On Wed, Apr 28, 2021 at 9:46 AM Johannes Thumshirn
<johannes.thumshirn@xxxxxxx> wrote:
>
> Add a test for the patch titled "btrfs: zoned: automatically reclaim
> zones".
Since this is already in Linus' tree, we could specify here the commit.
>
> This test creates a two file on a newly created FS in a way that when we
> delete the first one, an auto reclaim process will be triggered by the FS.
>
> After the reclaim process, it verifies that the data was moved to another
> zone and old zone was successfully reset.
>
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@xxxxxxx>
> ---
> common/config | 1 +
> tests/btrfs/236 | 103 ++++++++++++++++++++++++++++++++++++++++++++
> tests/btrfs/236.out | 2 +
> tests/btrfs/group | 1 +
> 4 files changed, 107 insertions(+)
> create mode 100755 tests/btrfs/236
> create mode 100644 tests/btrfs/236.out
>
> diff --git a/common/config b/common/config
> index a47e462c7792..1a26934985dd 100644
> --- a/common/config
> +++ b/common/config
> @@ -226,6 +226,7 @@ export FSVERITY_PROG="$(type -P fsverity)"
> export OPENSSL_PROG="$(type -P openssl)"
> export ACCTON_PROG="$(type -P accton)"
> export E2IMAGE_PROG="$(type -P e2image)"
> +export BLKZONE_PROG="$(type -P blkzone)"
>
> # 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/btrfs/236 b/tests/btrfs/236
> new file mode 100755
> index 000000000000..3be74196ec5d
> --- /dev/null
> +++ b/tests/btrfs/236
> @@ -0,0 +1,103 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2021 Western Digital Corporation. All Rights Reserved.
> +#
> +# FS QA Test 236
> +#
> +# Test that zone autoreclaim works as expected, that is: if the dirty
> +# threashold is exceeded the data gets relocated to new block group and the
threashold -> threshold
> +# old block group gets deleted. On block group deletion, the underlying device
> +# zone also needs to be reset.
> +#
> +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
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +
> +_supported_fs btrfs
> +_require_scratch
> +_require_btrfs_command inspect-internal dump-tree
> +_require_btrfs_command filesystem sync
> +_require_command "$BLKZONE_PROG" blkzone
> +_require_zoned_device "$SCRATCH_DEV"
> +
> +get_data_bg()
> +{
> + $BTRFS_UTIL_PROG inspect-internal dump-tree -t CHUNK $SCRATCH_DEV |\
> + grep -A 1 "CHUNK_ITEM" | grep -B 1 "type DATA" |\
> + grep -Eo "CHUNK_ITEM [[:digit:]]+" | cut -d ' ' -f 2
> +}
> +
> +zonesize=$(cat /sys/block/$(_short_dev $SCRATCH_DEV)/queue/chunk_sectors)
> +zonesize=$((zonesize << 9))
> +
> +_scratch_mkfs >/dev/null 2>&1
> +_scratch_mount -o commit=5 # 5s commit time to speed up test
Why not 1s? Would even be faster.
> +
> +uuid=$(findmnt -n -o UUID "$SCRATCH_MNT")
> +reclaim_threshold=75
> +echo $reclaim_threshold > /sys/fs/btrfs/"$uuid"/bg_reclaim_threshold
> +fill_percent=$((reclaim_threshold + 2))
> +rest_percent=$((90 - fill_percent)) # make sure we're not creating a new BG
> +fill_size=$((zonesize * fill_percent / 100))
> +rest=$((zonesize * rest_percent / 100))
> +
> +# step 1, fill FS over $fillsize
> +$XFS_IO_PROG -fdc "pwrite -W 0 $fill_size" $SCRATCH_MNT/$seq.test1 >> $seqres.full
> +$XFS_IO_PROG -fdc "pwrite -W 0 $rest" $SCRATCH_MNT/$seq.test2 >> $seqres.full
> +sleep 5 # need to make sure the transaction got committed
I don't get this.
Why not just call 'sync'? That commits the transaction too.
Doing regular buffered writes followed by sync is more intuitive and faster.
If for some reason direct IO is really needed, it also misses a
_require_odirect.
> +
> +zones_before=$($BLKZONE_PROG report $SCRATCH_DEV | grep -v -e em -e nw | wc -l)
> +echo "Before reclaim: $zones_before zones open" >> $seqres.full
> +old_data_zone=$(get_data_bg)
> +old_data_zone=$((old_data_zone >> 9))
> +printf "Old data zone 0x%x\n" $old_data_zone >> $seqres.full
> +
> +# step 2, delete the 1st $fill_size sized file to trigger reclaim
> +rm $SCRATCH_MNT/$seq.test1
> +$BTRFS_UTIL_PROG filesystem sync $SCRATCH_MNT
> +sleep 10 # 1 transaction commit for 'rm' and one for balance
> +
> +# check that we don't have more zones open than before
> +zones_after=$($BLKZONE_PROG report $SCRATCH_DEV | grep -v -e em -e nw | wc -l)
> +echo "After reclaim: $zones_after zones open" >> $seqres.full
> +
> +# Check that old data zone was reset
> +old_wptr=$($BLKZONE_PROG report -o $old_data_zone -c 1 $SCRATCH_DEV |\
> + grep -Eo "wptr 0x[[:xdigit:]]+" | cut -d ' ' -f 2)
> +if [ "$old_wptr" != "0x000000" ]; then
> + _fail "Old wptr still at $old_wptr"
> +fi
> +
> +new_data_zone=$(get_data_bg)
> +new_data_zone=$((new_data_zone >> 9))
> +printf "New data zone 0x%x\n" $new_data_zone >> $seqres.full
> +
> +# Check that data was really relocated to a different zone
> +if [ $old_data_zone -eq $new_data_zone ]; then
> + _fail "New zone same as old zone"
The use of _fail is usually discouraged. A simple echo would suffice here.
Other than that, it looks good, but I don't have an environment to
test zoned devices, so it's just eyeballing.
Thanks.
> +fi
> +
> +# success, all done
> +echo "Silence is golden"
> +status=0
> +exit
> diff --git a/tests/btrfs/236.out b/tests/btrfs/236.out
> new file mode 100644
> index 000000000000..b6b6e0cad9a7
> --- /dev/null
> +++ b/tests/btrfs/236.out
> @@ -0,0 +1,2 @@
> +QA output created by 236
> +Silence is golden
> diff --git a/tests/btrfs/group b/tests/btrfs/group
> index 331dd432fac3..62c9c761e974 100644
> --- a/tests/btrfs/group
> +++ b/tests/btrfs/group
> @@ -238,3 +238,4 @@
> 233 auto quick subvolume
> 234 auto quick compress rw
> 235 auto quick send
> +236 auto quick balance
> --
> 2.31.1
>
--
Filipe David Manana,
“Whether you think you can, or you think you can't — you're right.”
