From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: eguan@redhat.com
Cc: linux-xfs@vger.kernel.org, fstests@vger.kernel.org
Subject: [PATCH v4 8/8] xfs/068: fix variability problems in file/dir count output
Date: Thu, 14 Dec 2017 18:17:27 -0800 [thread overview]
Message-ID: <20171215021727.GV19219@magnolia> (raw)
In-Reply-To: <151314505158.18893.11894289091110903029.stgit@magnolia>
From: Darrick J. Wong <darrick.wong@oracle.com>
In these tests we use a fixed sequence of operations in fsstress to
create a directory tree and then exercise xfsdump/xfsrestore on that.
However, this changes every time someone adds a new fsstress command,
or someone runs with FSSTRESS_AVOID, etc.
Therefore, check the counts directly from xfsrestore output instead
of relying on the golden output; check that the paths of every file
in the directory tree match; and check that the md5 of every regular
file matches.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
v4: add lots more checking of everything we restored
---
common/dump | 36 ++++++++++++++++++++++++++++++++++--
tests/xfs/027.out | 2 +-
tests/xfs/068 | 9 +++++++++
tests/xfs/068.out | 2 +-
4 files changed, 45 insertions(+), 4 deletions(-)
diff --git a/common/dump b/common/dump
index 898aaa4..f4edd53 100644
--- a/common/dump
+++ b/common/dump
@@ -316,7 +316,8 @@ _create_dumpdir_stress_num()
echo "-----------------------------------------------"
echo "fsstress : $_param"
echo "-----------------------------------------------"
- if ! $here/ltp/fsstress $_param -s 1 $FSSTRESS_AVOID -n $_count -d $dump_dir >$tmp.out 2>&1
+ echo $FSSTRESS_PROG $_param -s 1 $FSSTRESS_AVOID -n $_count -d $dump_dir >> $seqres.full
+ if ! $FSSTRESS_PROG $_param -s 1 $FSSTRESS_AVOID -n $_count -d $dump_dir >$tmp.out 2>&1
then
echo " fsstress (count=$_count) returned $? - see $seqres.full"
@@ -1240,9 +1241,40 @@ _do_dump_restore()
echo "xfsdump|xfsrestore ..."
restore_opts="$_restore_debug$restore_args - $restore_dir"
dump_opts="$_dump_debug$dump_args -s $dump_sdir - $SCRATCH_MNT"
+
+ # We expect there to be one more dir (and inode) than what's in dump_dir.
+ # Construct the string we expect to see in the output, since fsstress
+ # will create different directory structures every time someone adds
+ # a new command, runs with a different FSSTRESS_AVOID, etc.
+ expected_dirs=$(find $dump_dir -type d | wc -l)
+ expected_inodes=$(find $dump_dir | wc -l)
+ expected_str=": $((expected_dirs + 1)) directories and $((expected_inodes + 1)) entries processed"
+
+ # Measure the md5 of every file...
+ (cd $dump_dir ; find . -type f -print0 | xargs -0 md5sum) > $seqres.md5
+ (cd $dump_dir ; find | sort) > $seqres.fstree
+
echo "xfsdump $dump_opts | xfsrestore $restore_opts" | _dir_filter
- $XFSDUMP_PROG $dump_opts 2>$tmp.dump.mlog | $XFSRESTORE_PROG $restore_opts 2>&1 | tee -a $seqres.full | _dump_filter
+ $XFSDUMP_PROG $dump_opts 2>$tmp.dump.mlog | $XFSRESTORE_PROG $restore_opts > $tmp.restore.mlog 2>&1
+ cat $tmp.restore.mlog >> $seqres.full
+ echo "xfsrestore output should contain$expected_str" >> $seqres.full
+ cat $tmp.restore.mlog | _dump_filter | sed -e 's/: \([0-9]*\) directories and \([0-9]*\) entries/: XXX directories and YYY entries/g'
_dump_filter <$tmp.dump.mlog
+
+ # Did we actually restore as many dirs/files as we had?
+ if ! grep -q "$expected_str" $tmp.restore.mlog; then
+ echo "mismatch counts between directory tree and restored filesystem"
+ grep "directories and.*entries processed" $tmp.restore.mlog | sed -e 's/^.*:/found:/g'
+ echo "expected$expected_str"
+ fi
+
+ # Does the directory tree match?
+ diff -u $seqres.fstree <(cd $restore_dir/$dump_sdir ; find | sort)
+
+ # Measure the md5 of every restored file...
+ (cd $restore_dir/$dump_sdir ; md5sum --quiet -c $seqres.md5)
+
+ rm -rf $seqres.md5 $seqres.fstree $tmp.restore.mlog $tmp.dump.mlog
}
#
diff --git a/tests/xfs/027.out b/tests/xfs/027.out
index ba425a3..7665021 100644
--- a/tests/xfs/027.out
+++ b/tests/xfs/027.out
@@ -19,7 +19,7 @@ xfsrestore: session id: ID
xfsrestore: media ID: ID
xfsrestore: searching media for directory dump
xfsrestore: reading directories
-xfsrestore: 3 directories and 39 entries processed
+xfsrestore: XXX directories and YYY entries processed
xfsrestore: directory post-processing
xfsrestore: restoring non-directory files
xfsrestore: restore complete: SECS seconds elapsed
diff --git a/tests/xfs/068 b/tests/xfs/068
index 7151e28..9ecb836 100755
--- a/tests/xfs/068
+++ b/tests/xfs/068
@@ -44,6 +44,15 @@ _supported_fs xfs
_supported_os Linux
_create_dumpdir_stress_num 4096
+
+# Let's make sure there's a largeish number of dirs/files here.
