Re: [PATCH] xfstests/shared: dedup integrity test by duperemove

["Darrick J. Wong" <darrick.wong@oracle.com>]

On Mon, May 28, 2018 at 12:54:27PM +0800, Zorro Lang wrote:
> Duperemove is a tool for finding duplicated extents and submitting
> them for deduplication, and it supports XFS. This case trys to
> verify the integrity of XFS after running duperemove.
> 
> Signed-off-by: Zorro Lang <zlang@redhat.com>
> ---
> 
> Hi,
> 
> There's not many softwares support XFS dedup now, duperemove is a rare one.
> So I write this case by using duperemove.
> 
> I use fsstress to make many files and data randomly, I don't know if there're
> better things I can use? Because fsstress only write '0xff' into files, maybe
> I should add an option to make fsstress can write random character?

Heh.  But you probably don't want totally random contents because then
dupremove doesn't do much.

> 
> Please tell me, if you have better ideas:)
> 
> PS: This case test passed on XFS(with reflink=1) and btrfs. And the duperemove
> can reclaim some space in the test, see below:
> 
>   Before duperemove
>     Filesystem                 1K-blocks    Used Available Use% Mounted on
>     /dev/mapper/xxxx-xfscratch 31441920K 583692K 30858228K   2% /mnt/scratch
> 
>   After duperemove
>     Filesystem                 1K-blocks    Used Available Use% Mounted on
>     /dev/mapper/xxxx-xfscratch 31441920K 345728K 31096192K   2% /mnt/scratch
> 
> Thanks,
> Zorro
> 
>  common/config        |  1 +
>  tests/shared/008     | 88 ++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/shared/008.out |  2 ++
>  tests/shared/group   |  1 +
>  4 files changed, 92 insertions(+)
>  create mode 100755 tests/shared/008
>  create mode 100644 tests/shared/008.out
> 
> diff --git a/common/config b/common/config
> index 02c378a9..def559c1 100644
> --- a/common/config
> +++ b/common/config
> @@ -207,6 +207,7 @@ export SQLITE3_PROG="`set_prog_path sqlite3`"
>  export TIMEOUT_PROG="`set_prog_path timeout`"
>  export SETCAP_PROG="`set_prog_path setcap`"
>  export GETCAP_PROG="`set_prog_path getcap`"
> +export DUPEREMOVE_PROG="`set_prog_path duperemove`"
>  
>  # use 'udevadm settle' or 'udevsettle' to wait for lv to be settled.
>  # newer systems have udevadm command but older systems like RHEL5 don't.
> diff --git a/tests/shared/008 b/tests/shared/008
> new file mode 100755
> index 00000000..dace5429
> --- /dev/null
> +++ b/tests/shared/008
> @@ -0,0 +1,88 @@
> +#! /bin/bash
> +# FS QA Test 008
> +#
> +# Dedup integrity test by duperemove
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2018 Red Hat Inc.  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 -f $tmp.*
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/reflink
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +
> +# duperemove only supports btrfs and xfs (with reflink feature).
> +# Add other filesystems if it supports more later.
> +_supported_fs xfs btrfs
> +_supported_os Linux

_require_command "$DUPEREMOVE_PROG" duperemove ?

> +_require_scratch_reflink

_require_scratch_dedupe

> +
> +[ "$DUPEREMOVE_PROG" = "" ] && _notrun "duperemove not found"
> +_scratch_mkfs > $seqres.full 2>&1
> +_scratch_mount >> $seqres.full 2>&1
> +
> +testdir=$SCRATCH_MNT/test-$seq
> +mkdir $testdir
> +
> +fsstress_opts="-w -r -f mknod=0"
> +# Create some files and make a duplicate
> +$FSSTRESS_PROG $fsstress_opts -d $testdir \
> +	       -n $((500 * LOAD_FACTOR)) -p 10 >/dev/null 2>&1
> +duptestdir=${testdir}.dup
> +cp -a $testdir $duptestdir
> +
> +# Make some difference in two directories
> +$FSSTRESS_PROG $fsstress_opts -d $testdir -n 200 -p 5 >/dev/null 2>&1
> +$FSSTRESS_PROG $fsstress_opts -d $duptestdir -n 200 -p 5 >/dev/null 2>&1
> +
> +# Record all files' md5 checksum
> +find $testdir -type f -exec md5sum {} \; > $TEST_DIR/${seq}md5.sum
> +find $duptestdir -type f -exec md5sum {} \; > $TEST_DIR/dup${seq}md5.sum
> +
> +# Dedup
> +echo "== Duperemove output ==" >> $seqres.full
> +$DUPEREMOVE_PROG -dr $SCRATCH_MNT/ >>$seqres.full 2>&1
> +
> +# Verify all files' integrity
> +md5sum -c --quiet $TEST_DIR/${seq}md5.sum
> +md5sum -c --quiet $TEST_DIR/dup${seq}md5.sum

Can we _scratch_mount_cycle and md5sum -c again so that we test that the
pagecache contents don't mutate and a fresh read from the disk also
doesn't show mutations?

--D

> +
> +echo "Silence is golden"
> +
> +status=0
> +exit
> diff --git a/tests/shared/008.out b/tests/shared/008.out
> new file mode 100644
> index 00000000..dd68d5a4
> --- /dev/null
> +++ b/tests/shared/008.out
> @@ -0,0 +1,2 @@
> +QA output created by 008
> +Silence is golden
> diff --git a/tests/shared/group b/tests/shared/group
> index b3663a03..de7fe79f 100644
> --- a/tests/shared/group
> +++ b/tests/shared/group
> @@ -10,6 +10,7 @@
>  005 dangerous_fuzzers
>  006 auto enospc
>  007 dangerous_fuzzers
> +008 auto quick dedupe
>  032 mkfs auto quick
>  272 auto enospc rw
>  289 auto quick
> -- 
> 2.14.3
> 
> --
> 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