Re: [Qemu-devel] [PATCH v6 20/21] iotests: add incremental backup failure recovery test

[John Snow <jsnow@redhat.com>]

On 11/27/2015 12:14 PM, Kevin Wolf wrote:
> Am 18.04.2015 um 01:50 hat John Snow geschrieben:
>> Test the failure case for incremental backups.
>>
>> Signed-off-by: John Snow <jsnow@redhat.com>
>> Reviewed-by: Max Reitz <mreitz@redhat.com>
>> ---
>>  tests/qemu-iotests/124     | 57 ++++++++++++++++++++++++++++++++++++++++++++++
>>  tests/qemu-iotests/124.out |  4 ++--
>>  2 files changed, 59 insertions(+), 2 deletions(-)
>>
>> diff --git a/tests/qemu-iotests/124 b/tests/qemu-iotests/124
>> index 5c3b434..95f6de5 100644
>> --- a/tests/qemu-iotests/124
>> +++ b/tests/qemu-iotests/124
>> @@ -240,6 +240,63 @@ class TestIncrementalBackup(iotests.QMPTestCase):
>>          self.check_backups()
>>  
>>  
>> +    def test_incremental_failure(self):
>> +        '''Test: Verify backups made after a failure are correct.
>> +
>> +        Simulate a failure during an incremental backup block job,
>> +        emulate additional writes, then create another incremental backup
>> +        afterwards and verify that the backup created is correct.
>> +        '''
>> +
>> +        # Create a blkdebug interface to this img as 'drive1',
>> +        # but don't actually create a new image.
>> +        drive1 = self.add_node('drive1', self.drives[0]['fmt'],
>> +                               path=self.drives[0]['file'],
>> +                               backup=self.drives[0]['backup'])
>> +        result = self.vm.qmp('blockdev-add', options={
>> +            'id': drive1['id'],
>> +            'driver': drive1['fmt'],
>> +            'file': {
>> +                'driver': 'blkdebug',
>> +                'image': {
>> +                    'driver': 'file',
>> +                    'filename': drive1['file']
>> +                },
>> +                'set-state': [{
>> +                    'event': 'flush_to_disk',
>> +                    'state': 1,
>> +                    'new_state': 2
>> +                }],
>> +                'inject-error': [{
>> +                    'event': 'read_aio',
>> +                    'errno': 5,
>> +                    'state': 2,
>> +                    'immediately': False,
>> +                    'once': True
>> +                }],
>> +            }
>> +        })
>> +        self.assert_qmp(result, 'return', {})
> 
> John, how naughty of you!
> 

And here I thought it was OK because nobody yelled!

The yell was just delayed.

> It's interesting how many tests break now that I tried to add some
> advisory qcow2 locking so that people don't constantly break their
> images with 'qemu-img snapshot' while the VM is running.
> 
> I think this one is a bug in the test case. I'm not completely sure how
> to fix it, though. Can we move the blkdebug layer to the top level? I
> think reusing the same qcow2 BDS (using a node name reference) would be
> okay. We just need to avoid opening the qcow2 layer twice for the same
> image.
> 

I can either do that, or just fall back to fully allocating two images
and modify the test accordingly.

> Kevin
> 

Is this for 2.5?

--js