yahoo-eng-team team mailing list archive

[OpenStack Infra <1549401@xxxxxxxxxxxxxxxxxx>]

Reviewed:  https://review.openstack.org/284287
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=654f02948256dd6e3e1c021b5e613590d268a332
Submitter: Jenkins
Branch:    master

commit 654f02948256dd6e3e1c021b5e613590d268a332
Author: Matt Riedemann <mriedem@xxxxxxxxxx>
Date:   Wed Feb 24 12:33:01 2016 -0500

    Gracefully handle cancelling all events more than once
    
    Check to see if we can iterate over events if cancel_all_events
    is called more than once, so we can avoid a NoneType error.
    
    Change-Id: Ie0baeb17ffb1a6421bcc7d0fbcdf24afc6ede14c
    Closes-Bug: #1549401


** 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/1549401

Title:
  compute: calling cancel_all_events twice raises NoneType error

Status in OpenStack Compute (nova):
  Fix Released

Bug description:
  Seen in running the nova functional tests:

  2016-02-24 11:50:26,686 INFO [nova.osapi_compute.wsgi.server] 127.0.0.1 "POST /v2.1/openstack//servers/6adf14e3-5a6b-4562-a3c9-e27d1c59a81e/action HTTP/1.1" status: 400 len: 490 time: 0.0028141
      2016-02-24 11:50:26,687 ERROR [nova.service] Service error occurred during cleanup_host
      Traceback (most recent call last):
        File "nova/service.py", line 301, in stop
          self.manager.cleanup_host()
        File "nova/compute/manager.py", line 1329, in cleanup_host
          self.instance_events.cancel_all_events()
        File "nova/compute/manager.py", line 586, in cancel_all_events
          for instance_uuid, events in our_events.items():
      AttributeError: 'NoneType' object has no attribute 'items'
      2016-02-24 11:50:26,743 ERROR [nova.service] Service error occurred during cleanup_host
      Traceback (most recent call last):
        File "nova/service.py", line 301, in stop
          self.manager.cleanup_host()
        File "nova/compute/manager.py", line 1329, in cleanup_host
          self.instance_events.cancel_all_events()
        File "nova/compute/manager.py", line 586, in cancel_all_events
          for instance_uuid, events in our_events.items():
      AttributeError: 'NoneType' object has no attribute 'items'
      2016-02-24 11:50:26,745 WARNING [nova.service] Service killed that has no database entry
      2016-02-24 11:50:26,745 ERROR [nova.service] Service error occurred during cleanup_host
      Traceback (most recent call last):
        File "nova/service.py", line 301, in stop
          self.manager.cleanup_host()
        File "nova/compute/manager.py", line 1329, in cleanup_host
          self.instance_events.cancel_all_events()
        File "nova/compute/manager.py", line 586, in cancel_all_events
          for instance_uuid, events in our_events.items():
      AttributeError: 'NoneType' object has no attribute 'items'
      2016-02-24 11:50:26,746 WARNING [nova.service] Service killed that has no database entry
      2016-02-24 11:50:26,885 INFO [nova.wsgi] Stopping WSGI server.
      2016-02-24 11:50:26,885 INFO [nova.osapi_compute.wsgi.server] (15418) wsgi exited, is_accepting=True

  This is because the tests are using the ServiceFixture to start a
  service which has it's own cleanup callback to kill the service, but
  some tests are redundantly killing the same service when tearing down.
  This results in the NoneType.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1549401/+subscriptions