Nova服务重启之后过几分钟Down

检查NOVA的服务LOG如下：

2017-05-04 08:21:49.453 21323 WARNING nova.compute.monitors [req-84b9acc2-c0b2-4069-a573-5a1f29c8dbfb - - - - -] Excluding nova.compute.monitors.cpu monitor virt_driver. Not in the list of enabled monitors (CONF.compute_monitors).
2017-05-04 08:21:49.454 21323 INFO nova.compute.resource_tracker [req-84b9acc2-c0b2-4069-a573-5a1f29c8dbfb - - - - -] Auditing locally available compute resources for node compute1
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager [req-84b9acc2-c0b2-4069-a573-5a1f29c8dbfb - - - - -] Error updating resources for node compute1.
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager Traceback (most recent call last):
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 6460, in update_available_resource
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager     rt.update_available_resource(context)
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/compute/resource_tracker.py", line 500, in update_available_resource
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager     resources = self.driver.get_available_resource(self.nodename)
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 5394, in get_available_resource
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager     disk_over_committed = self._get_disk_over_committed_size_total()
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 7052, in _get_disk_over_committed_size_total
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager     local_instances[guest.uuid], bdms[guest.uuid])
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager KeyError: '0a07a562-67b5-42ab-bb6d-4bdd5a42abb5'
2017-05-04 08:21:49.685 21323 ERROR nova.compute.manager
2017-05-04 08:21:51.802 21323 WARNING nova.compute.manager [req-6d99d2b9-b14f-4e3c-80fe-10fa9db98b73 - - - - -] While synchronizing instance power states, found 2 instances in the database and 4 instances on the hypervisor.
2017-05-04 08:22:22.592 21323 INFO nova.compute.resource_tracker [req-6d99d2b9-b14f-4e3c-80fe-10fa9db98b73 - - - - -] Auditing locally available compute resources for node compute1
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager [req-6d99d2b9-b14f-4e3c-80fe-10fa9db98b73 - - - - -] Error updating resources for node compute1.
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager Traceback (most recent call last):
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 6460, in update_available_resource
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager     rt.update_available_resource(context)
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/compute/resource_tracker.py", line 500, in update_available_resource
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager     resources = self.driver.get_available_resource(self.nodename)
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 5394, in get_available_resource
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager     disk_over_committed = self._get_disk_over_committed_size_total()
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 7052, in _get_disk_over_committed_size_total
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager     local_instances[guest.uuid], bdms[guest.uuid])
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager KeyError: '0a07a562-67b5-42ab-bb6d-4bdd5a42abb5'
2017-05-04 08:22:22.801 21323 ERROR nova.compute.manager
2017-05-04 08:23:23.671 21323 INFO nova.compute.resource_tracker [req-6d99d2b9-b14f-4e3c-80fe-10fa9db98b73 - - - - -] Auditing locally available compute resources for node compute1
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager [req-6d99d2b9-b14f-4e3c-80fe-10fa9db98b73 - - - - -] Error updating resources for node compute1.
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager Traceback (most recent call last):
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 6460, in update_available_resource
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager     rt.update_available_resource(context)
2017-05-04 08:23:23.671 21323 INFO nova.compute.resource_tracker [req-6d99d2b9-b14f-4e3c-80fe-10fa9db98b73 - - - - -] Auditing locally available compute resources for node compute1
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager [req-6d99d2b9-b14f-4e3c-80fe-10fa9db98b73 - - - - -] Error updating resources for node compute1.
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager Traceback (most recent call last):
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 6460, in update_available_resource
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager     rt.update_available_resource(context)
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/compute/resource_tracker.py", line 500, in update_available_resource
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager     resources = self.driver.get_available_resource(self.nodename)
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 5394, in get_available_resource
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager     disk_over_committed = self._get_disk_over_committed_size_total()
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 7052, in _get_disk_over_committed_size_total
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager     local_instances[guest.uuid], bdms[guest.uuid])
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager KeyError: '0a07a562-67b5-42ab-bb6d-4bdd5a42abb5'
2017-05-04 08:23:23.827 21323 ERROR nova.compute.manager
2017-05-04 08:24:22.572 21323 INFO nova.compute.resource_tracker [req-6d99d2b9-b14f-4e3c-80fe-10fa9db98b73 - - - - -] Auditing locally available compute resources for node compute1
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager [req-6d99d2b9-b14f-4e3c-80fe-10fa9db98b73 - - - - -] Error updating resources for node compute1.
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager Traceback (most recent call last):
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 6460, in update_available_resource
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager     rt.update_available_resource(context)
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/compute/resource_tracker.py", line 500, in update_available_resource
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager     resources = self.driver.get_available_resource(self.nodename)
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 5394, in get_available_resource
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager     disk_over_committed = self._get_disk_over_committed_size_total()
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 7052, in _get_disk_over_committed_size_total
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager     local_instances[guest.uuid], bdms[guest.uuid])
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager KeyError: '0a07a562-67b5-42ab-bb6d-4bdd5a42abb5'
2017-05-04 08:24:22.728 21323 ERROR nova.compute.manager

问题解决：

Virsh list中的虚拟机数目与OpenStack数据库中的不匹配
使用

virsh destroy Domain
virsh undefine Domain

从物理计算节点上删除虚拟机
检查OpenStack数据库中的虚拟机数目与物理计算节点上的虚拟机数目是否一致