Cannot delete vm instance if send duplicate delete requests

[zhangjialong]

I deployed openstack with icehouse rc1 and booted 100 vms on my nodes. After my testing, i tried to delete my vms at the same time. Then i fount all of my vms` status change to deleting but cannot be deleted. I checked my openstack, the rabbitmq-server crashed . Then i restart rabbitmq-server and my openstack nova services, sended the delete requests again and again, the vms still cannot be deleted. While , in havana, the vms can be deleted if received duplicate delete requests .
I think icehouse should handle duplicate delete requests like havana .

[Openstack Gerrit (openstack-gerrit)]

Fix proposed to branch: master
Review: https://review.openstack.org/87868

[Takahiro Shida (shida)]

Duplicated delete request fixed with a purpose following report.
https://bugs.launchpad.net/nova/+bug/1248563

I think that the ignoring the duplicate delete is raise such as this problem. So, we should discuss more about this problem.

[Tushar Patil (tpatil)]

IMO, there are 2 ways to fix this problem.
1. nova-api, remove the check to allow instance deletion when the task_state is DELETING.
2. nova-compute, add a new periodic task to delete the instances whose task_state is deleting.

[zhangjialong (zhangjl)]

@Tushar Patil

Aggree with you. Your suggestion is better than mine.

[OpenStack Infra (hudson-openstack)]

Change abandoned by zhangjl (<email address hidden>) on branch: master
Review: https://review.openstack.org/87868
Reason: According to Tushar Patil`s suggestion, I think we should fix this problem by periodic task

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/121800

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/121800
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=222d44532c65ddf3f26532ced217890628352536
Submitter: Jenkins
Branch: master

commit 222d44532c65ddf3f26532ced217890628352536
Author: Rajesh Tailor <email address hidden>
Date: Mon Sep 15 07:20:14 2014 -0700

    Allow force-delete irrespective of VM task_state

    Cannot delete vm instance if duplicate delete requests are sent.

    When user sends request to delete an instance, its task_state gets
    changed to 'deleting' state. When an instance task_state is already in
    'deleting' state and if at that moment the rabbitmq-server get crashed
    by some reasons, then the instance task_state remains in 'deleting'
    state and user won't be able to delete the instance forever.

    At this moment, there is only one way to delete the instances, whose
    task_state is in 'deleting' state, by restarting the nova-compute
    services where these instances are running.

    To avoid restarting the nova-compute service manually, modified the
    force-delete api to allow instance deletion irrespective of
    instance task_state.

    Added new module to use delete_types as constants and replaced all
    delete_type string occurances with new constants.

    Closes-Bug: #1308342

    Change-Id: I4d0e47662a80109ef9622d85455587d487e47c01

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/juno
Review: https://review.openstack.org/131948

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/163798

[OpenStack Infra (hudson-openstack)]

Change abandoned by Matt Riedemann (<email address hidden>) on branch: stable/juno
Review: https://review.openstack.org/131948
Reason: This broke cells RPC API on master (kilo) so we had to revert this in kilo and then fix the bug, so no point in backporting this to juno.