delete a updating stack may cause that some resources can't be deleted

[Zhang Yang]

if delete a updating stack before the new template ID is updated,
the new resources added in stack won't be deleted, I think
we need save the new template earlier.

I have a autoscalinggroup, and it have a scaleout policy which add 2 VMs
everytime. and if you delete the stack which is scaling, the added VMs won't be
deleted.

heat=> select resource_name, resource_action,resource_status,resource_type,created_at from event where stack_id='3ea4e2d1-4506-459a-a582-adc7708cab07' order by created_at;
  resource_name | resource_action | resource_status | resource_type | created_at
------------------+-----------------+-----------------+--------------------+----------------------------
 WebServerGroup-0 | CREATE | IN_PROGRESS | AWS::EC2::Instance | 2014-04-04 11:27:49.916327
 WebServerGroup-0 | CREATE | COMPLETE | AWS::EC2::Instance | 2014-04-04 11:27:57.586147
 WebServerGroup-1 | CREATE | IN_PROGRESS | AWS::EC2::Instance | 2014-04-04 11:28:34.832987
 WebServerGroup-1 | CREATE | COMPLETE | AWS::EC2::Instance | 2014-04-04 11:28:42.827186
 WebServerGroup-2 | CREATE | IN_PROGRESS | AWS::EC2::Instance | 2014-04-04 11:29:34.886439
 WebServerGroup-3 | CREATE | IN_PROGRESS | AWS::EC2::Instance | 2014-04-04 11:29:36.098389
 WebServerGroup-1 | DELETE | IN_PROGRESS | AWS::EC2::Instance | 2014-04-04 11:29:43.906241
 WebServerGroup-0 | DELETE | IN_PROGRESS | AWS::EC2::Instance | 2014-04-04 11:29:44.523359
 WebServerGroup-2 | CREATE | COMPLETE | AWS::EC2::Instance | 2014-04-04 11:29:44.638489
 WebServerGroup-3 | CREATE | COMPLETE | AWS::EC2::Instance | 2014-04-04 11:29:44.90781
 WebServerGroup-1 | DELETE | COMPLETE | AWS::EC2::Instance | 2014-04-04 11:29:47.702711
 WebServerGroup-0 | DELETE | COMPLETE | AWS::EC2::Instance | 2014-04-04 11:29:50.058942
(12 rows)

heat=> select resource_name, resource_action,resource_status,resource_type,created_at from event where stack_id='3ce80dd6-aba2-4108-b142-7d68fe63df21' order by created_at;
     resource_name | resource_action | resource_status | resource_type | created_at
------------------------+-----------------+-----------------+---------------------------------------+----------------------------
 LaunchConfig | CREATE | IN_PROGRESS | AWS::AutoScaling::LaunchConfiguration | 2014-04-04 11:27:49.408229
 LaunchConfig | CREATE | COMPLETE | AWS::AutoScaling::LaunchConfiguration | 2014-04-04 11:27:49.448942
 WebServerGroup | CREATE | IN_PROGRESS | AWS::AutoScaling::AutoScalingGroup | 2014-04-04 11:27:49.482674
 WebServerGroup | CREATE | COMPLETE | AWS::AutoScaling::AutoScalingGroup | 2014-04-04 11:27:57.63526
 WebServerScaleUpPolicy | CREATE | IN_PROGRESS | AWS::AutoScaling::ScalingPolicy | 2014-04-04 11:27:57.667967
 WebServerScaleUpPolicy | CREATE | COMPLETE | AWS::AutoScaling::ScalingPolicy | 2014-04-04 11:27:58.88479
 CPUAlarmHigh | CREATE | IN_PROGRESS | OS::Ceilometer::Alarm | 2014-04-04 11:27:58.951199
 CPUAlarmHigh | CREATE | COMPLETE | OS::Ceilometer::Alarm | 2014-04-04 11:28:00.052574
 WebServerScaleUpPolicy | signal | COMPLETE | AWS::AutoScaling::ScalingPolicy | 2014-04-04 11:28:34.259004
 WebServerScaleUpPolicy | signal | COMPLETE | AWS::AutoScaling::ScalingPolicy | 2014-04-04 11:29:34.035764
 CPUAlarmHigh | DELETE | IN_PROGRESS | OS::Ceilometer::Alarm | 2014-04-04 11:29:42.424311
 CPUAlarmHigh | DELETE | COMPLETE | OS::Ceilometer::Alarm | 2014-04-04 11:29:42.490193
 WebServerScaleUpPolicy | DELETE | IN_PROGRESS | AWS::AutoScaling::ScalingPolicy | 2014-04-04 11:29:42.540389
 WebServerScaleUpPolicy | DELETE | COMPLETE | AWS::AutoScaling::ScalingPolicy | 2014-04-04 11:29:43.692144
 WebServerGroup | DELETE | IN_PROGRESS | AWS::AutoScaling::AutoScalingGroup | 2014-04-04 11:29:43.74047
 WebServerGroup | DELETE | COMPLETE | AWS::AutoScaling::AutoScalingGroup | 2014-04-04 11:29:51.208128
 LaunchConfig | DELETE | IN_PROGRESS | AWS::AutoScaling::LaunchConfiguration | 2014-04-04 11:29:51.274372
 LaunchConfig | DELETE | COMPLETE | AWS::AutoScaling::LaunchConfiguration | 2014-04-04 11:29:51.32382
(18 rows)

[Sergey Kraynev (skraynev)]

Could you please upload example for reproducing this problem.

[Zhang Yang (neil-zhangyang)]

the reason is that we won't save the new template until the stack update finish, so if delete a updating stack, the loaded stack will use the old template which don't have the new resources.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/86043

[Zane Bitter (zaneb)]

The plan is for https://blueprints.launchpad.net/heat/+spec/update-failure-recovery to fix this issue in Juno.

[Angus Salkeld (asalkeld)]

resetting state as no activity for a year.