+expected_dirs=$(find $dump_dir -type d | wc -l)
+expected_inodes=$(find $dump_dir | wc -l)
+
+if [ $expected_dirs -lt 100 ] || [ $expected_inodes -lt 600 ]; then
+ echo "Oddly small dir tree: $((expected_dirs + 1)) directories and $((expected_inodes + 1)) entries processed"
+fi
+
_do_dump_restore
# success, all done
diff --git a/tests/xfs/068.out b/tests/xfs/068.out
index fa3a552..f53c555 100644
--- a/tests/xfs/068.out
+++ b/tests/xfs/068.out
@@ -22,7 +22,7 @@ xfsrestore: session id: ID
xfsrestore: media ID: ID
xfsrestore: searching media for directory dump
xfsrestore: reading directories
-xfsrestore: 383 directories and 1335 entries processed
+xfsrestore: XXX directories and YYY entries processed
xfsrestore: directory post-processing
xfsrestore: restoring non-directory files
xfsrestore: restore complete: SECS seconds elapsed
next prev parent reply other threads:[~2017-12-15 2:24 UTC|newest]
Thread overview: 55+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-12-13 6:03 [PATCH 0/8] weekly fstests changes Darrick J. Wong
2017-12-13 6:03 ` [PATCH 1/8] common/rc: report kmemleak errors Darrick J. Wong
2017-12-14 9:37 ` Eryu Guan
2017-12-14 18:15 ` Darrick J. Wong
2018-01-05 8:02 ` Eryu Guan
2018-01-05 17:02 ` Darrick J. Wong
2018-01-07 15:25 ` Eryu Guan
2017-12-13 6:03 ` [PATCH 2/8] common/xfs: fix scrub support probing again Darrick J. Wong
2017-12-13 6:03 ` [PATCH 3/8] generic/45[34]: test line draw characters in file/attr names Darrick J. Wong
2017-12-13 6:03 ` [PATCH 4/8] xfs: fix tests to handle removal of no-alloc create nonfeature Darrick J. Wong
2017-12-13 22:12 ` Dave Chinner
2017-12-13 22:45 ` [PATCH v2 " Darrick J. Wong
2017-12-13 6:03 ` [PATCH 5/8] generic: test error shutdown while stressing filesystem Darrick J. Wong
2017-12-13 6:03 ` [PATCH 6/8] fsstress: implement the clonerange/deduperange ioctls Darrick J. Wong
2017-12-14 6:39 ` Amir Goldstein
2017-12-14 7:32 ` Eryu Guan
2017-12-14 20:20 ` Darrick J. Wong
2017-12-15 2:07 ` [PATCH v2 " Darrick J. Wong
2018-01-03 8:48 ` Eryu Guan
2018-01-03 17:12 ` Darrick J. Wong
2018-01-05 4:35 ` Eryu Guan
2018-01-05 4:54 ` Darrick J. Wong
2018-01-06 1:46 ` Darrick J. Wong
2018-01-09 7:09 ` Darrick J. Wong
2018-02-22 16:06 ` Luis Henriques
2018-02-22 17:27 ` Darrick J. Wong
2018-02-22 18:17 ` Luis Henriques
2018-02-22 18:34 ` Darrick J. Wong
2018-02-23 10:17 ` Luis Henriques
2017-12-13 6:04 ` [PATCH 7/8] generic: run a long-soak write-only fsstress test Darrick J. Wong
2018-01-07 15:34 ` Eryu Guan
2017-12-13 6:04 ` [PATCH 8/8] xfs/068: fix variability problems in file/dir count output Darrick J. Wong
2017-12-13 22:20 ` Dave Chinner
2017-12-13 22:23 ` Darrick J. Wong
2017-12-13 22:45 ` Dave Chinner
2017-12-13 23:17 ` Darrick J. Wong
2017-12-13 23:42 ` Dave Chinner
2017-12-13 23:28 ` [PATCH v2 8/8] xfs/068: fix clonerange " Darrick J. Wong
2017-12-13 23:44 ` Dave Chinner
2017-12-14 6:52 ` Amir Goldstein
2017-12-14 7:37 ` Amir Goldstein
2017-12-14 7:49 ` Eryu Guan
2017-12-14 8:15 ` Amir Goldstein
2017-12-14 21:35 ` Dave Chinner
2017-12-15 2:04 ` Darrick J. Wong
2017-12-15 4:37 ` Dave Chinner
2017-12-15 7:06 ` Amir Goldstein
2017-12-15 2:08 ` [PATCH v3 8/8] xfs/068: fix variability " Darrick J. Wong
2017-12-15 2:16 ` Darrick J. Wong
2017-12-15 2:17 ` Darrick J. Wong [this message]
2017-12-15 8:55 ` [PATCH 0/8] weekly fstests changes Eryu Guan
2018-01-03 19:22 ` [PATCH 9/8] xfs: find libxfs api violations Darrick J. Wong
2018-01-03 19:26 ` [PATCH 10/8] xfs: check that fs freeze minimizes required recovery Darrick J. Wong
2018-01-09 11:33 ` Eryu Guan
2018-01-10 0:03 ` Darrick J. Wong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20171215021727.GV19219@magnolia \
--to=darrick.wong@oracle.com \
--cc=eguan@redhat.com \
--cc=fstests@vger.kernel.org \
--cc=linux-xfs@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).