On Mon, Jul 10, 2017 at 03:34:18PM +0200, Carlos Maiolino wrote:
> Hi folks,
>
> this is the 2nd version of this test case. I still keep the number as
> 999 to avoid conflicts while it's still WIP, next version should be ok
> to make it into a real patch with the right number.
>
> Reason of this patch is still to check if AIL items are being properly
> resubmitted after an error during writeback from buffers containing AIL
> items.
>
> This V2 includes:
>
> - Better indentation
> - More comments
> - Add copyright
> - use $SCRATCH_MNT instead of local created mount point
> - redirect outputs to $seqres.full instead of dev/null
>
> - It doesn't change sleep by wait, because AFAIK, wait requires the process
> being waited to return. Fsfreeze here is expected to hang until the
> underlying device is extended.
>
Thinking more about this, we could keep the sleep and also add the wait
right before the unfreeze, right?
> Let me know your thoughts.
>
> cheers.
>
> Signed-off-by: Carlos Maiolino <cmaiolino@xxxxxxxxxx>
> ---
Mostly looks good to me. A few minor notes...
> tests/xfs/999 | 119 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
> tests/xfs/999.out | 2 +
> tests/xfs/group | 1 +
The test is still under xfs (rather than generic).
> 3 files changed, 122 insertions(+)
> create mode 100755 tests/xfs/999
> create mode 100644 tests/xfs/999.out
>
> diff --git a/tests/xfs/999 b/tests/xfs/999
> new file mode 100755
> index 0000000..b46f1cc
> --- /dev/null
> +++ b/tests/xfs/999
> @@ -0,0 +1,119 @@
> +#! /bin/bash
> +# FS QA Test 999
> +#
> +# Test buffer resubmission after a failed writeback with to a full overcommited
> +# dm-thin device.
> +#
> +# When a dm-thin device reaches its full capacity, but the virtual device still
> +# shows available space, XFS loops indefinitely in xfsaild due items still in
> +# AIL. The buffers containing such items couldn't be resubmitted because the
> +# items were flush locked. Test the kernel fix and ensure the buffers are
> +# properly resubmitted.
> +#
> +# This test will hang the filesystem when ran on an unpatched kernel
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2017 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.*
> + $UMOUNT_PROG $SCRATCH_MNT >>$seqres.full 2>&1
> + $LVM_PROG vgremove -ff $vgname >>$seqres.full 2>&1
> + $LVM_PROG pvremove -ff $SCRATCH_DEV >>$seqres.full 2>&1
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
Is filter used anywhere?
> +
> +# real QA test starts here
> +
> +# Modify as appropriate.
> +_supported_fs xfs
> +_supported_os Linux
> +_require_scratch_nocheck
> +_require_dm_target thin-pool
> +_require_command $LVM_PROG lvm
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +vgname=vg_$seq
> +lvname=lv_$seq
> +poolname=pool_$seq
> +snapname=snap_$seq
> +origpsize=100
> +virtsize=200
> +newpsize=200
> +
> +# Ensure we have enough disk space
> +_scratch_mkfs_sized $((250 * 1024 * 1024)) >>$seqres.full 2>&1
> +
> +# Create a 100MB dm-thin POOL
> +$LVM_PROG pvcreate -f $SCRATCH_DEV >>$seqres.full 2>&1
> +$LVM_PROG vgcreate -f $vgname $SCRATCH_DEV >>$seqres.full 2>&1
> +
> +$LVM_PROG lvcreate --thinpool $poolname --errorwhenfull y \
> + --zero n -L $origpsize \
> + --poolmetadatasize 4M $vgname >>$seqres.full 2>&1
> +
> +# Create a overprovisioned 200MB dm-thin virt. device
> +$LVM_PROG lvcreate --virtualsize $virtsize \
> + -T $vgname/$poolname \
> + -n $lvname >>$seqres.full 2>&1
> +
> +_mkfs_dev /dev/mapper/$vgname-$lvname >>$seqres.full 2>&1
> +
> +
> +$LVM_PROG lvcreate -k n -s $vgname/$lvname \
> + -n $snapname >>$seqres.full 2>&1
What's the reason for using a snapshot? Is the original thin vol not
sufficient?
> +
> +_mount /dev/mapper/$vgname-$snapname $SCRATCH_MNT
> +
> +# Consume all space available in the volume and freeze to ensure everything
> +# required to make the fs consistent is flushed to disk.
> +xfs_io -f -d -c 'pwrite -b 1m 0 120m' $SCRATCH_MNT/f1 >>$seqres.full 2>&1
$XFS_IO_PROG
> +
> +# This freeze will never complete until the dm-thin POOL device is extended.
> +# This is expected, it is only used so xfsaild is triggered to flush AIL items.
> +fsfreeze -f $SCRATCH_MNT &
> +
> +# Wait enough so xfsaild can run
> +sleep 10
> +
> +# Make some extra space available so the freeze above can proceed
> +lvextend -L $newpsize $vgname/$poolname >>$seqres.full 2>&1
$LVM_PROG lvmextend ?
Brian
> +
> +# Try to thaw the filesystem, and complete test if if succeed.
> +# NOTE: This will hang on affected XFS filesystems.
> +fsfreeze -u $SCRATCH_MNT
> +echo "Test OK"
> +
> +status=0
> +exit
> diff --git a/tests/xfs/999.out b/tests/xfs/999.out
> new file mode 100644
> index 0000000..8c3c938
> --- /dev/null
> +++ b/tests/xfs/999.out
> @@ -0,0 +1,2 @@
> +QA output created by 999
> +Test OK
> diff --git a/tests/xfs/group b/tests/xfs/group
> index 792161a..2bde916 100644
> --- a/tests/xfs/group
> +++ b/tests/xfs/group
> @@ -416,3 +416,4 @@
> 416 dangerous_fuzzers dangerous_scrub dangerous_repair
> 417 dangerous_fuzzers dangerous_scrub dangerous_online_repair
> 418 dangerous_fuzzers dangerous_scrub dangerous_repair
> +999 dangerous
> --
> 2.9.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe fstests" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
