[1/3] generic/459: use xfs_freeze instead of fsfreeze

Message ID 1526276991-23897-2-git-send-email-amir73il@gmail.com
State Not Applicable, archived
Headers show
Series
  • ext4 freeze regression test
Related show

Commit Message

Amir Goldstein May 14, 2018, 5:49 a.m.
This is what all other tests use as well as _require_freeze

Signed-off-by: Amir Goldstein <amir73il@gmail.com>
---
 tests/generic/459 | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

Comments

Dave Chinner May 14, 2018, 11:01 p.m. | #1
On Mon, May 14, 2018 at 08:49:49AM +0300, Amir Goldstein wrote:
> This is what all other tests use as well as _require_freeze
> 
> Signed-off-by: Amir Goldstein <amir73il@gmail.com>
> ---
>  tests/generic/459 | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/tests/generic/459 b/tests/generic/459
> index e755cf7..22b8e62 100755
> --- a/tests/generic/459
> +++ b/tests/generic/459
> @@ -113,7 +113,7 @@ $XFS_IO_PROG -f -d -c 'pwrite -b 1m 0 120m' $SCRATCH_MNT/f1 >>$seqres.full 2>&1
>  # extended. It is expected, and is only used so xfsaild is triggered to
>  # flush AIL items, other filesystems usually get remounted as read-only during
>  # the above write process.
> -fsfreeze -f $SCRATCH_MNT >>$seqres.full 2>&1 &
> +xfs_freeze -f $SCRATCH_MNT >>$seqres.full 2>&1 &
>  freezeid=$!
>  
>  # Wait enough so xfsaild can run
> @@ -132,7 +132,7 @@ ret=$?
>  #	- The filesystem stays in Read-Write mode, but can be frozen/thawed
>  #	  without getting stuck.
>  if [ $ret -ne 0 ]; then
> -	# fsfreeze failed, filesystem should reject further writes and remount
> +	# freeze failed, filesystem should reject further writes and remount
>  	# as readonly. Sometimes the previous write process won't trigger
>  	# ro-remount, e.g. on ext3/4, do additional touch here to make sure
>  	# filesystems see the metadata I/O error.
> @@ -148,7 +148,7 @@ if [ $ret -ne 0 ]; then
>  else
>  	# Try to thaw the filesystem, and complete test if if succeed.
>  	# NOTE: This will hang on affected XFS filesystems.
> -	fsfreeze -u $SCRATCH_MNT >>$seqres.full 2>&1
> +	xfs_freeze -u $SCRATCH_MNT >>$seqres.full 2>&1
>  	echo "Test OK"
>  fi

Shouldn't these all be converted to XFS_FREEZE_PROG w/ a setup call
to export XFS_FREEZE_PROG="`set_prog_path xfs_freeze`"?

Cheers,

Dave.
Amir Goldstein May 15, 2018, 5:25 a.m. | #2
On Tue, May 15, 2018 at 2:01 AM, Dave Chinner <david@fromorbit.com> wrote:
> On Mon, May 14, 2018 at 08:49:49AM +0300, Amir Goldstein wrote:
>> This is what all other tests use as well as _require_freeze
>>
>> Signed-off-by: Amir Goldstein <amir73il@gmail.com>
>> ---
>>  tests/generic/459 | 6 +++---
>>  1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/tests/generic/459 b/tests/generic/459
>> index e755cf7..22b8e62 100755
>> --- a/tests/generic/459
>> +++ b/tests/generic/459
>> @@ -113,7 +113,7 @@ $XFS_IO_PROG -f -d -c 'pwrite -b 1m 0 120m' $SCRATCH_MNT/f1 >>$seqres.full 2>&1
>>  # extended. It is expected, and is only used so xfsaild is triggered to
>>  # flush AIL items, other filesystems usually get remounted as read-only during
>>  # the above write process.
>> -fsfreeze -f $SCRATCH_MNT >>$seqres.full 2>&1 &
>> +xfs_freeze -f $SCRATCH_MNT >>$seqres.full 2>&1 &
>>  freezeid=$!
>>
>>  # Wait enough so xfsaild can run
>> @@ -132,7 +132,7 @@ ret=$?
>>  #    - The filesystem stays in Read-Write mode, but can be frozen/thawed
>>  #      without getting stuck.
>>  if [ $ret -ne 0 ]; then
>> -     # fsfreeze failed, filesystem should reject further writes and remount
>> +     # freeze failed, filesystem should reject further writes and remount
>>       # as readonly. Sometimes the previous write process won't trigger
>>       # ro-remount, e.g. on ext3/4, do additional touch here to make sure
>>       # filesystems see the metadata I/O error.
>> @@ -148,7 +148,7 @@ if [ $ret -ne 0 ]; then
>>  else
>>       # Try to thaw the filesystem, and complete test if if succeed.
>>       # NOTE: This will hang on affected XFS filesystems.
>> -     fsfreeze -u $SCRATCH_MNT >>$seqres.full 2>&1
>> +     xfs_freeze -u $SCRATCH_MNT >>$seqres.full 2>&1
>>       echo "Test OK"
>>  fi
>
> Shouldn't these all be converted to XFS_FREEZE_PROG w/ a setup call
> to export XFS_FREEZE_PROG="`set_prog_path xfs_freeze`"?
>

