>/dev/null`
> > > +
> > > +	if [[ $agcount =~ $num && $agsize =~ $num && $dbsize =~ $num ]];then
> > > +		for ((i = 0; i < agcount; i++)); do
> > > +			$XFS_IO_PROG -c "pwrite $((i * dbsize * agsize)) $dbsize" \
> > > +				$SCRATCH_DEV >/dev/null;
> > > +		done
> > > +	fi
> > 
> > What happened to the loop that simulates a _scratch_mkfs_xfs run (to get
> > the AG geometry) and then zaps that too?  You need both zeroing loops to
> > make sure xfs/030 doesn't trip over old superblocks, right?
> 
> Hi,
> 
> Sorry I'm not sure what you mean. Do you mean I'd better to do two zeroing
> loops, one is as above, the other is for default mkfs.xfs geometry as
> below[1]?

Yes.

--D

> Thanks,
> Zorro
> 
> [1]
>   local tmp=`mktemp -u`
>   _scratch_mkfs_xfs -N 2>/dev/null | perl -ne '
>     if (/^meta-data=.*\s+agcount=(\d+), agsize=(\d+) blks/) {
>       print STDOUT "agcount=$1\nagsize=$2\n";
>     }
>     if (/^data\s+=\s+bsize=(\d+)\s/) {
>       print STDOUT "dbsize=$1\n";
>     }' > $tmp.mkfs
> 
>   . $tmp.mkfs
>   if [ -n "$agcount" -a -n "$agsize" -a -n "$dbsize" ];then
>     for ((i = 0; i < agcount; i++)); do
>       $XFS_IO_PROG -c "pwrite $((i * dbsize * agsize)) $dbsize" \
>          $SCRATCH_DEV >/dev/null;
>     done
>   fi
>   rm -f $tmp.mkfs
> 
> 
> 
> > 
> > --D
> > 
> > > +}
> > > -- 
> > > 2.20.1
> > >
diff mbox series

Patch

diff --git a/common/rc b/common/rc
index 66c7fd4d..56329747 100644
--- a/common/rc
+++ b/common/rc
@@ -4045,6 +4045,14 @@  _try_wipe_scratch_devs()
 {
 	test -x "$WIPEFS_PROG" || return 0
 
+	# Do specified filesystem wipe at first
+	case "$FSTYP" in
+	"xfs")
+		_try_wipe_scratch_xfs
+		;;
+	esac
+
+	# Then do wipefs on all scratch devices
 	for dev in $SCRATCH_DEV_POOL $SCRATCH_DEV $SCRATCH_LOGDEV $SCRATCH_RTDEV; do
 		test -b $dev && $WIPEFS_PROG -a $dev
 	done
diff --git a/common/xfs b/common/xfs
index 1bce3c18..082a1744 100644
--- a/common/xfs
+++ b/common/xfs
@@ -884,3 +884,23 @@  _xfs_mount_agcount()
 {
 	$XFS_INFO_PROG "$1" | grep agcount= | sed -e 's/^.*agcount=\([0-9]*\),.*$/\1/g'
 }
+
+# Wipe the superblock of each XFS AGs
+_try_wipe_scratch_xfs()
+{
+	local num='^[0-9]+$'
+	local agcount
+	local agsize
+	local dbsize
+
+	agcount=`_scratch_xfs_get_sb_field agcount 2>/dev/null`
+	agsize=`_scratch_xfs_get_sb_field agblocks 2>/dev/null`
+	dbsize=`_scratch_xfs_get_sb_field blocksize 2>/dev/null`
+
+	if [[ $agcount =~ $num && $agsize =~ $num && $dbsize =~ $num ]];then
+		for ((i = 0; i < agcount; i++)); do
+			$XFS_IO_PROG -c "pwrite $((i * dbsize * agsize)) $dbsize" \
+				$SCRATCH_DEV >/dev/null;
+		done
+	fi
+}