| Message ID | 20170927010348.GK5020@magnolia (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
On Tue, Sep 26, 2017 at 06:03:48PM -0700, Darrick J. Wong wrote: > As of 2007, metadump has an interesting "feature" where it discards > directory extents that are longer than 1000 (originally 20) blocks. > This ostensibly was to protect metadump from corrupt bmbt records, but > it also has the effect of omitting from the metadump valid long extents. > The end result is that we create incomplete metadumps, which is > exacerbated by the lack of warning unless -w is passed. > > So now that we've fixed the default threshold to MAXEXTLEN, check that > the installed metadump no longer exhibits this behavior. > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> > --- > tests/xfs/707 | 101 +++++++++++++++++++++++++++++++++++++++++++++++++++++ > tests/xfs/707.out | 6 +++ > tests/xfs/group | 1 + > 3 files changed, 108 insertions(+) > create mode 100755 tests/xfs/707 > create mode 100644 tests/xfs/707.out > > diff --git a/tests/xfs/707 b/tests/xfs/707 > new file mode 100755 > index 0000000..f97d029 > --- /dev/null > +++ b/tests/xfs/707 > @@ -0,0 +1,101 @@ > +#! /bin/bash > +# FS QA Test No. 707 > +# > +# Ensure that metadump copies large directory extents > +# > +# Metadump helpfully discards directory (and xattr) extents that are > +# longer than 1000 blocks. This is a little silly since a hardlink farm > +# can easily create such a monster. > +# > +# Now that we've upped metadump's default too-long-extent discard > +# threshold to 2^21 blocks, make sure we never do that again. > +# > +#----------------------------------------------------------------------- > +# Copyright (c) 2017, Oracle and/or its affiliates. All Rights Reserved. > +# > +# This program is free software; you can redistribute it and/or > +# modify it under the terms of the GNU General Public License as > +# published by the Free Software Foundation. > +# > +# This program is distributed in the hope that it would be useful, > +# but WITHOUT ANY WARRANTY; without even the implied warranty of > +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the > +# GNU General Public License for more details. > +# > +# You should have received a copy of the GNU General Public License > +# along with this program; if not, write the Free Software Foundation, > +# Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA > +#----------------------------------------------------------------------- > + > +seq=`basename "$0"` > +seqres="$RESULT_DIR/$seq" > +echo "QA output created by $seq" > + > +here=`pwd` > +tmp=/tmp/$$ > +status=1 # failure is the default! > +trap "_cleanup; exit \$status" 0 1 2 3 15 > + > +_cleanup() > +{ > + cd / > + rm -rf "$tmp".* $metadump_file $metadump_img 'rm -f' to be safe :) > +} > + > +# get standard environment, filters and checks > +. ./common/rc > +. ./common/filter > + > +# real QA test starts here > +_supported_os Linux > +_supported_fs xfs > +_require_scratch > + > +rm -f "$seqres.full" > + > +echo "Format and mount" > +_scratch_mkfs -b size=1k -n size=64k > "$seqres.full" 2>&1 Add some comments about the non-default mkfs options? > +_scratch_mount >> "$seqres.full" 2>&1 > + > +metadump_file="$TEST_DIR/meta-$seq" > +metadump_img="$TEST_DIR/img-$seq" > +rm -rf $metadump_file $metadump_img Same here, 'rm -f' > +testdir="$SCRATCH_MNT/test-$seq" > +max_fname_len=255 > +blksz=$(_get_block_size $SCRATCH_MNT) > + > +# Try to create a directory w/ extents > +blocks=1050 > +names=$((blocks * (blksz / max_fname_len))) > +echo "Create huge dir" > +mkdir -p $testdir > +touch $SCRATCH_MNT/a > +seq 0 $names | while read f; do > + name="$testdir/$(printf "%0${max_fname_len}d" $f)" > + ln $SCRATCH_MNT/a $name > +done > +dir_inum=$(stat -c %i $testdir) > + > +echo "Check for > 1000 block extent?" > +_scratch_unmount > +check_for_long_extent() { > + inum=$1 > + > + _scratch_xfs_db -x -c "inode $dir_inum" -c bmap | \ > + sed -e 's/^.