[PATCH v3 1/3] xfstests: dedupe a single big file and verify integrity

[PATCH v3 1/3] xfstests: dedupe a single big file and verify integrity

[Zorro Lang]

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>
---

V3 update description about SPDX-License.

Thanks,
Zorro

 common/config        |  1 +
 tests/shared/008     | 64 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/shared/008.out |  3 +++
 tests/shared/group   |  1 +
 4 files changed, 69 insertions(+)
 create mode 100755 tests/shared/008
 create mode 100644 tests/shared/008.out

diff --git a/common/config b/common/config
index 09e7ffee..d02d6ed5 100644
--- a/common/config
+++ b/common/config
@@ -192,6 +192,7 @@ export SETCAP_PROG="$(type -P setcap)"
 export GETCAP_PROG="$(type -P getcap)"
 export CHECKBASHISMS_PROG="$(type -P checkbashisms)"
 export XFS_INFO_PROG="$(type -P xfs_info)"
+export DUPEREMOVE_PROG="$(type -P 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..a7d76083
--- /dev/null
+++ b/tests/shared/008
@@ -0,0 +1,64 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2018 Red Hat Inc.  All Rights Reserved.
+#
+# FS QA Test 008
+#
+# Dedupe a single big file and verify integrity
+#
+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 $seqres.full
+
+# duperemove only supports btrfs and xfs (with reflink feature).
+# Add other filesystems if it supports more later.
+_supported_fs xfs btrfs
+_supported_os Linux
+_require_scratch_dedupe
+_require_command "$DUPEREMOVE_PROG" duperemove
+
+fssize=$((2 * 1024 * 1024 * 1024))
+_scratch_mkfs_sized $fssize > $seqres.full 2>&1
+_scratch_mount >> $seqres.full 2>&1
+
+# fill the fs with a big file has same contents
+$XFS_IO_PROG -f -c "pwrite -S 0x55 0 $fssize" $SCRATCH_MNT/${seq}.file \
+	>> $seqres.full 2>&1
+md5sum $SCRATCH_MNT/${seq}.file > ${tmp}.md5sum
+
+echo "= before cycle mount ="
+# Dedupe with 1M blocksize
+$DUPEREMOVE_PROG -dr --dedupe-options=same -b 1048576 $SCRATCH_MNT/ >>$seqres.full 2>&1
+# Verify integrity
+md5sum -c --quiet ${tmp}.md5sum
+# Dedupe with 64k blocksize
+$DUPEREMOVE_PROG -dr --dedupe-options=same -b 65536 $SCRATCH_MNT/ >>$seqres.full 2>&1
+# Verify integrity again
+md5sum -c --quiet ${tmp}.md5sum
+
+# umount and mount again, verify pagecache contents don't mutate
+_scratch_cycle_mount
+echo "= after cycle mount ="
+md5sum -c --quiet ${tmp}.md5sum
+
+status=0
+exit
diff --git a/tests/shared/008.out b/tests/shared/008.out
new file mode 100644
index 00000000..f29d478f
--- /dev/null
+++ b/tests/shared/008.out
@@ -0,0 +1,3 @@
+QA output created by 008
+= before cycle mount =
+= after cycle mount =
diff --git a/tests/shared/group b/tests/shared/group
index b3663a03..49ffa8dd 100644
--- a/tests/shared/group
+++ b/tests/shared/group
@@ -10,6 +10,7 @@
 005 dangerous_fuzzers
 006 auto enospc
 007 dangerous_fuzzers
+008 auto stress dedupe
 032 mkfs auto quick
 272 auto enospc rw
 289 auto quick
-- 
2.14.4

[PATCH v3 2/3] xfstests: iterate dedupe integrity test

[Zorro Lang]

This case does dedupe on a dir, then copy the dir to next dir. Dedupe
the next dir again, then copy this dir to next again, and dedupe
again ... At the end, verify the data in the last dir is still same
with the first one.

