yahoo-eng-team team mailing list archive

[melanie witt <1389899@xxxxxxxxxxxxxxxxxx>]

I agree with Alex -- this sounds like a semantic change. Currently in
nova, delete is always allowed to succeed i.e. we soft delete, even when
the compute node is unavailable. Quota will be returned to the user and
the ip address will be freed. The idea behind this is that for the user,
it should be transparent to them whether or not the compute node their
instance is on is up or not. They should anyway be allowed to delete
their instance as normal, and nova will handle reaping the soft deleted
instance when the compute node is available again or when the operator
migrates it to an up node.

** Changed in: nova
   Importance: Undecided => Wishlist

** Changed in: nova
       Status: New => Opinion

** Changed in: nova
       Status: Opinion => Confirmed

-- 
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/1389899

Title:
  nova delete shouldn't remove instance from DB if host is not up

Status in OpenStack Compute (Nova):
  Confirmed

Bug description:
  Under nova/compute/api.py, it will delete instance from DB if compute
  node is not up. I think we should utilize force-delete to handle
  compute node is not up scenario. So, if compute node is not up, only
  force-delete can delete the instance.

  Code flow:
  delete -> _delete_instance -> _delete

  _delete(...) code snippet: 
  ..
              if not is_up:
                  # If compute node isn't up, just delete from DB
                  self._local_delete(context, instance, bdms, delete_type, cb)
                  quotas.commit()

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1389899/+subscriptions