*count \([0-9]*\) flag.*$/\1/g' | \ > + awk '{if ($1 > 1000) { printf("yes, %d\n", $1); } }' > +} > +extlen="$(check_for_long_extent $dir_inum)" > +echo "qualifying extent: $extlen blocks" >> $seqres.full > +test -n "$extlen" || _fail "could not create dir extent > 1000 blocks" Is this really a failure? IMHO, _notrun makes more sense here. Thanks, Eryu > + > +echo "Try to metadump" > +_scratch_metadump $metadump_file -w > +xfs_mdrestore $metadump_file $metadump_img > + > +echo "Check restored metadump image" > +$XFS_REPAIR_PROG -n $metadump_img >> $seqres.full 2>&1 > + > +# success, all done > +status=0 > +exit > diff --git a/tests/xfs/707.out b/tests/xfs/707.out > new file mode 100644 > index 0000000..0d2a222 > --- /dev/null > +++ b/tests/xfs/707.out > @@ -0,0 +1,6 @@ > +QA output created by 707 > +Format and mount > +Create huge dir > +Check for > 1000 block extent? > +Try to metadump > +Check restored metadump image > diff --git a/tests/xfs/group b/tests/xfs/group > index 7353b9c..a70c884 100644 > --- a/tests/xfs/group > +++ b/tests/xfs/group > @@ -432,3 +432,4 @@ > 703 auto quick clone fsr > 704 auto quick clone > 705 auto quick clone fsr > +707 auto quick dir metadata -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, Sep 28, 2017 at 04:55:38PM +0800, Eryu Guan wrote: > On Tue, Sep 26, 2017 at 06:03:48PM -0700, Darrick J. Wong wrote: > > As of 2007, metadump has an interesting "feature" where it discards > > directory extents that are longer than 1000 (originally 20) blocks. > > This ostensibly was to protect metadump from corrupt bmbt records, but > > it also has the effect of omitting from the metadump valid long extents. > > The end result is that we create incomplete metadumps, which is > > exacerbated by the lack of warning unless -w is passed. > > > > So now that we've fixed the default threshold to MAXEXTLEN, check that > > the installed metadump no longer exhibits this behavior. > > > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> > > --- > > tests/xfs/707 | 101 +++++++++++++++++++++++++++++++++++++++++++++++++++++ > > tests/xfs/707.out | 6 +++ > > tests/xfs/group | 1 + > > 3 files changed, 108 insertions(+) > > create mode 100755 tests/xfs/707 > > create mode 100644 tests/xfs/707.out > > > > diff --git a/tests/xfs/707 b/tests/xfs/707 > > new file mode 100755 > > index 0000000..f97d029 > > --- /dev/null > > +++ b/tests/xfs/707 > > @@ -0,0 +1,101 @@ > > +#! /bin/bash > > +# FS QA Test No. 707 > > +# > > +# Ensure that metadump copies large directory extents > > +# > > +# Metadump helpfully discards directory (and xattr) extents that are > > +# longer than 1000 blocks. This is a little silly since a hardlink farm > > +# can easily create such a monster. > > +# > > +# Now that we've upped metadump's default too-long-extent discard > > +# threshold to 2^21 blocks, make sure we never do that again. > > +# > > +#----------------------------------------------------------------------- > > +# Copyright (c) 2017, Oracle and/or its affiliates. All Rights Reserved. > > +# > > +# This program is free software; you can redistribute it and/or > > +# modify it under the terms of the GNU General Public License as > > +# published by the Free Software Foundation. > > +# > > +# This program is distributed in the hope that it would be useful, > > +# but WITHOUT ANY WARRANTY; without even the implied warranty of > > +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the > > +# GNU General Public License for more details. > > +# > > +# You should have received a copy of the GNU General Public License > > +# along with this program; if not, write the Free Software Foundation, > > +# Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA > > +#----------------------------------------------------------------------- > > + > > +seq=`basename "$0"` > > +seqres="$RESULT_DIR/$seq" > > +echo "QA output created by $seq" > > + > > +here=`pwd` > > +tmp=/tmp/$$ > > +status=1 # failure is the default! > > +trap "_cleanup; exit \$status" 0 1 2 3 15 > > + > > +_cleanup() > > +{ > > + cd / > > + rm -rf "$tmp".