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

V3 update description about SPDX-License

Thanks,
Zorro

 tests/shared/009     | 104 +++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/shared/009.out |   4 ++
 tests/shared/group   |   1 +
 3 files changed, 109 insertions(+)
 create mode 100755 tests/shared/009
 create mode 100644 tests/shared/009.out

diff --git a/tests/shared/009 b/tests/shared/009
new file mode 100755
index 00000000..933d3098
--- /dev/null
+++ b/tests/shared/009
@@ -0,0 +1,104 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2018 Red Hat Inc.  All Rights Reserved.
+#
+# FS QA Test 009
+#
+# Iterate dedupe integrity test. Copy an original data0 several
+# times (d0 -> d1, d1 -> d2, ... dn-1 -> dn), dedupe dataN everytime
+# before copy. At last, verify dataN same with data0.
+#
+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 $seqres.full
+
+# real QA test starts here
+
+# duperemove only supports btrfs and xfs (with reflink feature).
+# Add other filesystems if it supports more later.
+_supported_fs xfs btrfs
+_supported_os Linux
+_require_scratch_dedupe
+_require_command "$DUPEREMOVE_PROG" duperemove
+
+_scratch_mkfs > $seqres.full 2>&1
+_scratch_mount >> $seqres.full 2>&1
+
+function iterate_dedup_verify()
+{
+	local src=$srcdir
+	local dest=$dupdir/1
+
+	for ((index = 1; index <= times; index++)); do
+		cp -a $src $dest
+		find $dest -type f -exec md5sum {} \; \
+			> $md5file$index
+		# Make some noise
+		$FSSTRESS_PROG $fsstress_opts -d $noisedir \
+			       -n 200 -p $((5 * LOAD_FACTOR)) >/dev/null 2>&1
+		# Too many output, so only save error output
+		$DUPEREMOVE_PROG -dr --dedupe-options=same $dupdir \
+			>/dev/null 2>$seqres.full
+		md5sum -c --quiet $md5file$index
+		src=$dest
+		dest=$dupdir/$((index + 1))
+	done
+}
+
+srcdir=$SCRATCH_MNT/src
+dupdir=$SCRATCH_MNT/dup
+noisedir=$dupdir/noise
+mkdir $srcdir $dupdir
+mkdir $dupdir/noise
+
+md5file=${tmp}.md5sum
+
+fsstress_opts="-w -r"
+# Create some files to be original data
+$FSSTRESS_PROG $fsstress_opts -d $srcdir \
+	       -n 500 -p $((5 * LOAD_FACTOR)) >/dev/null 2>&1
+
+# Calculate how many test cycles will be run
+src_size=`du -ks $srcdir | awk '{print $1}'`
+free_size=`df -kP $SCRATCH_MNT | grep -v Filesystem | awk '{print $4}'`
+times=$((free_size / src_size))
+if [ $times -gt $((4 * TIME_FACTOR)) ]; then
+	times=$((4 * TIME_FACTOR))
+fi
+
+echo "= Do dedup and verify ="
+iterate_dedup_verify
+
+# Use the last checksum file to verify the original data
+sed -e s#dup/$times#src#g $md5file$times > $md5file
+echo "= Backwords verify ="
+md5sum -c --quiet $md5file
+
+# read from the disk also doesn't show mutations.
+_scratch_cycle_mount
+echo "= Verify after cycle mount ="
+for ((index = 1; index <= times; index++)); do
+	md5sum -c --quiet $md5file$index
+done
+
+status=0
+exit
diff --git a/tests/shared/009.out b/tests/shared/009.out
new file mode 100644
index 00000000..44a78ba3
--- /dev/null
+++ b/tests/shared/009.out
@@ -0,0 +1,4 @@
+QA output created by 009
+= Do dedup and verify =
+= Backwords verify =
+= Verify after cycle mount =
diff --git a/tests/shared/group b/tests/shared/group
index 49ffa8dd..9c484794 100644
--- a/tests/shared/group
+++ b/tests/shared/group
@@ -11,6 +11,7 @@
 006 auto enospc
 007 dangerous_fuzzers
 008 auto stress dedupe
+009 auto stress dedupe
 032 mkfs auto quick
 272 auto enospc rw
 289 auto quick
-- 
2.14.4

[PATCH v3 3/3] xfstests: dedupe with random io race test

[Zorro Lang]

Run several duperemove processes with fsstress on same directory at
same time. Make sure the race won't break the fs or kernel.

