yahoo-eng-team team mailing list archive

[Launchpad Bug Tracker <1298061@xxxxxxxxxxxxxxxxxx>]

This bug was fixed in the package nova - 1:2014.1.5-0ubuntu1.6

---------------
nova (1:2014.1.5-0ubuntu1.6) trusty; urgency=medium

  * Allow evacuate for an instance in the Error state (LP: #1298061)
    - d/p/remove_useless_state_check.patch remove unnecessary task_state check
    - d/p/evacuate_error_vm.patch Allow evacuate from error state

 -- Liang Chen <liang.chen@xxxxxxxxxxxxx>  Fri, 09 Sep 2016 17:41:48
+0800

** Changed in: nova (Ubuntu Trusty)
       Status: Fix Committed => 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/1298061

Title:
  nova should allow evacuate for an instance in the Error state

Status in Ubuntu Cloud Archive:
  Fix Released
Status in Ubuntu Cloud Archive icehouse series:
  In Progress
Status in OpenStack Compute (nova):
  Fix Released
Status in nova package in Ubuntu:
  Fix Released
Status in nova source package in Trusty:
  Fix Released

Bug description:
  [Impact]

   * Instances in error state cannot be evacuated.

  [Test Case]

   * nova evacuate <error_state_instance> <another_compute_host>
   * nova refuses to evacuate the instance because of its state

  [Regression Potential]

   * Cherry picked from upstream
     - removed unnecessary argument passing
     - add allowing ERROR state before evacuating.
   * actually, in code, added one parameter, and removed unused one.
     so very low regression possibility.
   * Tested on juju+maas test env.
   * Passed tempest smoke tests locally.

  Note: one simple way to put an instance into error state is to
  directly change its database record, for example "update instances set
  vm_state='error' where uuid='XXXXXXXX'"

  We currently allow reboot/rebuild/rescue for an instance in the Error
  state if the instance has successfully booted at least once.

  We should allow "evacuate" as well, since it is essentially a
  "rebuild" on a different compute node.

  This would be useful in a number of cases, in particular if an initial
  evacuation attempt fails (putting the instance into the Error state).

To manage notifications about this bug go to:
https://bugs.launchpad.net/cloud-archive/+bug/1298061/+subscriptions