* $metadump_file $metadump_img > > 'rm -f' to be safe :) Ok. > > +} > > + > > +# get standard environment, filters and checks > > +. ./common/rc > > +. ./common/filter > > + > > +# real QA test starts here > > +_supported_os Linux > > +_supported_fs xfs > > +_require_scratch > > + > > +rm -f "$seqres.full" > > + > > +echo "Format and mount" > > +_scratch_mkfs -b size=1k -n size=64k > "$seqres.full" 2>&1 > > Add some comments about the non-default mkfs options? Will do. The goal here is to try to trick the directory code into allocating a huge extent by making directory blocks as large as possible. > > +_scratch_mount >> "$seqres.full" 2>&1 > > + > > +metadump_file="$TEST_DIR/meta-$seq" > > +metadump_img="$TEST_DIR/img-$seq" > > +rm -rf $metadump_file $metadump_img > > Same here, 'rm -f' > > > +testdir="$SCRATCH_MNT/test-$seq" > > +max_fname_len=255 > > +blksz=$(_get_block_size $SCRATCH_MNT) > > + > > +# Try to create a directory w/ extents > > +blocks=1050 > > +names=$((blocks * (blksz / max_fname_len))) > > +echo "Create huge dir" > > +mkdir -p $testdir > > +touch $SCRATCH_MNT/a > > +seq 0 $names | while read f; do > > + name="$testdir/$(printf "%0${max_fname_len}d" $f)" > > + ln $SCRATCH_MNT/a $name > > +done > > +dir_inum=$(stat -c %i $testdir) > > + > > +echo "Check for > 1000 block extent?" > > +_scratch_unmount > > +check_for_long_extent() { > > + inum=$1 > > + > > + _scratch_xfs_db -x -c "inode $dir_inum" -c bmap | \ > > + sed -e 's/^.*count \([0-9]*\) flag.*$/\1/g' | \ > > + awk '{if ($1 > 1000) { printf("yes, %d\n", $1); } }' > > +} > > +extlen="$(check_for_long_extent $dir_inum)" > > +echo "qualifying extent: $extlen blocks" >> $seqres.full > > +test -n "$extlen" || _fail "could not create dir extent > 1000 blocks" > > Is this really a failure? IMHO, _notrun makes more sense here. We failed to set up the preconditions for the test, and therefore did not run it. :) (Will change it to notrun.) --D > > Thanks, > Eryu > > > + > > +echo "Try to metadump" > > +_scratch_metadump $metadump_file -w > > +xfs_mdrestore $metadump_file $metadump_img > > + > > +echo "Check restored metadump image" > > +$XFS_REPAIR_PROG -n $metadump_img >> $seqres.full 2>&1 > > + > > +# success, all done > > +status=0 > > +exit > > diff --git a/tests/xfs/707.out b/tests/xfs/707.out > > new file mode 100644 > > index 0000000..0d2a222 > > --- /dev/null > > +++ b/tests/xfs/707.out > > @@ -0,0 +1,6 @@ > > +QA output created by 707 > > +Format and mount > > +Create huge dir > > +Check for > 1000 block extent? > > +Try to metadump > > +Check restored metadump image > > diff --git a/tests/xfs/group b/tests/xfs/group > > index 7353b9c..a70c884 100644 > > --- a/tests/xfs/group > > +++ b/tests/xfs/group > > @@ -432,3 +432,4 @@ > > 703 auto quick clone fsr > > 704 auto quick clone > > 705 auto quick clone fsr > > +707 auto quick dir metadata > -- > To unsubscribe from this list: send the line "unsubscribe linux-xfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/tests/xfs/707 b/tests/xfs/707 new file mode 100755 index 0000000..f97d029 --- /dev/null +++ b/tests/xfs/707 @@ -0,0 +1,101 @@ +#! /bin/bash +# FS QA Test No. 707 +# +# Ensure that metadump copies large directory extents +# +# Metadump helpfully discards directory (and xattr) extents that are +# longer than 1000 blocks. This is a little silly since a hardlink farm +# can easily create such a monster. +# +# Now that we've upped