| Message ID | 1464362482-7704-2-git-send-email-zlang@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 5/27/16 10:21 AM, Zorro Lang wrote: > XFS bring in a new configuration under /sys/fs/xfs/error, named > fail_at_unmount. It's used to stop unmount retrying forever when > it hit IO error. > > This case try to unmount an faulty dm device, and to sure unmount > won't retry forever if fail_at_unmount=1. > > Signed-off-by: Zorro Lang <zlang@redhat.com> > --- > > Hi, > > Thanks Carlos told this easy way(use dm error) to test this feature, or > I will try to use dm thinp to reproduce unmount hit an ENOSPC error. That's > more difficult to do. right, different errors from dm_thinp. But I think my simple dm_thinp test and the infrastructure it uses should make this easy? It was merged recently. > I haven't gotten some good ideas to test those /sys/fs/xfs/dm-X/error/ > metadata/... configurations. If someone have good ideas, please let me > know:) > > Thanks, > Zorro > > tests/xfs/006 | 80 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ > tests/xfs/006.out | 2 ++ > tests/xfs/group | 1 + > 3 files changed, 83 insertions(+) > create mode 100755 tests/xfs/006 > create mode 100644 tests/xfs/006.out > > diff --git a/tests/xfs/006 b/tests/xfs/006 > new file mode 100755 > index 0000000..dafab50 > --- /dev/null > +++ b/tests/xfs/006 > @@ -0,0 +1,80 @@ > +#! /bin/bash > +# FS QA Test 006 > +# > +# Test xfs' "fail at unmount" error handling configuration. Stop > +# XFS retrying to writeback forever when unmount. > +# > +#----------------------------------------------------------------------- > +# Copyright (c) 2016 YOUR NAME HERE. 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/dmerror > + > +# remove previous $seqres.full before test > +rm -f $seqres.full > + > +# real QA test starts here > +_supported_fs xfs > +_supported_os Linux > +_require_dm_target error > +_require_scratch > +_require_scratch_sys_fs error/fail_at_unmount > + > +# Format and mount > +# The device is still a linear device at here > +_scratch_mkfs > $seqres.full 2>&1 I guess my earlier suggestion was to not mkfs inside the _require function for sysfs, so maybe move the test here: _require_fs_sys_fs $SCRATCH_DEV error/fail_at_unmount > +_dmerror_init > +_dmerror_mount > + > +# Enable fail_at_unmount > +_enable_xfs_fail_at_unmount $DMERROR_DEV _set_fs_sys_fs_param $DMERROR_DEV error/fail_at_unmount 1 > + > +# make some IO at here. If these data still not writeback from > +# journal to data, it will hit writeback error after load error > +# tabel in next step. > +_pwrite_byte 0x61 0 100M $SCRATCH_MNT/file >/dev/null > +_dmerror_load_error_table > + > +# Even above data already all writeback, umount will cause XFS > +# try to writeback something to root inode. It still can trigger > +# umount fail. > +_dmerror_unmount > + > +echo "Silence is golden" > +_dmerror_cleanup > + > +# success, all done > +status=0 > +exit > diff --git a/tests/xfs/006.out b/tests/xfs/006.out > new file mode 100644 > index 0000000..675c1b7 > --- /dev/null > +++ b/tests/xfs/006.out > @@ -0,0 +1,2 @@ > +QA output created by 006 > +Silence is golden > diff --git a/tests/xfs/group b/tests/xfs/group > index f4c6816..39169ea 100644 > --- a/tests/xfs/group > +++ b/tests/xfs/group > @@ -4,6 +4,7 @@ > 004 db auto quick > 005 auto quick > 007 auto quota quick > +006 auto quick unmount > 008 rw ioctl auto quick > 009 rw ioctl auto prealloc quick > 010 auto quick repair > -- 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
On Fri, May 27, 2016 at 01:26:35PM -0500, Eric Sandeen wrote: > On 5/27/16 10:21 AM, Zorro Lang wrote: > > XFS bring in a new configuration under /sys/fs/xfs/error, named > > fail_at_unmount. It's used to stop unmount retrying forever when > > it hit IO error. > > > > This case try to unmount an faulty dm device, and to sure unmount > > won't retry forever if fail_at_unmount=1. > > > > Signed-off-by: Zorro Lang <zlang@redhat.com> > > --- > > > > Hi, > > > > Thanks Carlos told this easy way(use dm error) to test this feature, or > > I will try to use dm thinp to reproduce unmount hit an ENOSPC error. That's > > more difficult to do. > > right, different errors from dm_thinp. > > But I think my simple dm_thinp test and the infrastructure it uses > should make this easy? It was merged recently. Yeah, those dmthin functions looks fine for me, I haven't use them(maybe use them soon). I said use dmthin is "more difficult", not means use your functions is difficult, I mean use dmthin to test this "retry forever" issue is difficult. Because if use dmthinp, I should make sure when I do unmount, there're some journal have been recorded, but the data haven't be writeback. So when unmount try to flush these data, it will hit ENOSPC. right? I think it's too difficult to create this situation. But if use dmerror, it's easy to make unmount hit EIO error. Thanks, Zorro > > > I haven't gotten some good ideas to test those /sys/fs/xfs/dm-X/error/ > > metadata/... configurations. If someone have good ideas, please let me > > know:) > > > > Thanks, > > Zorro > > > > tests/xfs/006 | 80 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > tests/xfs/006.out | 2 ++ > > tests/xfs/group | 1 + > > 3 files changed, 83 insertions(+) > > create mode 100755 tests/xfs/006 > > create mode 100644 tests/xfs/006.out > > > > diff --git a/tests/xfs/006 b/tests/xfs/006 > > new file mode 100755 > > index 0000000..dafab50 > > --- /dev/null > > +++ b/tests/xfs/006 > > @@ -0,0 +1,80 @@ > > +#! /bin/bash > > +# FS QA Test 006 > > +# > > +# Test xfs' "fail at unmount" error handling configuration. Stop > > +# XFS retrying to writeback forever when unmount. > > +# > > +#----------------------------------------------------------------------- > > +# Copyright (c) 2016 YOUR NAME HERE. 