Skip to contentSkip to conversation

xfstests btrfs/284: shorten duration, fix output

PATCH In review Eric Sandeen 4 participants 6 messages Apr 26, 2013 – May 15, 2013 Btrfs file system

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 xfstests btrfs/284: shorten duration, fix output

test 284 had... some issues.

First, it took so long nobody ran it; so shorten the extent count by a factor of about 100.

Having fixed that, we see failures in 2 cases; when start or len is -1, but the golden output file didn't have error output, as if they should pass.

I'm going to argue that these *should* both fail; start = -1 has no real meaning. length = -1 might mean "the rest of the file" but if that's what you really want, just don't specify -l.

So add failure output for those cases.

Send all command output to $seq.full, in case that changes in the future; just capture the return value.

Then remove the return value echo on failure (50?) because who knows when that might change to some other magic value.

Ok, then when defrag actually works, old defrag returned "20" (because?) but a recent commit changed it to 0. So accommodate that too.

And remove a stray "HAVE_DEFRAG=1" while we're at it. That variable is never used.

---

tests/btrfs/284 +14 -11
old mode 100644
new mode 100755
index d952977..67161a3
@@ -26,7 +26,7 @@ seqres=$RESULT_DIR/$seq
 echo "QA output created by $seq"
 here="`pwd`"
 tmp=/tmp/$$
-cnt=11999
+cnt=119
 filesize=48000
 
 status=1	# failure is the default!
@@ -58,11 +58,12 @@ _create_file()
 _btrfs_online_defrag()
 {
 	str=""
+	# start = -1 is invalid, should fail
 	if [ "$2" = "2" ];then
 		str="$str -s -1 -l $((filesize / 2)) "
 	elif [ "$2" = "3" ];then
 		str="$str -s $((filesize + 1)) -l $((filesize / 2)) "
-		HAVE_DEFRAG=1
+	# len = -1 is invalid, should fail
 	elif [ "$2" = "4" ];then
 		str="$str -l -1 "
 	elif [ "$2" = "5" ];then
@@ -76,20 +77,22 @@ _btrfs_online_defrag()
 	fi
 
 	if [ "$str" != "" ]; then
-		$BTRFS_UTIL_PROG filesystem defragment $str $SCRATCH_MNT/tmp_file
+		$BTRFS_UTIL_PROG filesystem defragment $str $SCRATCH_MNT/tmp_file >> $seq.full 2>&1
 	else
 		if [ "$1" = "1" ];then
-			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT/tmp_file
+			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT/tmp_file >> $seq.full 2>&1
 		elif [ "$1" = "2" ];then
-			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT/tmp_dir
+			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT/tmp_dir >> $seq.full 2>&1
 		elif [ "$1" = "3" ];then
-			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT
+			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT >> $seq.full 2>&1
 		fi
 	fi
 	ret_val=$?
 	_scratch_remount
-	if [ $ret_val -ne 20 ];then
-		echo "btrfs filesystem defragment failed! err is $ret_val"
+	# Older defrag returned "20" for success
+	# e9393c2 btrfs-progs: defrag return zero on success
+	if [ $ret_val -ne 0 -a $ret_val -ne 20 ]; then
+		echo "btrfs filesystem defragment failed!"
 	fi
 }
 
@@ -140,19 +143,19 @@ _scratch_mount
 _require_defrag
 
 echo "defrag object | defragment range | defragment compress"
-echo "a single file |  default | off"
+echo "a single file | default | off"
 _rundefrag 1 1 1
 
 echo "a single file | default |  on"
 _rundefrag 1 1 2
 
-echo "a single file | start < 0 && 0 < len < file size | off"
+echo "a single file | start < 0 && 0 < len < file size | off (should fail)"
 _rundefrag 1 2 1
 
 echo "a single file | start > file size && 0 < len < file size | off"
 _rundefrag 1 3 1
 
-echo "a single file | start = 0 && len < 0 | off"
+echo "a single file | start = 0 && len < 0 | off (should fail)"
 _rundefrag 1 4 1
 
 echo "a single file | start = 0 && len > file size | off"
tests/btrfs/284.out +5 -3
index 4a69f82..c942271 100644
@@ -1,10 +1,12 @@
 QA output created by 284
 defrag object | defragment range | defragment compress
-a single file |  default | off
+a single file | default | off
 a single file | default |  on
-a single file | start < 0 && 0 < len < file size | off
+a single file | start < 0 && 0 < len < file size | off (should fail)
+btrfs filesystem defragment failed!
 a single file | start > file size && 0 < len < file size | off
-a single file | start = 0 && len < 0 | off
+a single file | start = 0 && len < 0 | off (should fail)
+btrfs filesystem defragment failed!
 a single file | start = 0 && len > file size | off
 a single file | start = 0 && 0 < len < file size | off
 a directory | default | off
#1 lore raw
Quoted 22 lines · Eric Sandeen

test 284 had... some issues.