metadump's default too-long-extent discard +# threshold to 2^21 blocks, make sure we never do that again. +# +#----------------------------------------------------------------------- +# Copyright (c) 2017, Oracle and/or its affiliates. All Rights Reserved. +# +# This program is free software; you can redistribute it and/or +# modify it under the terms of the GNU General Public License as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it would be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, write the Free Software Foundation, +# Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA +#----------------------------------------------------------------------- + +seq=`basename "$0"` +seqres="$RESULT_DIR/$seq" +echo "QA output created by $seq" + +here=`pwd` +tmp=/tmp/$$ +status=1 # failure is the default! +trap "_cleanup; exit \$status" 0 1 2 3 15 + +_cleanup() +{ + cd / + rm -rf "$tmp".* $metadump_file $metadump_img +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# real QA test starts here +_supported_os Linux +_supported_fs xfs +_require_scratch + +rm -f "$seqres.full" + +echo "Format and mount" +_scratch_mkfs -b size=1k -n size=64k > "$seqres.full" 2>&1 +_scratch_mount >> "$seqres.full" 2>&1 + +metadump_file="$TEST_DIR/meta-$seq" +metadump_img="$TEST_DIR/img-$seq" +rm -rf $metadump_file $metadump_img +testdir="$SCRATCH_MNT/test-$seq" +max_fname_len=255 +blksz=$(_get_block_size $SCRATCH_MNT) + +# Try to create a directory w/ extents +blocks=1050 +names=$((blocks * (blksz / max_fname_len))) +echo "Create huge dir" +mkdir -p $testdir +touch $SCRATCH_MNT/a +seq 0 $names | while read f; do + name="$testdir/$(printf "%0${max_fname_len}d" $f)" + ln $SCRATCH_MNT/a $name +done +dir_inum=$(stat -c %i $testdir) + +echo "Check for > 1000 block extent?" +_scratch_unmount +check_for_long_extent() { + inum=$1 + + _scratch_xfs_db -x -c "inode $dir_inum" -c bmap | \ + sed -e 's/^.*count \([0-9]*\) flag.*$/\1/g' | \ + awk '{if ($1 > 1000) { printf("yes, %d\n", $1); } }' +} +extlen="$(check_for_long_extent $dir_inum)" +echo "qualifying extent: $extlen blocks" >> $seqres.full +test -n "$extlen" || _fail "could not create dir extent > 1000 blocks" + +echo "Try to metadump" +_scratch_metadump $metadump_file -w +xfs_mdrestore $metadump_file $metadump_img + +echo "Check restored metadump image" +$XFS_REPAIR_PROG -n $metadump_img >> $seqres.full 2>&1 + +# success, all done +status=0 +exit diff --git a/tests/xfs/707.out b/tests/xfs/707.out new file mode 100644 index 0000000..0d2a222 --- /dev/null +++ b/tests/xfs/707.out @@ -0,0 +1,6 @@ +QA output created by 707 +Format and mount +Create huge dir +Check for > 1000 block extent? +Try to metadump +Check restored metadump image diff --git a/tests/xfs/group b/tests/xfs/group index 7353b9c..a70c884 100644 --- a/tests/xfs/group +++ b/tests/xfs/group @@ -432,3 +432,4 @@ 703 auto quick clone fsr 704 auto quick clone 705 auto quick clone fsr +707 auto quick dir metadata
As of 2007, metadump has an interesting "feature" where it discards directory extents that are longer than 1000 (originally 20) blocks. This ostensibly was to protect metadump from corrupt bmbt records, but it also has the effect of omitting from the metadump valid long extents. The end result is that we create incomplete metadumps, which is exacerbated by the lack of warning unless -w is passed. So now that we've fixed the default threshold to MAXEXTLEN, check that the installed metadump no longer exhibits this behavior. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> --- tests/xfs/707 | 101 +++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/707.out | 6 +++ tests/xfs/group | 1 + 3 files changed, 108 insertions(+) create mode 100755 tests/xfs/707 create mode 100644 tests/xfs/707.out -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html