Yeh, I suppose I can take this one as part of the "freeze" cleanup.

Thanks,
Amir.
Amir Goldstein May 15, 2018, 1:50 p.m. | #3
On Tue, May 15, 2018 at 8:25 AM, Amir Goldstein <amir73il@gmail.com> wrote:
> On Tue, May 15, 2018 at 2:01 AM, Dave Chinner <david@fromorbit.com> wrote:
>> On Mon, May 14, 2018 at 08:49:49AM +0300, Amir Goldstein wrote:
>>> This is what all other tests use as well as _require_freeze
>>>
>>> Signed-off-by: Amir Goldstein <amir73il@gmail.com>
>>> ---
[...]
>>
>> Shouldn't these all be converted to XFS_FREEZE_PROG w/ a setup call
>> to export XFS_FREEZE_PROG="`set_prog_path xfs_freeze`"?
>>
>
> Yeh, I suppose I can take this one as part of the "freeze" cleanup.
>

I looked again and xfs_freeze pattern repeats itself quite often with
other xfs_* progs, so I guess I'll leave that cleanup for another time...

Thanks,
Amir.

$ git grep -v '^[[:space:]]*[#;]' tests/*/???| grep -v echo > all
$ (cd /usr/sbin; ls xfs_*) > xfsprogs
$ for p in `cat xfsprogs`; do echo $p; grep -w $p all | wc -l; done
xfs_admin
1
xfs_bmap
24
xfs_copy
9
xfs_db
35
xfs_estimate
0
xfs_freeze
17
xfs_fsr
15
xfs_growfs
3
xfs_info
31
xfs_io
17
xfs_logprint
0
xfs_mdrestore
7
xfs_metadump
1
xfs_mkfile
1
xfs_ncheck
0
xfs_quota
43
xfs_rtcp
0
xfs_spaceman
0
Dave Chinner May 16, 2018, 12:47 a.m. | #4
On Tue, May 15, 2018 at 04:50:04PM +0300, Amir Goldstein wrote:
> On Tue, May 15, 2018 at 8:25 AM, Amir Goldstein <amir73il@gmail.com> wrote:
> > On Tue, May 15, 2018 at 2:01 AM, Dave Chinner <david@fromorbit.com> wrote:
> >> On Mon, May 14, 2018 at 08:49:49AM +0300, Amir Goldstein wrote:
> >>> This is what all other tests use as well as _require_freeze
> >>>
> >>> Signed-off-by: Amir Goldstein <amir73il@gmail.com>
> >>> ---
> [...]
> >>
> >> Shouldn't these all be converted to XFS_FREEZE_PROG w/ a setup call
> >> to export XFS_FREEZE_PROG="`set_prog_path xfs_freeze`"?
> >>
> >
> > Yeh, I suppose I can take this one as part of the "freeze" cleanup.
> >
> 
> I looked again and xfs_freeze pattern repeats itself quite often with
> other xfs_* progs, so I guess I'll leave that cleanup for another time...