First, it took so long nobody ran it; so shorten the extent count by a factor of about 100.

Having fixed that, we see failures in 2 cases; when start or len is -1, but the golden output file didn't have error output, as if they should pass.

I'm going to argue that these *should* both fail; start = -1 has no real meaning. length = -1 might mean "the rest of the file" but if that's what you really want, just don't specify -l.

So add failure output for those cases.

Send all command output to $seq.full, in case that changes in the future; just capture the return value.

Then remove the return value echo on failure (50?) because who knows when that might change to some other magic value.

Quoted 7 lines

Ok, then when defrag actually works, old defrag returned "20" (because?) but a recent commit changed it to 0. So accommodate that too.

And remove a stray "HAVE_DEFRAG=1" while we're at it. That variable is never used.

So should I be seeing failures with btrfs-progs-0.20-0.2.git91d9eec.el6.x86_64 installed?

./check btrfs/284

FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 cxfsxe4 3.9.0+
MKFS_OPTIONS  -- /dev/sdk2

MOUNT_OPTIONS -- /dev/sdk2 /mnt/scratch

btrfs/284 - output mismatch (see /usr/src/rcj/xfstests/results/btrfs/284.out.bad)

    --- tests/btrfs/284.out	2013-05-14 09:31:35.000000000 -0500
    +++ /usr/src/rcj/xfstests/results/btrfs/284.out.bad	2013-05-14 
10:10:45.000000000 -0500
    @@ -6,7 +6,6 @@
     btrfs filesystem defragment failed!
     a single file | start > file size && 0 < len < file size | off
     a single file | start = 0 && len < 0 | off (should fail)
    -btrfs filesystem defragment failed!
     a single file | start = 0 && len > file size | off
     a single file | start = 0 && 0 < len < file size | off
     a directory | default | off
     ...
     (Run 'diff -u tests/btrfs/284.out 

/usr/src/rcj/xfstests/results/btrfs/284.out.bad' to see the entire diff)
Ran: btrfs/284
Failures: btrfs/284
Failed 1 of 1 tests

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#2 lore raw
Quoted 27 lines · Rich Johnston
Quoted 26 lines · Eric Sandeen

test 284 had... some issues.

First, it took so long nobody ran it; so shorten the extent count by a factor of about 100.

Having fixed that, we see failures in 2 cases; when start or len is -1, but the golden output file didn't have error output, as if they should pass.

I'm going to argue that these *should* both fail; start = -1 has no real meaning. length = -1 might mean "the rest of the file" but if that's what you really want, just don't specify -l.

So add failure output for those cases.

Send all command output to $seq.full, in case that changes in the future; just capture the return value.

Then remove the return value echo on failure (50?) because who knows when that might change to some other magic value.

Ok, then when defrag actually works, old defrag returned "20" (because?) but a recent commit changed it to 0. So accommodate that too.

Quoted 6 lines
Quoted 3 lines

And remove a stray "HAVE_DEFRAG=1" while we're at it. That variable is never used.

So should I be seeing failures with btrfs-progs-0.20-0.2.git91d9eec.el6.x86_64 installed?

Maybe? ...if that's an old version in rhel6.

If you really want to investigate this, you could grab i.e.
http://kojipkgs.fedoraproject.org/packages/btrfs-progs/0.20.rc1.20130501git7854c8b/3.fc20/src/btrfs-progs-0.20.rc1.20130501git7854c8b-3.fc20.src.rpm
and rebuild it for something newer. (Or I could double check . . . )

But honestly as the sgi xfstests maintainer I think you are going well above and beyond your duties here.

Ideally, someone from the btrfs community could help out here, and test/review the change...

