stack update failed due to AttributeError: 'NoneType' object has no attribute 'properties'

[XiaoLiang Hu]

I try same stack update cases on different resource type , it is failed for below cases:

Note:
1. all the cases base on YAML format heat template
2. all the case pass the stack create , but failed on stack update.
3. all the exception trace back log in engine.log file show same reason for the failed cases.

case1 :
1) create a stack named "stack_test" with only one resource type "OS::Heat::InstanceGroup" named "xx" in template >> success
2) update this template, add a new resource type "OS::Heat::InstanceGroup" named "yy" in template
3) run "stack-update -f template_file stack_test" >>>> failed

I think the expected result should be an new resource "yy" create success and the nova instances which belong to this resource create success as well.

case2 :
1) create a stack named "stack_test" with only one resource type "OS::Heat::InstanceGroup" named "xx" in template >> success
2) update this template, change the value of "Size", such as 1 --- > 2
3) run "stack-update -f template_file stack_test" >>>> failed

I think the expected result should be one nova instance which belong to this resource create success.

case3:
1) create a stack named "stack_test" with resource type "AWS::AutoScaling::AutoScalingGroup" named "my_asg" in template as below >> success

[root@xhu01 heat_action_test]# more autoscalinggroup_aws.yaml
heat_template_version: 2013-05-23

resources:

  my_config:
    type: AWS::AutoScaling::LaunchConfiguration
    properties:
      ImageId: xx
      InstanceType: xhu.test

  my_asg:
    type: AWS::AutoScaling::AutoScalingGroup
    properties:
      AvailabilityZones: [nova]
      MinSize: 1
      MaxSize: 3
      DesiredCapacity: 1
      LaunchConfigurationName: my_config
[root@xhu01 heat_action_test]

2) update this template, add new resource "OS::Heat::InstanceGroup" named "your_asg" in template

[root@xhu01 heat_action_test]# more autoscalinggroup_aws.yaml
heat_template_version: 2013-05-23

resources:

  my_config:
    type: AWS::AutoScaling::LaunchConfiguration
    properties:
      ImageId: xx
      InstanceType: xhu.test

  my_asg:
    type: AWS::AutoScaling::AutoScalingGroup
    properties:
      AvailabilityZones: [nova]
      MinSize: 2
      MaxSize: 3
      DesiredCapacity: 3
      LaunchConfigurationName: my_config

  your_config:
    type: AWS::AutoScaling::LaunchConfiguration
    properties:
      ImageId: xx
      InstanceType: xhu.test

  your_asg:
    type: AWS::AutoScaling::AutoScalingGroup
    properties:
      AvailabilityZones: [nova]
      MinSize: 2
      MaxSize: 3
      DesiredCapacity: 3
      LaunchConfigurationName: your_config

3) run "stack-update -f template_file stack_test" >>>> failed

I think the expected result should be new resource "your_asg" create success, and 3 nova instances which belong to this resource create success.

case4:
1) base on case 3 step 1)
2) update this template, change the value of "DesiredCapacity", such as 1 --- > 2
3) run "stack-update -f template_file stack_test" >>>> failed

I think the expected result should be one nova instances which belong to "my_asg" create success.

the engine.log :

2014-09-30 18:08:29.029 21787 INFO heat.engine.resource [-] updating AutoScalingGroup "my_asg" [666353e9-537d-4ee1-8345-1e445f1c1120] Stack "xx" [98397eba-6c7f-4997-a9c3-cc5953378da7]
2014-09-30 18:08:29.055 21787 INFO heat.engine.environment [-] Registering OS::Heat::ScaledResource -> AWS::EC2::Instance
2014-09-30 18:08:29.100 21787 INFO heat.engine.resource [-] UPDATE: AutoScalingGroup "my_asg" [666353e9-537d-4ee1-8345-1e445f1c1120] Stack "xx" [98397eba-6c7f-4997-a9c3-cc5953378da7]
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource Traceback (most recent call last):
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/resource.py", line 435, in _action_recorder
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource yield
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/resource.py", line 690, in update
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource args=[after, tmpl_diff, prop_diff])
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/scheduler.py", line 286, in wrapper
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource step = next(subtask)
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/resource.py", line 476, in action_handler_task
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource handler_data = handler(*args)
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/resources/autoscaling.py", line 605, in handle_update
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource adjustment_type=EXACT_CAPACITY)
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/resources/autoscaling.py", line 654, in adjust
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource LOG.exception(_('Failed sending error notification'))
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/openstack/common/excutils.py", line 82, in __exit__
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource six.reraise(self.type_, self.value, self.tb)
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/resources/autoscaling.py", line 645, in adjust
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource self.resize(new_capacity)
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/resources/autoscaling.py", line 378, in resize
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource new_template = self._create_template(new_capacity)
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/resources/autoscaling.py", line 305, in _create_template
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource instance_definition = self._get_instance_definition()
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource File "/usr/lib/python2.6/site-packages/heat/engine/resources/autoscaling.py", line 281, in _get_instance_definition
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource props = function.resolve(conf.properties.data)
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource AttributeError: 'NoneType' object has no attribute 'properties'
2014-09-30 18:08:29.100 21787 TRACE heat.engine.resource

