AutoScalingGroup ignore "DesiredCapacity" in some conditions when updating resource

[Zhang Yang]

Update a AutoScalingGroup, "DesiredCapacity" is ignored in some conditions.

1. After auto scale, the value setted in creation will be ignored. For example:
    when creating group, the value of "DesiredCapacity" is 4, and after autoscale,
    the current capacity is 6. If you use 4 to update the group, the group will ignore it.

2. "DesiredCapacity" will be ignored if the value is zero.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/87204

[Zhang Yang (neil-zhangyang)]

I'm working on condition 1 for one day, but still have no idea. when you update the resource, the updater compares the old
template with the new template to decide whether the resource should be updated, but when autoscalegroup adjusts, the
stored template wouldn't be updated, and I don't think it is a good idea to update raw_template directly. Does any have a idea
about it?

[Zane Bitter (zaneb)]

The documentation for "DesiredCapacity" says that it is the "Desired initial number of instances". So I don't think it is a bug that it is ignored on updates when it hasn't changed.

[Openstack Gerrit (openstack-gerrit)]

Reviewed: https://review.openstack.org/87204
Committed: https://git.openstack.org/cgit/openstack/heat/commit/?id=a41442e8737b19b02f5c18cf004bd95e9bbde006
Submitter: Jenkins
Branch: master

commit a41442e8737b19b02f5c18cf004bd95e9bbde006
Author: Zhang Yang <email address hidden>
Date: Mon Apr 14 00:25:46 2014 -0700

    Allow DesiredCapacity to be zero

    When updating or validating a AutoScaleGroup,
    'DesiredCapacity' is ignored if the value is 0.

    - Add validation for 'DesiredCapacity' if the value is 0,
    - Don't ignore zero when updating.

    Change-Id: I830aaf97ebed6b2c77c72b18dc8d222ce23f9a28
    Partial-Bug: #1304423

[Steven Hardy (shardy)]

What's the status of this, the patch above is marked Partial-Bug but no subsequent patch has been proposed - is this bug fix actually completed?

[Zhang Yang (neil-zhangyang)]

shardy：
       To fix this, we may need to change the resource update logic .And then talked with zaneb, he didn't think It's a bug.

[Steven Hardy (shardy)]

Ref comment #6 I'm going to consider https://review.openstack.org/#/c/87204/ as the completed fix for this bug, and mark the bug fix committed.

Zhang Yang if you disagree with zaneb and believe there is additional work to do, please raise a new bug with details of what needs to be done, thanks!