Comment 3 for bug 1021340

Reviewed: https://review.openstack.org/9630
Committed: http://github.com/openstack/nova/commit/34f9d7e974d0e09c723e0a04ed6eecd1b482e77d
Submitter: Jenkins
Branch: master

commit 34f9d7e974d0e09c723e0a04ed6eecd1b482e77d
Author: Phil Day <email address hidden>
Date: Wed Jul 11 09:26:11 2012 +0100

    Convert remaining network API casts to calls

    Fix for Bug 1021340

    During compute/terminate_instance networking is de-allocated by a call to
    network/api/deallocate_for_instance(), which is implemented as an rpc.cast to
    the network manager. This in turn will eventually call
    network/manager/deallocate_fixed_ip(), which in turn call the compute API to
    trigger a security group refresh, which will get the instance from the database

    However because original call to the network manager is a cast there is a chance
    that the compute manager will delete the instance record in the DB before the
    compute API (in the network manager) tries to retrieve the instance. At this
    point the security group refresh fails (leaving rules in place which are a
    security risk when the IP is reused), and potentially stopping otheraspects
    of the network deallocation from completing.

    Changing this from rpc.call to rpc.cast will fix this issue.

    Aside from this specific use of a cast there are 4 other casts in the
    network API:
      add_fixed_ip_to_instance
      remove_fixed_ip_from_instance
      add_network_to_project
      release_floating_ip

    and to avoid other timing issues these will also be converted to calls.

    Change-Id: I5cdcc628293d3e7cf165c5ffe4883f138783f73f