I don't see the my test way/step has issue, basically, I refer to http://docs.openstack.org/developer/heat/template_guide/ as the a guide to generate my test case. I think this is the heat bug, heat engine should show the reason for why update failed in log file so that the end user can take follow up action, thanks.

[Angus Salkeld (asalkeld)]

I have reproduced this and can add:

heat event-list st
+---------------+--------------------------------------+-----------------------------------------------------------------+--------------------+----------------------+
| resource_name | id | resource_status_reason | resource_status | event_time |
+---------------+--------------------------------------+-----------------------------------------------------------------+--------------------+----------------------+
| your_asg | 3420cf21-5145-4d93-950e-bbf521231b6e | AttributeError: 'NoneType' object has no attribute 'properties' | CREATE_FAILED | 2014-10-01T00:46:13Z |
| your_asg | 36a43554-93cd-40da-83d9-a3debd9ccfbf | state changed | CREATE_IN_PROGRESS | 2014-10-01T00:46:13Z |
| my_asg | d492c470-4e65-4e86-8465-64dcb5e28e38 | state changed | CREATE_COMPLETE | 2014-10-01T00:44:12Z |
| my_config | 0f3c76b9-a9bb-4da1-b467-a7e83d401600 | state changed | CREATE_COMPLETE | 2014-10-01T00:44:11Z |
| my_config | 873a6220-dd86-43fb-841e-716f23f220e5 | state changed | CREATE_IN_PROGRESS | 2014-10-01T00:44:11Z |
| my_asg | fd11f4de-6fc2-490b-9c3b-39cbbde5a5b3 | state changed | CREATE_IN_PROGRESS | 2014-10-01T00:44:11Z |
+---------------+--------------------------------------+-----------------------------------------------------------------+--------------------+----------------------+

I added the your_asg and your_conf in the update.
Notice that your_conf has not been created, so it looks like the update dependencies are not working correctly.

[Angus Salkeld (asalkeld)]

silly me, the template is wrong.

Please change the following:
LaunchConfigurationName: my_config
to
LaunchConfigurationName: {Ref: my_config}

This is why you are not getting a dependency between the auto-scaling group and the launchconfig.
See: http://docs.aws.amazon.com/AWSCloudFormation/latest/UserGuide/quickref-autoscaling.html

[Angus Salkeld (asalkeld)]

This is a common error, I could add a validation check for this.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/125251

[Mike Spreitzer (mike-spreitzer)]

Let's also update the documentation of the ASG resource type to be clear on this point, since the name is misleading.

[XiaoLiang Hu (xlhuxa)]

Hi, all:
Thanks for your timely reply, I change my template according to your suggestion, all the case1- case 4 work well now.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/125251
Committed: https://git.openstack.org/cgit/openstack/heat/commit/?id=bbe47783b587c6da2ba1a86e332fbd37096005c4
Submitter: Jenkins
Branch: master

commit bbe47783b587c6da2ba1a86e332fbd37096005c4
Author: Angus Salkeld <email address hidden>
Date: Tue Oct 21 10:48:28 2014 +1000

    Make sure that AutoScalingGroup depends on the launch config

    It seems to be common error to do the following:
      LaunchConfigurationName: my_config
    and not include a depends_on.

    Largely because of the unfortunate property name.
    So this patch tries to make sure the user has the
    correct dependency in place.

    Change-Id: I4f7d282d0af8681dc2c8cd265e16a27fe23a3bdb
    Closes-bug: #1375777