No worries, just add it to the pile...

Cheers,

Dave.
Eryu Guan May 17, 2018, 2:36 a.m. | #5
On Mon, May 14, 2018 at 08:49:49AM +0300, Amir Goldstein wrote:
> This is what all other tests use as well as _require_freeze
> 
> Signed-off-by: Amir Goldstein <amir73il@gmail.com>

I'll drop this patch for now per discussion but take the other two
patches, the XFS_FREEZE_PROG conversion is not urgent and could wait I
think :)

Thanks,
Eryu

> ---
>  tests/generic/459 | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/tests/generic/459 b/tests/generic/459
> index e755cf7..22b8e62 100755
> --- a/tests/generic/459
> +++ b/tests/generic/459
> @@ -113,7 +113,7 @@ $XFS_IO_PROG -f -d -c 'pwrite -b 1m 0 120m' $SCRATCH_MNT/f1 >>$seqres.full 2>&1
>  # extended. It is expected, and is only used so xfsaild is triggered to
>  # flush AIL items, other filesystems usually get remounted as read-only during
>  # the above write process.
> -fsfreeze -f $SCRATCH_MNT >>$seqres.full 2>&1 &
> +xfs_freeze -f $SCRATCH_MNT >>$seqres.full 2>&1 &
>  freezeid=$!
>  
>  # Wait enough so xfsaild can run
> @@ -132,7 +132,7 @@ ret=$?
>  #	- The filesystem stays in Read-Write mode, but can be frozen/thawed
>  #	  without getting stuck.
>  if [ $ret -ne 0 ]; then
> -	# fsfreeze failed, filesystem should reject further writes and remount
> +	# freeze failed, filesystem should reject further writes and remount
>  	# as readonly. Sometimes the previous write process won't trigger
>  	# ro-remount, e.g. on ext3/4, do additional touch here to make sure
>  	# filesystems see the metadata I/O error.
> @@ -148,7 +148,7 @@ if [ $ret -ne 0 ]; then
>  else
>  	# Try to thaw the filesystem, and complete test if if succeed.
>  	# NOTE: This will hang on affected XFS filesystems.
> -	fsfreeze -u $SCRATCH_MNT >>$seqres.full 2>&1
> +	xfs_freeze -u $SCRATCH_MNT >>$seqres.full 2>&1
>  	echo "Test OK"
>  fi
>  
> -- 
> 2.7.4
>

Patch

diff --git a/tests/generic/459 b/tests/generic/459
index e755cf7..22b8e62 100755
--- a/tests/generic/459
+++ b/tests/generic/459
@@ -113,7 +113,7 @@  $XFS_IO_PROG -f -d -c 'pwrite -b 1m 0 120m' $SCRATCH_MNT/f1 >>$seqres.full 2>&1
 # extended. It is expected, and is only used so xfsaild is triggered to
 # flush AIL items, other filesystems usually get remounted as read-only during
 # the above write process.
-fsfreeze -f $SCRATCH_MNT >>$seqres.full 2>&1 &
+xfs_freeze -f $SCRATCH_MNT >>$seqres.full 2>&1 &
 freezeid=$!
 
 # Wait enough so xfsaild can run
@@ -132,7 +132,7 @@  ret=$?
 #	- The filesystem stays in Read-Write mode, but can be frozen/thawed
 #	  without getting stuck.
 if [ $ret -ne 0 ]; then
-	# fsfreeze failed, filesystem should reject further writes and remount
+	# freeze failed, filesystem should reject further writes and remount
 	# as readonly. Sometimes the previous write process won't trigger
 	# ro-remount, e.g. on ext3/4, do additional touch here to make sure
 	# filesystems see the metadata I/O error.
@@ -148,7 +148,7 @@  if [ $ret -ne 0 ]; then
 else
 	# Try to thaw the filesystem, and complete test if if succeed.
 	# NOTE: This will hang on affected XFS filesystems.
-	fsfreeze -u $SCRATCH_MNT >>$seqres.full 2>&1
+	xfs_freeze -u $SCRATCH_MNT >>$seqres.full 2>&1
 	echo "Test OK"
 fi