yahoo-eng-team team mailing list archive

[OpenStack Infra <1650328@xxxxxxxxxxxxxxxxxx>]

Reviewed:  https://review.openstack.org/411440
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=5eb8297f6144f2f941cab92c6298f6eaebafa202
Submitter: Jenkins
Branch:    master

commit 5eb8297f6144f2f941cab92c6298f6eaebafa202
Author: int32bit <krystism@xxxxxxxxx>
Date:   Fri Dec 16 02:13:52 2016 +0800

    Add soft delete wrinkle to api-ref
    
    The API reference for deleting a server is making the assumption that
    the instance is going to be (hard) deleted from the system, but it
    doesn't take into account the case that CONF.reclaim_instance_interval
    is set, which if > 0 makes the API 'soft delete' the instance, so that
    the instance's vm won't be deleted until the
    CONF.reclaim_instance_interval has expired since the instance was soft
    deleted. The instance marked as ``SOFT_DELETED`` can be recovered via
    ``restore`` action before it's really deleted from the system. We should
    update the reference to talk about the soft deleted wrinkle so we can set
    expectations.
    
    Closes-Bug: #1650328
    
    Change-Id: Ibbb1c225da11234aaac3634769db112a30dfef64


** Changed in: nova
       Status: In Progress => Fix Released

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1650328

Title:
  The delete server API reference doesn't mention the SOFT_DELETED
  wrinkle

Status in OpenStack Compute (nova):
  Fix Released

Bug description:
  http://developer.openstack.org/api-ref/compute/?expanded=delete-
  server-detail#delete-server

  The API reference for deleting a server is making the assumption that
  the instance is going to be (hard) deleted from the system, but it
  doesn't take into account the case that CONF.reclaim_instance_interval
  is set, which if > 0 makes the API 'soft delete' the instance, so that
  the instance's vm status is set to 'SOFT_DELETED' and at some later
  time (after the CONF.reclaim_instance_interval has expired since the
  instance was soft deleted), a periodic task on the compute node will
  delete the instance.

  So we should update the reference to talk about the soft deleted
  wrinkle so we can set expectations, and it's also just in general kind
  of a gross API behavior change based on configuration, but that's
  another story.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1650328/+subscriptions