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/dmerror > > + > > +# remove previous $seqres.full before test > > +rm -f $seqres.full > > + > > +# real QA test starts here > > +_supported_fs xfs > > +_supported_os Linux > > +_require_dm_target error > > +_require_scratch > > +_require_scratch_sys_fs error/fail_at_unmount > > + > > +# Format and mount > > +# The device is still a linear device at here > > +_scratch_mkfs > $seqres.full 2>&1 > > I guess my earlier suggestion was to not mkfs inside the > _require function for sysfs, so maybe move the test here: > > _require_fs_sys_fs $SCRATCH_DEV error/fail_at_unmount hmm... I think generally we do _require_* test before do real test. But do _require_* test after mkfs is still better than after mount:) > > > +_dmerror_init > > +_dmerror_mount > > + > > +# Enable fail_at_unmount > > +_enable_xfs_fail_at_unmount $DMERROR_DEV > > _set_fs_sys_fs_param $DMERROR_DEV error/fail_at_unmount 1 Sure Thanks again for your review, Zorro > > > + > > +# make some IO at here. If these data still not writeback from > > +# journal to data, it will hit writeback error after load error > > +# tabel in next step. > > +_pwrite_byte 0x61 0 100M $SCRATCH_MNT/file >/dev/null > > +_dmerror_load_error_table > > + > > +# Even above data already all writeback, umount will cause XFS > > +# try to writeback something to root inode. It still can trigger > > +# umount fail. > > +_dmerror_unmount > > + > > +echo "Silence is golden" > > +_dmerror_cleanup > > + > > +# success, all done > > +status=0 > > +exit > > diff --git a/tests/xfs/006.out b/tests/xfs/006.out > > new file mode 100644 > > index 0000000..675c1b7 > > --- /dev/null > > +++ b/tests/xfs/006.out > > @@ -0,0 +1,2 @@ > > +QA output created by 006 > > +Silence is golden > > diff --git a/tests/xfs/group b/tests/xfs/group > > index f4c6816..39169ea 100644 > > --- a/tests/xfs/group > > +++ b/tests/xfs/group > > @@ -4,6 +4,7 @@ > > 004 db auto quick > > 005 auto quick > > 007 auto quota quick > > +006 auto quick unmount > > 008 rw ioctl auto quick > > 009 rw ioctl auto prealloc quick > > 010 auto quick repair > > > -- > 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 -- 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
diff --git a/tests/xfs/006 b/tests/xfs/006 new file mode 100755 index 0000000..dafab50 --- /dev/null +++ b/tests/xfs/006 @@ -0,0 +1,80 @@ +#! /bin/bash +# FS QA Test 006 +# +# Test xfs' "fail at unmount" error handling configuration. Stop +# XFS retrying to writeback forever when unmount. +# +#----------------------------------------------------------------------- +# Copyright (c) 2016 YOUR NAME HERE. 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/dmerror + +# remove previous $seqres.full before test +rm -f $seqres.full + +# real QA test starts here +_supported_fs xfs +_supported_os Linux +_require_dm_target error +_require_scratch +_require_scratch_sys_fs error/fail_at_unmount + +# Format and mount +# The device is still a linear device at here +_scratch_mkfs > $seqres.full 2>&1 +_dmerror_init +_dmerror_mount + +# Enable fail_at_unmount +_enable_xfs_fail_at_unmount $DMERROR_DEV + +# make some IO at here. If these data still not writeback from +# journal to data, it will hit writeback error after load error +# tabel in next step. +_pwrite_byte 0x61 0 100M $SCRATCH_MNT/file >/dev/null +_dmerror_load_error_table + +# Even above data already all writeback, umount will cause XFS +# try to writeback something to root inode. It still can trigger +# umount fail. +_dmerror_unmount + +echo "Silence is golden" +_dmerror_cleanup + +# success, all done +status=0 +exit diff --git a/tests/xfs/006.out b/tests/xfs/006.out new file mode 100644 index 0000000..675c1b7 --- /dev/null +++ b/tests/xfs/006.out @@ -0,0 +1,2 @@ +QA output created by 006 +Silence is golden diff --git a/tests/xfs/group b/tests/xfs/group index f4c6816..39169ea 100644 --- a/tests/xfs/group +++ b/tests/xfs/group @@ -4,6 +4,7 @@ 004 db auto quick 005 auto quick 007 auto quota quick +006 auto quick unmount 008 rw ioctl auto quick 009 rw ioctl auto prealloc quick 010 auto quick repair
XFS bring in a new configuration under /sys/fs/xfs/error, named fail_at_unmount. It's used to stop unmount retrying forever when it hit IO error. This case try to unmount an faulty dm device, and to sure unmount won't retry forever if fail_at_unmount=1. Signed-off-by: Zorro Lang <zlang@redhat.com> --- Hi, Thanks Carlos told this easy way(use dm error) to test this feature, or I will try to use dm thinp to reproduce unmount hit an ENOSPC error. That's more difficult to do. I haven't gotten some good ideas to test those /sys/fs/xfs/dm-X/error/ metadata/... configurations. If someone have good ideas, please let me know:) Thanks, Zorro tests/xfs/006 | 80 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/006.out | 2 ++ tests/xfs/group | 1 + 3 files changed, 83 insertions(+) create mode 100755 tests/xfs/006 create mode 100644 tests/xfs/006.out