compute periodic task _poll_bandwidth_usage fails if no network attached to vif

[Brian Waldon]

I saw the following traceback in my compute manager log:

2011-12-07 17:26:09,500 ERROR nova.manager [-] Error during ComputeManager._poll_bandwidth_usage: 'VirtualInterface' object has no attribute 'network'
(nova.manager): TRACE: Traceback (most recent call last):
(nova.manager): TRACE: File "/usr/lib/python2.7/dist-packages/nova/manager.py", line 155, in periodic_tasks
(nova.manager): TRACE: task(self, context)
(nova.manager): TRACE: File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 1847, in _poll_bandwidth_usage
(nova.manager): TRACE: vif.network.label,
(nova.manager): TRACE: AttributeError: 'VirtualInterface' object has no attribute 'network'

[Chris Behrens (cbehrens)]

I just saw this too! Although it's not because there's no network attached to vif... it's because network was untied from vif in the models.

[Chris Behrens (cbehrens)]

Shoot. xenapi's get_all_bw_usage() does a db.virtual_interface_get_by_address() which now needs to be a call to the network manager.

[Jason Kölker (jason-koelker)]

https://review.openstack.org/#change,2267

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/3644

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/3644
Committed: http://github.com/openstack/nova/commit/9b1b65a4395511b07e15a521de3a3a66c4bdcfaa
Submitter: Jenkins
Branch: master

commit 9b1b65a4395511b07e15a521de3a3a66c4bdcfaa
Author: Jason Kölker <email address hidden>
Date: Mon Dec 12 16:41:40 2011 -0600

    Fix _poll_bandwidth_usage if no network on vif

    Bugfix for lp901819

    Move DB lookups out of the virt layer for bw usage.
    Migrate to storing the mac address in the table.
    Only lookup network label when sending usage events.
    Attempt to use info_cache for label lookup before
        punting to rpc call.

    Change-Id: Iffb4bc220804be7d5fe26d6416b90cd747b72753