On Tue, Nov 01, 2016 at 07:19:30PM +0800, Wang Xiaoguang wrote:
In btrfs, sometimes though the number of created files' consumed disk space
are not larger than fs's free space, we can still get some ENOSPC error, it
may be that btrfs does not try hard to reclaim disk space(I have sent kernel
patch to resolve this kind of enospc error. Note, this false enospc error
will not always happen even in kernel without my fixing patch).
Currently only in btrfs, I get this ENOSPC error, xfs and ext4 work well.
Signed-off-by: Wang Xiaoguang <wangxg.f...@cn.fujitsu.com>
---
tests/generic/389 | 78 +++++++++++++++++++++++++++++++++++++++++++++++++++
tests/generic/389.out | 2 ++
tests/generic/group | 1 +
3 files changed, 81 insertions(+)
create mode 100755 tests/generic/389
create mode 100644 tests/generic/389.out
diff --git a/tests/generic/389 b/tests/generic/389
new file mode 100755
index 0000000..96bc12e
--- /dev/null
+++ b/tests/generic/389
@@ -0,0 +1,78 @@
+#! /bin/bash
+# FS QA Test 389
+#
+# Create and delete files repeatedly to exercise ENOSPC behaviour.
Trailing whitespace in this line.
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2016 Fujitsu. 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
+
+# remove previous $seqres.full before test
+rm -f $seqres.full
+
+# Modify as appropriate.
+_supported_fs generic
+_supported_os Linux
+_require_scratch
+
+RUN_TIME=$((600 * $TIME_FACTOR))
Hmm, does it really need 600s to run? I think it's better to limit the
runtime within 300s and make it an 'auto' test. I, personally, prefer a
"loop count" based test, I'd find out a minimum loop count that could
reproduce the ENOSPC problem more reliably on btrfs (for example, say
75%) and make the count scale with LOAD_FACTOR.
+fs_size=$((15 * 1024 * 1024 * 1024))
And does it really need 15G on SCRATCH_DEV? A smaller fs size makes test
run faster, and gives the test more chance to be run, because not
everyone has a 15G SCRATCH_DEV.
+_scratch_mkfs_sized $fs_size > $seqres.full 2>&1
+_scratch_mount > $seqres.full 2>&1
Append to $seqres.full not overwrite.
+
+testfile1=$SCRATCH_MNT/testfile1
+testfile2=$SCRATCH_MNT/testfile2
+filesize1=$(((fs_size * 80) / 100))
+filesize2=$(((fs_size * 5) / 100))
Better to have some comments on the filesizes chosen here. e.g. someone
may wonder that why it's testing ENOSPC condition with 85% full, not 99%
or 100%.
+
+do_test()
+{
+ while [ -f $SCRATCH_MNT/run ]; do
+ $XFS_IO_PROG -fc "pwrite 0 $filesize1" $testfile1 > /dev/null
+ $XFS_IO_PROG -fc "pwrite 0 $filesize2" $testfile2 > /dev/null
+ rm -f $testfile1 $testfile2
Trailing whitespace here.
+ done
+}
+
+echo "Silence is golden"
+touch $SCRATCH_MNT/run
+do_test &
+sleep $RUN_TIME
+rm -f $SCRATCH_MNT/run
+wait
+
+status=0
+exit
diff --git a/tests/generic/389.out b/tests/generic/389.out
new file mode 100644
index 0000000..e8c24bb
--- /dev/null
+++ b/tests/generic/389.out
@@ -0,0 +1,2 @@
+QA output created by 389
+Silence is golden
Can you please rebase on top of current master? generic/389 is already
taken, and it makes applying & testing the patch a litter harder :)
diff --git a/tests/generic/group b/tests/generic/group
index fc32cfd..b6d4013 100644
--- a/tests/generic/group
+++ b/tests/generic/group
@@ -391,3 +391,4 @@
386 auto quick quota
387 auto clone
388 auto log metadata
+389 enospc
Perhaps we can add it to 'rw' group too.
Thanks,
Eryu
