On Tue, Aug 29, 2017 at 03:19:54PM +0100, Stefan Hajnoczi wrote: > On Tue, Aug 29, 2017 at 11:06:22AM +0200, Kashyap Chamarthy wrote:
[...] > > +iotests.log('Gracefully ending the `drive-mirror` job on source...') > > +iotests.log(source_vm.qmp('block-job-cancel', device='mirror-job0')) > > Two issues: > > 1. Migration may not have completed yet so drive-mirror cannot be > stopped here. We need to wait for migration to complete on the > source before cancelling the block job. > > 2. block-job-cancel is asynchronous. The block job may still be running > after this returns. Therefore it may still be using the NBD drive... > > > + > > +iotests.log('Stopping the NBD server on destination...') > > +iotests.log(dest_vm.qmp('nbd-server-stop')) > > ...and that races with this command on the destination. > > We need to wait for the BLOCK_JOB_CANCELLED/COMPLETED event on the > source QEMU before stopping the NBD server on the destination. Yes, both good points. How about the following `diff` to address both the issues you raised? #---------------------------------------------------------------------- diff --git a/tests/qemu-iotests/194 b/tests/qemu-iotests/194 index 9d81189..a3e3bad 100755 --- a/tests/qemu-iotests/194 +++ b/tests/qemu-iotests/194 @@ -67,20 +67,18 @@ dest_vm.qmp('migrate-set-capabilities', capabilities=[{'capability': 'events', 'state': True}]) iotests.log(source_vm.qmp('migrate', uri='unix:{0}'.format(migration_sock_path))) -iotests.log('Gracefully ending the `drive-mirror` job on source...') -iotests.log(source_vm.qmp('block-job-cancel', device='mirror-job0')) - -iotests.log('Stopping the NBD server on destination...') -iotests.log(dest_vm.qmp('nbd-server-stop')) - while True: event1 = source_vm.event_wait('MIGRATION') iotests.log(event1, filters=[iotests.filter_qmp_event]) if event1['data']['status'] in ('completed', 'failed'): + iotests.log('Gracefully ending the `drive-mirror` job on source...') + iotests.log(source_vm.qmp('block-job-cancel', device='mirror-job0')) break while True: event2 = source_vm.event_wait('BLOCK_JOB_COMPLETED') iotests.log(event2, filters=[iotests.filter_qmp_event]) if event2['event'] == 'BLOCK_JOB_COMPLETED': + iotests.log('Stopping the NBD server on destination...') + iotests.log(dest_vm.qmp('nbd-server-stop')) break #---------------------------------------------------------------------- And the output would be: #---------------------------------------------------------------------- Launching VMs... Launching NBD server on destination... {u'return': {}} {u'return': {}} Starting `drive-mirror` on source... {u'return': {}} Waiting for `drive-mirror` to complete... {u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'device': u'mirror-job0', u'type': u'mirror', u'speed': 0, u'len': 1073741824, u'offset': 1073741824}, u'event': u'BLOCK_JOB_READY'} Starting migration... {u'return': {}} {u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'setup'}, u'event': u'MIGRATION'} {u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'active'}, u'event': u'MIGRATION'} {u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'completed'}, u'event': u'MIGRATION'} Gracefully ending the `drive-mirror` job on source... {u'return': {}} {u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'device': u'mirror-job0', u'type': u'mirror', u'speed': 0, u'len': 1073741824, u'offset': 1073741824}, u'event': u'BLOCK_JOB_COMPLETED'} Stopping the NBD server on destination... {u'return': {}} #---------------------------------------------------------------------- -- /kashyap