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

V3 did below changes:
1) Use $TEST_DIR/${seq}-running file to loop run duperemove.
2) Change kill_all_stress function
3) Change $TEST_DIR/${seq}md5.sum to ${tmp}.md5sum

Thanks,
Zorro

 tests/shared/010     | 94 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/shared/010.out |  2 ++
 tests/shared/group   |  1 +
 3 files changed, 97 insertions(+)
 create mode 100755 tests/shared/010
 create mode 100644 tests/shared/010.out

diff --git a/tests/shared/010 b/tests/shared/010
new file mode 100755
index 00000000..3cf50ddd
--- /dev/null
+++ b/tests/shared/010
@@ -0,0 +1,94 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2018 Red Hat Inc.  All Rights Reserved.
+#
+# FS QA Test 010
+#
+# Dedup & random I/O race test, do multi-threads fsstress and dedupe on
+# same directory/files
+#
+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.*
+	end_test
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+. ./common/reflink
+
+# remove previous $seqres.full before test
+rm -f $seqres.full
+
+# real QA test starts here
+
+# duperemove only supports btrfs and xfs (with reflink feature).
+# Add other filesystems if it supports more later.
+_supported_fs xfs btrfs
+_supported_os Linux
+_require_scratch_dedupe
+_require_command "$DUPEREMOVE_PROG" duperemove
+_require_command "$KILLALL_PROG" killall
+
+_scratch_mkfs > $seqres.full 2>&1
+_scratch_mount >> $seqres.full 2>&1
+
+function end_test()
+{
+	local f=1
+
+	# stop duperemove running
+	if [ -e $dupe_run ]; then
+		rm -f $dupe_run
+		wait $dedup_pids
+	fi
+
+	# Make sure all fsstress get killed
+	while [ $f -ne 0 ]; do
+		$KILLALL_PROG -q $FSSTRESS_PROG > /dev/null 2>&1
+		sleep 1
+		f=`ps -eLf | grep $FSSTRESS_PROG | grep -v "grep" | wc -l`
+	done
+}
+
+sleep_time=$((50 * TIME_FACTOR))
+
+# Start fsstress
+fsstress_opts="-r -n 1000 -p $((5 * LOAD_FACTOR))"
+$FSSTRESS_PROG $fsstress_opts -d $SCRATCH_MNT -l 0 >> $seqres.full 2>&1 &
+dedup_pids=""
+dupe_run=$TEST_DIR/${seq}-running
+# Start several dedupe processes on same directory
+touch $dupe_run
+for ((i = 0; i < $((2 * LOAD_FACTOR)); i++)); do
+	while [ -e $dupe_run ]; do
+		$DUPEREMOVE_PROG -dr --dedupe-options=same $SCRATCH_MNT/
+			>>$seqres.full 2>&1
+	done &
+	dedup_pids="$! $dedup_pids"
+done
+
+# End the test after $sleep_time seconds
+sleep $sleep_time
+end_test
+
+# umount and mount again, verify pagecache contents don't mutate and a fresh
+# read from the disk also doesn't show mutations.
+find $testdir -type f -exec md5sum {} \; > ${tmp}.md5sum
+_scratch_cycle_mount
+md5sum -c --quiet ${tmp}.md5sum
+
+echo "Silence is golden"
+status=0
+exit
diff --git a/tests/shared/010.out b/tests/shared/010.out
new file mode 100644
index 00000000..1d83a8d6
--- /dev/null
+++ b/tests/shared/010.out
@@ -0,0 +1,2 @@
+QA output created by 010
+Silence is golden
diff --git a/tests/shared/group b/tests/shared/group
index 9c484794..094da27d 100644
--- a/tests/shared/group
+++ b/tests/shared/group
@@ -12,6 +12,7 @@
 007 dangerous_fuzzers
 008 auto stress dedupe
 009 auto stress dedupe
+010 auto stress dedupe
 032 mkfs auto quick
 272 auto enospc rw
 289 auto quick
-- 
2.14.4

Re: [PATCH v3 3/3] xfstests: dedupe with random io race test

[Eryu Guan]

