no event recorded for INIT_* resources

[Robert Collins]

We have a resource that is getting stuck on INIT_COMPLETE. There are no events visible in event-list <stack> for that resource (or any others) for the INIT state - I'm wondering if this is deliberate or oversight (because - its hard to debug the problem using the API without such events).

[Steven Hardy (shardy)]

This is a similar issue to https://bugs.launchpad.net/heat/+bug/1273796, which I started looking at but haven't had time to revisit.

Basically we need to store the state regardless of what happens at INIT time, and probably generate an event via the storage of that state.

I guess it's more of a historical known issue than an oversight, because "INIT" was never really a thing until https://review.openstack.org/#/c/36385/ landed a while back.

Before then, the first transition was from the resource not existing to CREATE_IN_PROGRESS, but since then we have a slightly weird interim state which is unrelated to the state of the actual thing backing the resource.

I'll see if I can get back to 1273796 soon and solve it in a way which also fixes this.

[Zane Bitter (zaneb)]

INIT_COMPLETE is a Claytons state.

[Zane Bitter (zaneb)]

This will actually be useful for convergence, since that will require us to store all of the resources in the DB before kicking off the stack creation (unlike now, where they are created when we begin creating each actual, physical resource).

However, I don't think this will be as helpful for debugging as the bug report assumed, because INIT_COMPLETE isn't a state, and doesn't correspond to any particular activity or roll up exceptions from it. Therefore, I bumped this bug to kilo-1.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/164651

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/165622

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/165623

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/165624

[OpenStack Infra (hudson-openstack)]

Change abandoned by Steven Hardy (<email address hidden>) on branch: master
Review: https://review.openstack.org/165624
Reason: Obsoleted by https://review.openstack.org/#/c/165686/

[OpenStack Infra (hudson-openstack)]

Change abandoned by Steven Hardy (<email address hidden>) on branch: master
Review: https://review.openstack.org/164651
Reason: Obsoleted by https://review.openstack.org/#/c/165686/

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/165686
Committed: https://git.openstack.org/cgit/openstack/heat/commit/?id=073a9d34048a29fab099c2886c274eacf94301dd
Submitter: Jenkins
Branch: master

commit 073a9d34048a29fab099c2886c274eacf94301dd
Author: Zane Bitter <email address hidden>
Date: Wed Mar 18 21:30:31 2015 -0400

    Store Resources before they are created

    Store all Resources in an initial INIT_COMPLETE state prior to creating a
    stack, rather than waiting until each Resource is create()d. When updating
    a stack, store each new Resource just prior to create()ing it.

    Change-Id: Ifeee8213c1962cbf45fd0dd59f1832b0eeff1a47
    Closes-Bug: #1319813

[OpenStack Infra (hudson-openstack)]

Change abandoned by Steven Hardy (<email address hidden>) on branch: master
Review: https://review.openstack.org/165623

[OpenStack Infra (hudson-openstack)]

Change abandoned by Steven Hardy (<email address hidden>) on branch: master
Review: https://review.openstack.org/165622

[Zane Bitter (zaneb)]

Note that while the patch provides what was requested, the real issues here are bug #1492427 and bug #1492433