Tech Debt: Store stack_id in component after stack create

[Adrian Otto]

See: [1] https://review.openstack.org/89935

When we create a new heat stack, we create a component to represent it within the assembly. On git push, the trigger workflow gets called. It then then calls the deployer service which currently tries to redeploy the assembly by creating a new heat stack instead of updating the existing one. The review[1] above works around a "stack already exists" error by using the Heat API to look up the Heat stack, and if it exists, to update it.

We should not query Heat. Instead we should store the stack_id in the component resource that is part of the assembly being updated. We should do a select on assembly+components (joined on assembly_id) to determine if we know the stack id, and if we do, then update it. If there is no stack_id yet, then create one.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/96928

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/96928
Committed: https://git.openstack.org/cgit/stackforge/solum/commit/?id=6c9cfe93de3dce164eda9f68d95f6d88b1a3bdd4
Submitter: Jenkins
Branch: master

commit 6c9cfe93de3dce164eda9f68d95f6d88b1a3bdd4
Author: arati.mahimane <email address hidden>
Date: Fri May 30 15:13:23 2014 -0500

    Store heat_stack_id and component_type in component table

    Added two attributes to component - heat_stack_id and
    component_type.

    When we need to update a heat stack (while updating
    an assembly), the stack_id can be obtained from the
    component table rather than querying the Heat API.
    However, during assembly delete we still need to
    query the Heat API to ensure that the stack is
    actually deleted before deleting the assembly.

    Closes-Bug: #1312856

    Change-Id: I6ebf9b32d80d6c9f5f51e7f77d7c5ff344ef29ef