-Eric

Quoted 23 lines

./check btrfs/284

FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 cxfsxe4 3.9.0+
MKFS_OPTIONS  -- /dev/sdk2

MOUNT_OPTIONS -- /dev/sdk2 /mnt/scratch

btrfs/284 - output mismatch (see /usr/src/rcj/xfstests/results/btrfs/284.out.bad)

    --- tests/btrfs/284.out    2013-05-14 09:31:35.000000000 -0500
    +++ /usr/src/rcj/xfstests/results/btrfs/284.out.bad    2013-05-14 10:10:45.000000000 -0500
    @@ -6,7 +6,6 @@
     btrfs filesystem defragment failed!
     a single file | start > file size && 0 < len < file size | off
     a single file | start = 0 && len < 0 | off (should fail)
    -btrfs filesystem defragment failed!
     a single file | start = 0 && len > file size | off
     a single file | start = 0 && 0 < len < file size | off
     a directory | default | off
     ...
     (Run 'diff -u tests/btrfs/284.out /usr/src/rcj/xfstests/results/btrfs/284.out.bad' to see the entire diff)

Ran: btrfs/284
Failures: btrfs/284
Failed 1 of 1 tests

Reviewed-by
#3 lore raw
In reply to Rich Johnston
Quoted 55 lines · Rich Johnston
Quoted 22 lines · Eric Sandeen

test 284 had... some issues.

First, it took so long nobody ran it; so shorten the extent count by a factor of about 100.

Having fixed that, we see failures in 2 cases; when start or len is -1, but the golden output file didn't have error output, as if they should pass.

I'm going to argue that these *should* both fail; start = -1 has no real meaning. length = -1 might mean "the rest of the file" but if that's what you really want, just don't specify -l.

So add failure output for those cases.

Send all command output to $seq.full, in case that changes in the future; just capture the return value.

Then remove the return value echo on failure (50?) because who knows when that might change to some other magic value.

Quoted 7 lines

Ok, then when defrag actually works, old defrag returned "20" (because?) but a recent commit changed it to 0. So accommodate that too.

And remove a stray "HAVE_DEFRAG=1" while we're at it. That variable is never used.

So should I be seeing failures with btrfs-progs-0.20-0.2.git91d9eec.el6.x86_64 installed?

./check btrfs/284

FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 cxfsxe4 3.9.0+
MKFS_OPTIONS  -- /dev/sdk2

MOUNT_OPTIONS -- /dev/sdk2 /mnt/scratch

btrfs/284 - output mismatch (see
/usr/src/rcj/xfstests/results/btrfs/284.out.bad)

     --- tests/btrfs/284.out	2013-05-14 09:31:35.000000000 -0500
     +++ /usr/src/rcj/xfstests/results/btrfs/284.out.bad	2013-05-14 
10:10:45.000000000 -0500
     @@ -6,7 +6,6 @@
      btrfs filesystem defragment failed!
      a single file | start > file size && 0 < len < file size | off
      a single file | start = 0 && len < 0 | off (should fail)
     -btrfs filesystem defragment failed!
      a single file | start = 0 && len > file size | off
      a single file | start = 0 && 0 < len < file size | off
      a directory | default | off
      ...
      (Run 'diff -u tests/btrfs/284.out 
Quoted 5 lines

/usr/src/rcj/xfstests/results/btrfs/284.out.bad' to see the entire diff)
Ran: btrfs/284
Failures: btrfs/284
Failed 1 of 1 tests

Yeah defrag used to spit out a return value on success, that has been fixed recently. This patch looks good, I just ran it on my box and it ran fine, you can add

Thanks,

Josef

Reviewed-by on v1 1/1
#4 lore raw
In reply to Eric Sandeen
Quoted 22 lines · Eric Sandeen

test 284 had... some issues.

First, it took so long nobody ran it; so shorten the extent count by a factor of about 100.

Having fixed that, we see failures in 2 cases; when start or len is -1, but the golden output file didn't have error output, as if they should pass.

I'm going to argue that these *should* both fail; start = -1 has no real meaning. length = -1 might mean "the rest of the file" but if that's what you really want, just don't specify -l.

So add failure output for those cases.

Send all command output to $seq.full, in case that changes in the future; just capture the return value.

Then remove the return value echo on failure (50?) because who knows when that might change to some other magic value.

Quoted 6 lines

Ok, then when defrag actually works, old defrag returned "20" (because?) but a recent commit changed it to 0. So accommodate that too.

And remove a stray "HAVE_DEFRAG=1" while we're at it. That variable is never used.

Thanks for working on this, Eric!

