resume stack failed after suspend a stack at the same time autoscaling updating the nested stack

[wujiajie618]

the stack has resouce of autoscalinggroup, when suspending a stack, there is no check about the nested stack status, so if autoscaling group add one instance, at the same time suspend the father stack, the new instance status is in active, no suspend, resource status is CREATE COMPLETE, so then resume the stack will failed because the new instance resource status is CREATE COMPLETE

[Anant Patil (ananta)]

wujiajie618, could you please upload the template? Also, if you could give more details on how to reproduce, it will be helpful.

[Angus Salkeld (asalkeld)]

Please confirm that this will fix the issue:
https://github.com/openstack/heat/commit/664f70383bfdd704322e136de226e9627c48d1a7

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/169576

[Angus Salkeld (asalkeld)]

Once the above jenkins job passes, we can mark this bug as Fix Committed.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/169576
Committed: https://git.openstack.org/cgit/openstack/heat/commit/?id=bafb803a7fb4d00602fa373de222da7e79272311
Submitter: Jenkins
Branch: master

commit bafb803a7fb4d00602fa373de222da7e79272311
Author: Angus Salkeld <email address hidden>
Date: Wed Apr 1 11:22:40 2015 +1000

    Prove that we can't signal during a suspend

    This is to just make sure that bug 1334333 is fixed.

    Change-Id: I7a54e7f089b247997fb5cc5e696629f6415d009f
    Partial-bug: #1334333

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/169725

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/169725
Committed: https://git.openstack.org/cgit/openstack/heat/commit/?id=9468e87e417f3b0d2fde5979fc4eb5b5027a4970
Submitter: Jenkins
Branch: master

commit 9468e87e417f3b0d2fde5979fc4eb5b5027a4970
Author: Angus Salkeld <email address hidden>
Date: Thu Apr 2 09:14:39 2015 +1000

    Create an event when signalling fails

    - otherwise it is hard to know that signalling happened and failed
    - remove the LOG.exception() - doesn't seem like a programming error.
    - don't change the resource action/state
    - this makes a functional test much more reliable

    Change-Id: I84c73b06bade4dfcadba240eb3dddac698194acf
    Partial-bug: #1334333