On Sun, Jun 24, 2018 at 02:00:29AM +0800, Zorro Lang wrote:
> Run several duperemove processes with fsstress on same directory at
> same time. Make sure the race won't break the fs or kernel.
> 
> Signed-off-by: Zorro Lang <zlang@redhat.com>
> ---
> 
> V3 did below changes:
> 1) Use $TEST_DIR/${seq}-running file to loop run duperemove.
> 2) Change kill_all_stress function
> 3) Change $TEST_DIR/${seq}md5.sum to ${tmp}.md5sum
> 
> Thanks,
> Zorro
> 
>  tests/shared/010     | 94 ++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/shared/010.out |  2 ++
>  tests/shared/group   |  1 +
>  3 files changed, 97 insertions(+)
>  create mode 100755 tests/shared/010
>  create mode 100644 tests/shared/010.out
> 
> diff --git a/tests/shared/010 b/tests/shared/010
> new file mode 100755
> index 00000000..3cf50ddd
> --- /dev/null
> +++ b/tests/shared/010
> @@ -0,0 +1,94 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2018 Red Hat Inc.  All Rights Reserved.
> +#
> +# FS QA Test 010
> +#
> +# Dedup & random I/O race test, do multi-threads fsstress and dedupe on
> +# same directory/files
> +#
> +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.*
> +	end_test
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/reflink
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +
> +# duperemove only supports btrfs and xfs (with reflink feature).
> +# Add other filesystems if it supports more later.
> +_supported_fs xfs btrfs
> +_supported_os Linux
> +_require_scratch_dedupe
> +_require_command "$DUPEREMOVE_PROG" duperemove
> +_require_command "$KILLALL_PROG" killall
> +
> +_scratch_mkfs > $seqres.full 2>&1
> +_scratch_mount >> $seqres.full 2>&1
> +
> +function end_test()
> +{
> +	local f=1
> +
> +	# stop duperemove running
> +	if [ -e $dupe_run ]; then
> +		rm -f $dupe_run
> +		wait $dedup_pids
> +	fi
> +
> +	# Make sure all fsstress get killed
> +	while [ $f -ne 0 ]; do
> +		$KILLALL_PROG -q $FSSTRESS_PROG > /dev/null 2>&1
> +		sleep 1
> +		f=`ps -eLf | grep $FSSTRESS_PROG | grep -v "grep" | wc -l`
> +	done
> +}
> +
> +sleep_time=$((50 * TIME_FACTOR))
> +
> +# Start fsstress
> +fsstress_opts="-r -n 1000 -p $((5 * LOAD_FACTOR))"
> +$FSSTRESS_PROG $fsstress_opts -d $SCRATCH_MNT -l 0 >> $seqres.full 2>&1 &
> +dedup_pids=""
> +dupe_run=$TEST_DIR/${seq}-running
> +# Start several dedupe processes on same directory
> +touch $dupe_run
> +for ((i = 0; i < $((2 * LOAD_FACTOR)); i++)); do
> +	while [ -e $dupe_run ]; do
> +		$DUPEREMOVE_PROG -dr --dedupe-options=same $SCRATCH_MNT/

This needs a "\" at the end of line, otherwise the output all go to
stdout then breaks golden image. I fixed it on commit.

Thanks,
Eryu

> +			>>$seqres.full 2>&1
> +	done &
> +	dedup_pids="$! $dedup_pids"
> +done
> +
> +# End the test after $sleep_time seconds
> +sleep $sleep_time
> +end_test
> +
> +# umount and mount again, verify pagecache contents don't mutate and a fresh
> +# read from the disk also doesn't show mutations.
> +find $testdir -type f -exec md5sum {} \; > ${tmp}.md5sum
> +_scratch_cycle_mount
> +md5sum -c --quiet ${tmp}.md5sum
> +
> +echo "Silence is golden"
> +status=0
> +exit
> diff --git a/tests/shared/010.out b/tests/shared/010.out
> new file mode 100644
> index 00000000..1d83a8d6
> --- /dev/null
> +++ b/tests/shared/010.out
> @@ -0,0 +1,2 @@
> +QA output created by 010
> +Silence is golden
> diff --git a/tests/shared/group b/tests/shared/group
> index 9c484794..094da27d 100644
> --- a/tests/shared/group
> +++ b/tests/shared/group
> @@ -12,6 +12,7 @@
>  007 dangerous_fuzzers
>  008 auto stress dedupe
>  009 auto stress dedupe
> +010 auto stress dedupe
>  032 mkfs auto quick
>  272 auto enospc rw
>  289 auto quick
> -- 
> 2.14.4
> 
> --
> To unsubscribe from this list: send the line "unsubscribe fstests" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html