You can add

thanks,
liubo

Quoted 111 lines

---

tests/btrfs/284 +14 -11
old mode 100644
new mode 100755
index d952977..67161a3
@@ -26,7 +26,7 @@ seqres=$RESULT_DIR/$seq
 echo "QA output created by $seq"
 here="`pwd`"
 tmp=/tmp/$$
-cnt=11999
+cnt=119
 filesize=48000
 
 status=1	# failure is the default!
@@ -58,11 +58,12 @@ _create_file()
 _btrfs_online_defrag()
 {
 	str=""
+	# start = -1 is invalid, should fail
 	if [ "$2" = "2" ];then
 		str="$str -s -1 -l $((filesize / 2)) "
 	elif [ "$2" = "3" ];then
 		str="$str -s $((filesize + 1)) -l $((filesize / 2)) "
-		HAVE_DEFRAG=1
+	# len = -1 is invalid, should fail
 	elif [ "$2" = "4" ];then
 		str="$str -l -1 "
 	elif [ "$2" = "5" ];then
@@ -76,20 +77,22 @@ _btrfs_online_defrag()
 	fi
 
 	if [ "$str" != "" ]; then
-		$BTRFS_UTIL_PROG filesystem defragment $str $SCRATCH_MNT/tmp_file
+		$BTRFS_UTIL_PROG filesystem defragment $str $SCRATCH_MNT/tmp_file >> $seq.full 2>&1
 	else
 		if [ "$1" = "1" ];then
-			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT/tmp_file
+			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT/tmp_file >> $seq.full 2>&1
 		elif [ "$1" = "2" ];then
-			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT/tmp_dir
+			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT/tmp_dir >> $seq.full 2>&1
 		elif [ "$1" = "3" ];then
-			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT
+			$BTRFS_UTIL_PROG filesystem defragment $SCRATCH_MNT >> $seq.full 2>&1
 		fi
 	fi
 	ret_val=$?
 	_scratch_remount
-	if [ $ret_val -ne 20 ];then
-		echo "btrfs filesystem defragment failed! err is $ret_val"
+	# Older defrag returned "20" for success
+	# e9393c2 btrfs-progs: defrag return zero on success
+	if [ $ret_val -ne 0 -a $ret_val -ne 20 ]; then
+		echo "btrfs filesystem defragment failed!"
 	fi
 }
 
@@ -140,19 +143,19 @@ _scratch_mount
 _require_defrag
 
 echo "defrag object | defragment range | defragment compress"
-echo "a single file |  default | off"
+echo "a single file | default | off"
 _rundefrag 1 1 1
 
 echo "a single file | default |  on"
 _rundefrag 1 1 2
 
-echo "a single file | start < 0 && 0 < len < file size | off"
+echo "a single file | start < 0 && 0 < len < file size | off (should fail)"
 _rundefrag 1 2 1
 
 echo "a single file | start > file size && 0 < len < file size | off"
 _rundefrag 1 3 1
 
-echo "a single file | start = 0 && len < 0 | off"
+echo "a single file | start = 0 && len < 0 | off (should fail)"
 _rundefrag 1 4 1
 
 echo "a single file | start = 0 && len > file size | off"
tests/btrfs/284.out +5 -4
index 4a69f82..c942271 100644
@@ -1,10 +1,12 @@
 QA output created by 284
 defrag object | defragment range | defragment compress
-a single file |  default | off
+a single file | default | off
 a single file | default |  on
-a single file | start < 0 && 0 < len < file size | off
+a single file | start < 0 && 0 < len < file size | off (should fail)
+btrfs filesystem defragment failed!
 a single file | start > file size && 0 < len < file size | off
-a single file | start = 0 && len < 0 | off
+a single file | start = 0 && len < 0 | off (should fail)
+btrfs filesystem defragment failed!
 a single file | start = 0 && len > file size | off
 a single file | start = 0 && 0 < len < file size | off
 a directory | default | off
 
 
--

To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Reviewed-by Josef Bacik Reviewed-by Liu Bo on v1 1/1
#5 lore raw
In reply to Eric Sandeen

tests/btrfs/284

Quoted 11 lines · Eric Sandeen

test 284 had... some issues.

tests/btrfs/284 +0 -0
old mode 100644
new mode 100755
index d952977..67161a3

This patch has been committed:

commit 91f87e3b89c0f7350a56d397ba7255d0e3d6b486

Author: Eric Sandeen <[email protected]>
Date:   Wed May 15 07:18:07 2013 -0500

xfstests btrfs/284: shorten duration, fix output

Thanks
--Rich