From: Brian Foster Date: Wed, 2 Aug 2017 16:36:13 +0000 (-0400) Subject: xfs: test for log recovery failure after tail overwrite X-Git-Tag: v2022.05.01~1915 X-Git-Url: https://www.infradead.org/git/?a=commitdiff_plain;h=1cc5517771ed4ee133f71d06e1f5d62c22e8ae82;p=users%2Fhch%2Fxfstests-dev.git xfs: test for log recovery failure after tail overwrite XFS is susceptible to log recovery problems if the fs crashes under certain circumstances. If the tail has been pinned for long enough to the log to fill and the next batch of log buffer submissions happen to fail, the filesystem shuts down having potentially overwritten part of the range between the last good tail->head range in the log. This causes log recovery to fail with crc mismatch or invalid log record errors. Add a test that uses XFS DEBUG mode error injection to force the tail overwrite condition with a known bad (crc mismatch) log write and tests that log recovery succeeds. Note that this problem is currently only reproducible with larger (non-default) log buffer sizes (i.e., '-o logbsize=256k') or smaller block sizes (1k). Signed-off-by: Brian Foster Reviewed-by: Eryu Guan Signed-off-by: Eryu Guan --- diff --git a/tests/xfs/057 b/tests/xfs/057 new file mode 100755 index 000000000..74a253e0e --- /dev/null +++ b/tests/xfs/057 @@ -0,0 +1,116 @@ +#! /bin/bash +# FS QA Test No. 057 +# +# Attempt to reproduce log recovery failure by writing corrupt log records over +# the last good tail in the log. The tail is force pinned while a workload runs +# the head as close as possible behind the tail. Once the head is pinned, +# corrupted log records are written to the log and the filesystem shuts down. +# +# While log recovery should handle the corrupted log records, it has historical +# problems dealing with the situation where the corrupted log records may have +# overwritten the tail of the previous good record in the log. If this occurs, +# log recovery may fail. +# +# This can be reproduced more reliably under non-default conditions such as with +# the smallest supported FSB sizes and/or largest supported log buffer sizes and +# counts (logbufs and logbsize mount options). +# +# Note that this test requires a DEBUG mode 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.* + $KILLALL_PROG -9 fsstress > /dev/null 2>&1 + [ -e /sys/fs/xfs/$sdev/errortag/log_item_pin ] && + echo 0 > /sys/fs/xfs/$sdev/errortag/log_item_pin + wait > /dev/null 2>&1 +} + +rm -f $seqres.full + +# get standard environment, filters and checks +. ./common/rc +. ./common/inject + +# real QA test starts here + +# Modify as appropriate. +_supported_fs xfs +_supported_os Linux +_require_xfs_io_error_injection log_item_pin +_require_xfs_io_error_injection log_bad_crc +_require_scratch +_require_command "$KILLALL_PROG" killall + +echo "Silence is golden." + +sdev=$(_short_dev $SCRATCH_DEV) + +# use a small log fs +_scratch_mkfs_sized $((1024 * 1024 * 500)) >> $seqres.full 2>&1 || + _fail "mkfs failed" +_scratch_mount || _fail "mount failed" + +# populate the fs with some data and cycle the mount to reset the log head/tail +$FSSTRESS_PROG -d $SCRATCH_MNT -z -fcreat=1 -p 4 -n 100000 > /dev/null 2>&1 +_scratch_cycle_mount || _fail "cycle mount failed" + +# Pin the tail and start a file removal workload. File removal tends to +# reproduce the corruption more reliably. +_scratch_inject_error log_item_pin 1 + +rm -rf $SCRATCH_MNT/* > /dev/null 2>&1 & +workpid=$! + +# wait for the head to stop pushing forward +prevhead=-1 +head=`cat /sys/fs/xfs/$sdev/log/log_head_lsn` +while [ "$head" != "$prevhead" ]; do + sleep 5 + prevhead=$head + head=`cat /sys/fs/xfs/$sdev/log/log_head_lsn` +done + +# Once the head is pinned behind the tail, enable log record corruption and +# unpin the tail. All subsequent log buffer writes end up corrupted on-disk and +# result in log I/O errors. +_scratch_inject_error log_bad_crc 1 +_scratch_inject_error log_item_pin 0 + +# wait for fs shutdown to kill the workload +wait $workpid + +# cycle mount to test log recovery +_scratch_cycle_mount + +# success, all done +status=0 +exit diff --git a/tests/xfs/057.out b/tests/xfs/057.out new file mode 100644 index 000000000..e0d648661 --- /dev/null +++ b/tests/xfs/057.out @@ -0,0 +1,2 @@ +QA output created by 057 +Silence is golden. diff --git a/tests/xfs/group b/tests/xfs/group index cf876a298..617c3d2ad 100644 --- a/tests/xfs/group +++ b/tests/xfs/group @@ -54,6 +54,7 @@ 054 auto quick 055 dump ioctl remote tape 056 dump ioctl auto quick +057 auto log 059 dump ioctl auto quick 060 dump ioctl auto quick 061 dump ioctl auto quick